Serving Two Masters, Yet out of Control

Serving Two Masters, Yet Out of Control

Serving Two Masters, Yet Out of Control Fannie Mae and Freddie Mac

Peter J. Wallison, editor

The AEI Press Publisher for the American Enterprise Institute WASHINGTON, D.C.

2001 Available in the United States from the AEI Press, c/o Publisher Resources Inc., 1224 Heil Quaker Blvd., P.O. Box 7001, La Vergne, TN 37086-7001. To order, call 1-800-269-6267. Distributed outside the United States by arrange- ment with Eurospan, 3 Henrietta Street, London WC2E 8LU, England.

Library of Congress Cataloging-in-Publication Data

Serving two masters, yet out of control : Fannie Mae and Freddie Mac / Peter J. Wallison, editor. p. cm. Papers from three AEI conferences. ISBN 0-8447-4166-3 (pbk.) 1. Fannie Mae—Congresses. 2. Federal Home Loan Mortgage Corporation—Congresses. 3. Secondary mortgage market—United States—Congresses. 4. Mortgage loans—United States—Congresses. 5. Housing—United States—Finance—Congresses. I. Wallison, Peter J. II. American Enterprise Institute for Public Policy Research.

HG2040.5.U5 S473 2001 332.7’22—dc21 2001045196

13579108642

© 2001 by the American Enterprise Institute for Public Policy Research, Wash- ington, D.C. All rights reserved. No part of this publication may be used or reproduced in any manner whatsoever without permission in writing from the American Enterprise Institute except in cases of brief quotations embodied in news articles, critical articles, or reviews. The views expressed in the publications of the American Enterprise Institute are those of the authors and do not necessarily reflect the views of the staff, advisory panels, officers, or trustees of AEI.

THE AEI PRESS Publisher for the American Enterprise Institute 1150 17th Street, N.W., Washington, D.C. 20036

Printed in the United States of America Contents

CONTRIBUTORS vii

INTRODUCTION 1 1 Peter J. Wallison

PART ONE BASICS

ESTIMATING THE VALUE AND ALLOCATION OF FEDERAL SUBSIDIES 8 2 Robert S. Seiler Jr.

THE ECONOMICS OF FANNIE MAE AND FREDDIE MAC 41 3 Robert Van Order

FEDERAL DEPOSIT INSURANCE AND FEDERAL SPONSORSHIP OF 4 FANNIE MAE AND FREDDIE MAC: THE STRUCTURE OF SUBSIDY 56 Richard Scott Carnell

AN ECONOMIST’S CASE FOR GSE REFORM 85 5 Charles W. Calomiris

PART TWO CONTROLS AND RISKS

HOW FANNIE AND FREDDIE INFLUENCE THE POLITICAL PROCESS 110 6 Ralph Nader

SETTING GSE POLICY THROUGH CHARTERS, LAWS, 7 AND REGULATIONS 120 John C. Weicher

v vi CONTENTS

IMPROVING CONTROL OVER FANNIE MAE 8 AND FREDDIE MAC 139 Ron Feldman

REFORM OF GSE HOUSING GOALS 153 9 Jonathan Brown

PART THREE IDEAS FOR CHANGE

THE PRIVATIZATION OF SALLIE MAE 170 10 Mark Overend

AN ALTERNATIVE APPROACH TO GSE REFORM 11 THROUGH NATIONAL MORTGAGE ASSOCIATIONS 175 Thomas H. Stanton Contributors

JONATHAN BROWN has been the director of the Geographic Informa- tion System and Financial Research Project at Essential Information since 1991. The project conducts research and publishes reports on consumer protection regulation in the financial services sector, access to housing and small business credit, community reinvestment performance of banking institutions, and similar concerns. From 1989 to 1991 Mr. Brown was the director of Research on Financial Democ- racy, a consulting firm to nonprofit organizations on financial regulation issues. From 1985 to 1988 he was the director of BankWatch, a consumer advocacy project of the Center for Study of Responsive Law. Mr. Brown has served on several federal advisory committees including the GSE Working Group, Department of Housing and Urban Development; Consumer Advisory Council, Board of Gover- nors of the Federal Reserve System; and Advisory Committee on National Banks, Comptroller of the Currency.

CHARLES W. CALOMIRIS is the Paul M. Montrone Professor of Finance in the Division of Finance and Economics and the director of the Program on Financial Institutions, both at Columbia University’s Graduate School of Business. He is a visiting scholar and the codirector of the Project on Financial Deregulation at the American Enterprise Institute and a research associate of the National Bureau of Economic Research. Mr. Calomiris has written papers and books on financial institutions, economics, and history. He has been a con- sultant on financial regulation for the Federal Reserve Board; Fed- eral Reserve Banks of New York, Chicago, and St. Louis; World Bank; Central Bank of Argentina; and governments of Mexico, El Salva- dor, China, and Japan.

RICHARD SCOTT CARNELL is an associate professor of law at Fordham University School of Law. From 1993 to 1999 he was the assistant

vii viii CONTRIBUTORS secretary for financial institutions at the Treasury Department, where he helped secure legislation authorizing interstate banking and branching, establishing community development financial institutions, and reducing the costs and improving the quality of depository institution regulation. He also worked on Treasury’s financial modernization proposal. From 1987 to 1993 he was the counsel and then senior counsel to the Senate Committee on Banking, Housing, and Urban Affairs, where he helped shape legislation reforming thrift regulation and federal deposit insurance. After a private law practice in San Francisco, he was an attorney for the Board of Governors of the Federal Reserve System.

RON FELDMAN is an assistant vice president at the Federal Reserve Bank of Minneapolis, where he oversees applications, surveillance, and special studies. His research focuses on government-sponsored enterprises, deposit insurance, and federal budgeting, as well as trends affecting financial institutions and banking supervision policies; he has published several articles in that area. Mr. Feldman had earlier analyzed federal credit and insurance programs for the Con- gressional Budget Office.

RALPH NADER is a consumer activist. Time magazine has named him one of the 100 most influential Americans of the twentieth century. He has worked for the creation of the Occupational Safety and Health Administration, the Environmental Protection Agency, and the Con- sumer Product Safety Commission and has advocated many laws, including the Safe Drinking Water Act and the Freedom of Informa- tion Act. Mr. Nader has formed numerous citizen groups, including the Center for Auto Safety, Public Citizen, Pension Rights Center, Coalition for Universities in the Public Interest, and student public interest research groups (PIRGs). In 1992, 1996, and 2000 Mr. Nader organized presidential campaigns to challenge the duopoly of the two-party system. He has written many books, including Unsafe at Any Speed.

MARK OVEREND is a senior vice president of SLM Holding Corpora- tion (Sallie Mae). He joined Sallie Mae in 1986 as a director in the controller’s division and was promoted to assistant vice president (1988), vice president (1989), controller (1990), and senior vice president and chief financial officer (1997). His responsibilities have included oversight of capital market activity, financial planning and CONTRIBUTORS ix reporting, and tax planning and compliance functions. In March 2000 Mr. Overend assumed the leadership of the e-commerce and Internet initiatives at Sallie Mae. He had been on the staff of Arthur Young & Company in Washington, D.C., for seven years.

ROBERT S. SEILER JR. is the manager of policy analysis at the Office of Federal Housing Enterprise Oversight, the safety and soundness regulator of Fannie Mae and Freddie Mac. From 1986 to 1994 and 1997 to 2000 he was a senior analyst with the Congressional Budget Office, where he specialized in issues related to controlling and measuring the cost of federal credit programs, especially government- sponsored enterprises. Mr. Seiler was the principal author of CBO’s 1991 report on safety and soundness regulation of GSEs and has written about federal subsidies to Fannie Mae and Freddie Mac and federal regulation of those two enterprises. He has also written CBO studies on federal credit support for the financing of U.S. exports, infrastructure, and higher education and on the federal budget process and concepts. From 1994 to 1996 Mr. Seiler was a senior policy analyst at OFHEO. Before joining CBO, he worked for the Senate Budget Committee, for two members of the House of Representa- tives, and in the private sector.

THOMAS H. STANTON, an attorney in Washington, D.C., provides legal and policy counsel relating to the design and operation of public institutions and the delivery of public services, particularly federal credit and benefits programs, government corporations, and financial regulation. He is a fellow of the Center for the Study of Ameri- can Government at Johns Hopkins University, where he teaches the law of public institutions. Mr. Stanton is the chairman of the Stand- ing Panel on Executive Organization and Management of the National Academy of Public Administration. His writings include A State of Risk (1991) about government-sponsored enterprises. He is a member of the advisory board of the Financier: Analyses of Capi- tal and Money Market Transactions.

ROBERT VAN ORDER was named the chief economist at Freddie Mac in June 1987 with primary responsibility in research. Other responsibilities include macroeconomic analysis and forecasting and work with the media on developments in mortgage and other markets. Mr. Van Order had been the director of the Housing Finance Analy- sis Division at the Department of Housing and Urban Development x CONTRIBUTORS and a visiting professor of real estate at the Graduate School of Man- agement, University of California, Los Angeles.

PETER J. WALLISON joined the American Enterprise Institute in Janu- ary 1999 as a resident fellow and as the codirector of AEI’s program on financial market deregulation. As a partner of Gibson, Dunn & Crutcher LLP, he had practiced banking, corporate, and financial law in the firm’s Washington and New York offices. As the general counsel of the Treasury Department from 1981 to 1985, Mr. Wallison helped develop the Reagan administration’s proposals for deregulating the financial services industry. During 1986 and 1987 he was the counsel to President Ronald Reagan. He is the author of Back from the Brink and the coauthor of Nationalizing Mortgage Risk: The Growth of Fannie Mae and Freddie Mac and The GAAP Gap: Corpo- rate Disclosure in the Age of the Internet (all published by AEI Press).

JOHN C. WEICHER is the director of urban policy studies with the Hudson Institute. From 1997 to 1999 he was the project director for Hudson’s Michigan Urban Policy Initiative, which designed the first state urban homeownership strategy in the United States. Before joining Hudson in 1993, he was the assistant secretary for policy development and research at the Department of Housing and Urban De- velopment for four years. Mr. Weicher was the chief economist at HUD from 1975 to 1977 and at the Office of Management and Bud- get from 1987 to 1989. He is a past president of the American Real Estate and Urban Economics Association and received the association’s annual award for career achievement in 1993. Mr. Weicher served on the Committee on Urban Policy of the National Academy of Sciences and on the Advisory Committee on Popula- tion Statistics of the U.S. Census Bureau. He held the F. K. Weyerhaeuser Chair in Public Policy Research at AEI and has directed the Housing and Financial Markets Program at the Urban Institute. Mr. Weicher has written and edited twelve books and numerous articles. 1 Introduction

Peter J. Wallison

In 1999 and 2000 the American Enterprise Institute sponsored three conferences on the Federal National Mortgage Association (Fannie Mae) and the Federal Home Loan Mortgage Corporation (Freddie Mac), each exploring a different aspect of the public policy problem presented by these government-sponsored enterprises (GSEs). The series began with a conference entitled Fannie Mae and Freddie Mac: Public Mission and Private Interests; in a sense this title encapsulates the contradiction that lies at the root of the controversy over these organizations. The GSEs have two ultimately irreconcilable roles. In effect they serve two masters. As publicly owned corporations, they are required to maximize profitability for their shareholders; as government agencies, they have been assigned public missions for which they receive an implicit government subsidy. The size of this subsidy and how it is divided between their public and private objectives, the risks that those objectives create for taxpayers, and the agencies’ resistance to efforts by any regulatory body—or even Congress—to gain control over their activities exemplify the internal tensions created by their conflicting objectives. Unsurprisingly these issues should also be the focus of any serious inquiry into whether these organizations should retain the favored positions that have allowed them to grow to unprecedented size, to realize extraordinary profitability, and to achieve unparal- leled influence over the political process. Because they address the key issues most directly, the essays in this volume were selected from the many excellent contributions to the three AEI conferences. Contradictory goals lead to contradictory results. The Congres- sional Budget Office estimated in 1996 that Fannie and Freddie were simply dividing their subsidy between their shareholders and the

1 2 INTRODUCTION mortgage markets they were created to serve; about two-thirds was being used to lower mortgage rates, while the remaining third was used to make both companies highly profitable. CBO found that about 42 percent of the agencies’ profitability in 1995 resulted from government backing in that year. The Treasury Department prepared a study with similar results in the same year. In 2001 CBO did a somewhat more sophisticated study. This time the subsidy was even larger, 10.6 billion, and the percentages that went to shareholders rose to almost 37. All the essays in this volume were prepared before this second CBO report was made public. Two analyses delivered at the first conference were particularly noteworthy. Robin Seiler, then at CBO, reviewed and expanded on the CBO and Treasury findings. He noted that both studies under- stated the size of the implicit subsidy because they focused only on the subsidy’s effect in lowering the agencies’ borrowing costs. Ad- ditional subsidy may be implicit in the fact that they function with far lower capital ratios than other financial intermediaries. At the same conference Robert Van Order, chief economist at Freddie Mac, argued that the GSEs did not substantially differ from national banks. Van Order pictured the housing market as a case of “dueling charters” with two kinds of GSEs—national banks on one hand and Fannie Mae and Freddie Mac on the other—performing different roles. Van Order did not deny that the GSEs received some government backing but contended that they used their support more efficiently than national banks and thus—to the extent that they outcompeted national banks—increased overall welfare. The Van Order thesis is by all accounts the best and most sophisticated justification for the continued existence of the GSEs and provoked two contrary analyses at the second conference. Richard Scott Carnell argued that the GSEs and national banks cannot really be considered similar or competitive entities because of so many significant differences between the two. Charles Calomiris argued that Van Order’s welfare analysis failed to consider that the GSEs would use their subsidy elsewhere with no welfare benefit to society if GSE competition displaced banks from the mortgage market. Calomiris also presented an economist’s discussion of the other benefits frequently attributed to the GSEs and proposed a better way to allocate the agencies’ subsidy. The second conference also considered the question of whether and how Congress could effectively control Fannie and Freddie. Ralph Nader outlined the difficulties of doing so; he sees the GSEs PETER J. WALLISON 3 as an egregious case of corporate welfare. His essay, however, goes beyond the critique of Fannie and Freddie as feeders at the trough to delineate the many ways in which they use their wealth and power to perpetuate their advantages. The third conference focused on choices available to policymakers who are willing to consider modifying the status of Fannie Mae and Freddie Mac. Several possibilities presented themselves. The most popular idea has traditionally been full privatization, in which all the GSEs’ ties to the federal government would be cut. As a precedent for this, the Student Loan Marketing Association, a government-sponsored enterprise known as Sallie Mae, voluntarily privatized by cutting its government connections in the mid-1990s. Mark Overend of Sallie Mae discussed why and how the company took this step. Other proposals were discussed at the conference. Thomas Stanton proposed chartering additional GSEs as competitors for Fannie and Freddie. John Weicher of the Hudson Institute discussed the feasibility of restraining Fannie and Freddie through restrictions in their charters or through laws or regulations and concluded that such a course would likely be unsuccessful. Ron Feldman of the Minneapolis Federal Reserve Bank reached a similar conclusion; he explored the possibility that required credit ratings, imposition of a guarantee fee, or inclusion of the agencies in the federal budget might control Fannie and Freddie or reduce their risks. Finally, Jonathan Brown of Essential Information considered whether requiring the GSEs to meet stricter affordable and low- income housing guidelines might recapture all or a portion of the agencies’ subsidy. Brown illustrated that the GSEs do considerably less for low-income homebuyers than ordinary banks and savings- and-loan institutions despite congressional efforts to get them to share a greater proportion of their government benefits. That conclusion raises the interesting question of why so many key congressional supporters of the GSEs come from minority communities that receive the fewest benefits from Fannie and Freddie. More generally, however, the failure of Fannie and Freddie to respond to prodding by Congress and the Department of Housing and Urban Development illustrates why shareholder-owned corporations should not be charged with public missions. Fannie Mae and Freddie Mac, as privately owned companies, must seek every avenue for profit and every opportunity to escape those government mandates that will reduce their profitability. This objective is reinforced 4 INTRODUCTION by the compensation that their managements receive—through stock options and other benefits—from increases in their stock prices. Moreover, as Nader makes clear, Fannie Mae has used its wealth and power in Washington to escape virtually all forms of control. And it has done this through means that could—if fully exposed to public view—be seen as verging on the corrupt. Because of the agencies’ dual public and private form, various efforts to force Fannie Mae and Freddie Mac to fulfill their public mission at the cost of their profitability have failed—and will likely continue to fail. The only viable solution would seem to be full privatization or the adoption of policies that would force the agencies to adopt this course themselves. PART ONE Basics

In 1996 the Congressional Budget Office and the Treasury Department— responding to a request from Congress—produced groundbreaking studies of the implicit subsidies that Fannie Mae and Freddie Mac derived from their unique links to the federal government. Both analyses found that during 1995 Fannie and Freddie received aggregate subsidies of approximately $6.5 billion and—most significant—that they passed through only a portion of these subsidies to the mortgage markets. The balance was retained for the benefit of their shareholders and management. A somewhat more sophisticated study done by CBO in 2001 showed the subsidy even higher and a larger percentage of the higher subsidy going to shareholders and management. The GSEs contend that these studies are flawed, but they have produced no detailed analyses comparable to the CBO work. In the following essay Robert S. Seiler Jr.—an economist at CBO when he wrote this chapter—outlines the methodology used in CBO’s original analysis and further notes that the original methodology did not include certain additional factors that would increase the size of the 1995 subsidy estimate. Since the new CBO study will likely address those additional elements, the Seiler essay offers an insight into what to expect from the future CBO report.

7 2 Estimating the Value and Allocation of Federal Subsidies

Robert S. Seiler Jr.

Fannie Mae and Freddie Mac are government-sponsored enterprises (GSEs or enterprises) that dominate the nation’s system for financing home mortgages that carry no government insurance or guarantee, the so-called conventional loans. Government sponsorship provides each GSE with economic benefits that reduce its funding and operating costs by billions of dollars a year and give it a significant competitive advantage in financial markets. Those economic benefits result from explicit provisions of law and the perception of investors that the federal government has provided an implicit guarantee of the enterprise’s obligations. Federal law explicitly exempts Fannie Mae and Freddie Mac from state and local corporate income taxes and from the requirement to register their securities with the Securities and Exchange Commission (SEC). Each GSE has a conditional line of credit ($2.25 billion) with the Treasury Department, can use the Federal Reserve as a transfer agent, can borrow in the federal agency debt market, and is exempt from restrictive state laws on doing business. The securities of each enterprise are eligible for Federal Reserve open- market purchases, for collateralization of all federal and most state and local deposits, and for unlimited investment by most depository institutions. Fannie Mae and Freddie Mac also benefit from lower capital requirements than those for potential competitors.

The views in this essay are those of the author and should not be interpreted as those of the Congressional Budget Office. The author is grateful to Austin Kelly, Charles Capone, Marvin Phaup, Mario Ugoletti, and Robert Van Order for helpful comments.

8 ROBERT S. SEILER JR. 9

Investors infer an implicit federal guarantee of Fannie Mae and Freddie Mac obligations from their explicit legal benefits, congressional support for their public purposes, and the volume of their outstanding obligations. This perception reduces the interest rates that investors require on the GSEs’ debt and mortgage-backed securities (MBSs), and thus the enterprises are not subject to as much market discipline of their risk-taking as would be imposed in the absence of federal credit support. The perception of an implicit guarantee also allows Fannie Mae and Freddie Mac to issue much larger volumes of securities, sell a much larger proportion of debt that is callable, and enter into much larger volumes of hedging transactions than private firms with comparable levels of capital. Several studies by academic economists and federal agencies have analyzed the value of the economic benefits that government sponsorship provides to Fannie Mae and Freddie Mac. In 1996 four federal agencies—the Congressional Budget Office, the General Ac- counting Office, and the Departments of the Treasury and of Hous- ing and Urban Development—prepared the most recent estimates of benefits (CBO 1996; HUD 1996; Treasury 1996; GAO 1996).1 In response to a statutory request, each of those agencies submitted a report to Congress on the desirability and feasibility of repealing the federal charters of the two GSEs, eliminating their government sponsorship, and allowing them to operate as fully private corporations. In the context of the privatization issue, the reports estimated the value of the economic benefits received by Fannie Mae and Freddie Mac and assessed how the GSEs’ activities benefit the residential mortgage market. The CBO and Treasury reports differ from those of GAO and HUD in estimating the portion of their federal benefits that Fannie Mae and Freddie Mac retain for their shareholders. The CBO and the Treasury reports compare the value of the GSEs’ federal benefits with the value of the reduction in the interest rates on the mortgages that they finance. The two reports also argue that all economic benefits received by Fannie Mae and Freddie Mac, not merely those con- veyed by the implicit federal guarantee of their securities, are costly federal subsidies. Fannie Mae and Freddie Mac strongly criticize the 1996 privatization studies on two general grounds. First, the GSEs contend that the CBO and the Treasury define the federal subsidies received by the enterprises much too broadly. Fannie Mae and Freddie Mac argue that the government incurs tiny, indirect costs in provid- 10 VALUE OF FEDERAL SUBSIDIES ing the large economic benefits that they receive, and that only those costs provide subsidies. Second, the GSEs argue that the four reports overestimated the economic benefits that they receive and underestimated the economic benefits that their activities provide to the mortgage market. When a more appropriate definition of their federal subsidies and more accurate estimating methods are used, the enterprises contend, the inescapable conclusion is that the benefits provided to the mortgage market vastly exceed the subsidies that government sponsorship conveys. This essay evaluates the CBO and the Treasury estimates of the value and allocation of the federal subsidies provided to Fannie Mae and Freddie.2 The next section examines the definitions of the GSE federal subsidies adopted by the CBO and Treasury and advocated by the enterprises. Next comes an overview of the CBO and Trea- sury method for estimating the value and allocation of the GSE federal subsidies. Then I review and evaluate the two agencies’ estimates of their cost advantage in securitizing mortgages and issuing debt and review estimates of the differential between the interest rates on conventional, single-family, fixed-rate mortgages (FRMs) that are too large to be eligible for purchase by Fannie Mae and Freddie Mac and those that are small enough to be eligible for purchase by the GSEs. The former are often called jumbo mortgages, and the latter, conforming loans.3 Finally, I examine how the behavior of the enterprises influences the size and allocation of their federal subsidies and offer some conclusions.

Defining the Federal Subsidies

The CBO and Treasury privatization reports argue that the federal government provides subsidies to Fannie Mae and Freddie Mac equal to the value of the economic benefits of government sponsorship. The GSEs define their federal subsidies much more narrowly.4 According to the CBO report, whenever the government conveys economic benefits to private parties without receiving equivalent compensation in return, it incurs an opportunity cost equal to the market value of the benefits. Government sponsorship provides subsidies to Fannie Mae and Freddie Mac equal to the opportunity cost of giving away their charters free of charge. Although the Trea- sury report does not discuss the issue at length, it takes the same position (p. 29): “The benefits that Fannie Mae and Freddie Mac ROBERT S. SEILER JR. 11 receive from government sponsorship have real economic value, and thus provide an in-kind subsidy.” Fannie Mae and Freddie Mac argue that they impose no direct cost on the government because they receive no checks from the Treasury. They claim that the only indirect cost associated with government sponsorship is the potential future cost should Congress choose to provide financial assistance to them if they ever became insolvent.5 To buttress their view, the GSEs cite Bosworth, Carron, and Rhyne (1987), who argue that the budgetary cost of federal direct loans and loan guarantees should be estimated on the basis of the government’s expected cash outlays and receipts, rather than on the basis of the opportunity cost of charging recipients less than the market value of the assistance they enjoy. (The Federal Credit Re- form Act of 1990, which reformed the budgetary treatment of loans and loan guarantees, requires agencies to estimate the cost of such assistance on the basis of expected cash flows.) The enterprises also claim that the likelihood of either of them becoming insolvent and receiving federal assistance in the foreseeable future is extremely remote. In support of this assessment, they cite the president’s annual budget submissions to Congress in recent years, which have estimated that the implicit federal guarantee of their obligations will have no future budgetary costs, as well as the activities of the regulator of their safety and soundness, the Office of Federal Housing Enterprise Oversight (OFHEO).6 The view that government sponsorship of the GSEs is essentially costless to taxpayers is consistent with the fact that investors accept lower interest rates on Fannie Mae and Freddie Mac debt and MBSs than on comparable securities issued by non-GSEs. Most of the enterprises’ funding advantage results from the high liquidity of their obligations rather than from government risk-bearing. Woodward (1996) observes that the implicit federal guarantee allows investors to avoid nearly all normal costs associated with col- lecting or analyzing information about the default risk of a firm.7 Further, as the volume of Fannie Mae and Freddie Mac obligations steadily increases, the number of investors who are informed about the securities also rises. This steady growth in the size of the market continually reduces the cost to new investors of becoming informed.8 These reductions in the transaction costs associated with investing in GSE debt and MBSs make the obligations cheaper and easier to buy and sell; thus, investors will accept lower interest rates on the securities than on comparable securities issued by non- 12 VALUE OF FEDERAL SUBSIDIES

GSEs. Also contributing to the greater liquidity of enterprise obligations are the legal exemptions that broaden the universe of investors that buy the securities and increase the amounts that investors are willing to hold. Independent economists who have reviewed the literature on federal subsidies to Fannie Mae and Freddie Mac (Feldman 1999 and Kane 1999) concur with the broad definition of those subsidies favored by the CBO and Treasury. Aida Alvarez, the first director of OFHEO, also supported that view in testimony before Congress on the four privatization reports (Alvarez 1996). Downs (1980) supports the narrow view. This essay adopts the broad definition of the cost of government sponsorship of Fannie Mae and Freddie Mac on three grounds. First, the narrow definition assumes that the resources that government sponsorship conveys to the GSEs have no alternative uses. In fact, those resources could otherwise be sold to private investors, and the proceeds could finance other government programs. Char- acterized in this way, the opportunity cost of giving away the benefits free of charge is as real as the cost that the government would incur if it gave away permits to harvest timber in national forests, licenses to broadcast on the radio spectrum, or the authority to produce hydroelectric power on federally owned waterways. Second, even if most of the reduction in the interest rates on Fannie Mae and Freddie Mac obligations results from lower transaction costs that represent efficiency gains, those gains are not a public good because the government allows only two firms to enjoy them. The narrow definition would be appropriate only if any firm could obtain, free of charge, the same statutory authority and exemptions as the two GSEs possess or if the government auctioned off the charters and the enterprises’ related benefits to the highest bidders. Since neither of those conditions obtains, the broad definition is appropriate. Third, the broad definition focuses on how the GSEs allocate the scarce resources that they receive. The enterprises’ protected position gives their owners and managers greater control over the allocation of the economic benefits of government sponsorship than if identical charters were freely available to all their potential competitors. The issue of how Fannie Mae and Freddie Mac allocate the resources that they receive is worthy of policy analysis and public debate. Adopting the narrow definition of the GSEs’ federal subsidies obscures that issue. ROBERT S. SEILER JR. 13

In the decade before the four privatization reports, academic researchers published several studies of the cost to the government of the implicit guarantee of the obligations of Fannie Mae and Freddie Mac. Feldman (1999) and Kane (1999) provide surveys of this literature. Those studies give useful insights into how the GSEs’ behavior and financial market conditions affect the cost of the implicit federal guarantee. As noted, however, that guarantee accounts for only a small portion of the economic benefits of federal sponsorship. For that reason, the studies’ methods are not well suited to estimating the value of the GSEs’ subsidies as broadly defined by the CBO, Treasury, and this essay.

The CBO-Treasury Estimates

The CBO and the Treasury privatization reports each provide estimates of the aggregate federal subsidies that Fannie Mae and Freddie Mac received in 1995, the portion of those subsidies that the GSEs passed through to the mortgage market in that year, and the portion that the enterprises retained. Each report estimates sepa- rately the subsidies that Fannie Mae and Freddie Mac received from their explicit legal exemptions and their cost advantage in issuing debt and MBSs. The CBO estimates that the exemption from state and local income taxation cost federal taxpayers $263 million in 1995. That estimate reflects the net effect of the exemption on the total GSE tax liability; the CBO estimated the gross effect on their state and local tax liability at $394 million. Both agencies cited a previous GAO estimate that the exemption from Securities and Exchange Commission registration is worth more than $100 million annually to the GSEs. To measure the value of the funding-cost subsidies that Fannie Mae and Freddie Mac receive, the CBO and the Treasury make two assumptions for the purpose of simplification:

• That the risk posed by Fannie Mae and Freddie Mac to the federal government can be expressed in terms of a credit rating.9 TheCBO assumes that the GSEs would each be rated double-A. Treasury assumes that each would receive a single-A rating.10 • That, without government sponsorship, the enterprises would have the same capital structure and could issue debt in the same volume, issue and provide corporate guarantees of MBSs in the same volume, issue the same proportion of callable debt, and 14 VALUE OF FEDERAL SUBSIDIES

enter into comparable amounts of interest rate swaps and other hedging transactions.

On the basis of these assumptions, the CBO and the Treasury use market data to compare the interest rates of debt and MBSs issued by Fannie Mae and Freddie Mac with the rates of similar obligations issued by private financial institutions with the same credit ratings. The CBO and the Treasury estimate each enterprise’s funding subsidies in 1995 as in equation 2–1:

funding subsidies = (average cost savings in issuing (2–1) debt x average debt outstanding) + (average cost savings in issuing MBSs x average MBSs outstanding)

The CBO and the Treasury also estimate the proportion of the federal subsidies that Fannie Mae and Freddie Mac passed through to the mortgage market in 1995. To do so, the agencies assume that the GSEs’ activities reduce the interest rates on all mortgages that they purchase by an amount equal to the estimated differential between the interest rates on jumbo and conforming FRMs. As discussed later, the CBO and the Treasury rely on previous studies to estimate that differential. Based on these assumptions, the agencies estimate the federal subsidies that the enterprises pass through to the mortgage market as in equation 2–2:

subsidies passed through = (average reduction (2– 2) in mortgage rates x average mortgages financed)

The CBO and the Treasury estimate the federal subsidies that Fannie Mae and Freddie Mac retain by subtracting equation 2–2 from equation 2–1, which yields equation 2–3:

subsidies retained = (average cost savings in issuing (2–3) debt x average debt outstanding) + (average cost savings in issuing MBSs x average MBSs outstanding) – (average reduction in mortgage rates x average mortgages financed)

Using this simple approach, the CBO estimates that Fannie Mae and Freddie Mac received $6.5 billion in funding subsidies in 1995 (table 2–1). Adding the value of the GSEs’ explicit legal exemptions, the budget office estimates the total subsidies for that year at $6.9 ROBERT S. SEILER JR. 15

TABLE 2–1 CBO’s Estimates of Federal Subsidies to Fannie Mae and Freddie Mac, 1995 (billions of dollars) Fannie Mae Freddie Mac Total Gross subsidy Average debt outstanding 278.3 105.7 384.0 Subsidy (70 or 68 basis points)a 1.9 0.7 2.6 Average MBS outstanding 494.7 450.5 945.2 Subsidy (40 basis points) 2.0 1.8 3.8 Total funding subsidy 3.9 2.6 6.5 Subsidy pass-through Mortgages financed 719.1 529.9 1,249.0 Pass-through (35 basis points) 2.5 1.9 4.4 Funding subsidy retained (total subsidy minus pass-through) 1.4 0.7 2.1 Net income before taxes and gift 3.4 1.6 4.9 Retained subsidy (percentage of net income before taxes and gift) 41.1 44.9 42.3 (average) MBS = mortgage-backed securities. a. CBO estimates average savings for 1995 of 70 basis points for Fannie Mae and 68 basis points for Freddie Mac. SOURCE: CBO 1996, table 8, 19. billion. The CBO also estimates that the enterprises passed through $4.4 billion of their funding subsidies to the mortgage market and retained $2.1 billion of those subsidies for their shareholders and other purposes. The CBO notes that the estimated retained funding subsidies amount to about 42 percent of the combined pretax net income of Fannie Mae and Freddie Mac in 1995.11 The office provides the results of a sensitivity analysis of its estimate of the GSEs’ retained subsidies as a percentage of their pretax net income. Under alternative assumptions about the size of the enterprises’ funding- cost advantage and the reduction in mortgage rates resulting from their activities, the retained subsidies range from 4 percent to 48 percent of pretax income in 1995. Using the same method but somewhat different assumptions, the Treasury estimates the total subsidies received by Fannie Mae 16 VALUE OF FEDERAL SUBSIDIES

TABLE 2–2 Treasury’s Estimates of Federal Subsidies to Fannie Mae and Freddie Mac, 1995 (billions of dollars)

Average Low Estimate High Estimate Obligations Outstanding Basis Basis ($) points $ points $ Gross Subsidies Debt Long-term 327 36 1.2 74 2.4 Short-term 57 3 a 33 0.2 MBS 960 30 2.9 40 3.8 Funding subsidies 4.1 6.4 Operating subsidies 0.5 0.5 Total gross subsidies 4.6 6.9

Subsidies Passed through Mortgages financed 1,287 20 2.6 40 5.1 MBS = mortgage-backed securities. a. Less than $500 million. SOURCE: Treasury 1996, 29–35.

and Freddie Mac in 1995 at $4.6–6.9 billion (table 2–2). Treasury estimates the amount of the subsidies that the enterprises’ passed through to the mortgage market at $2.6–5.1 billion. Treasury’s figures imply that Fannie Mae and Freddie Mac retained $1.8–2.0 billion of their total federal subsidies in 1995. The CBO sensitivity analysis of its estimates and the range of estimates provided by the Treasury suggest that the numbers do not have much value as point estimates. As Feldman (1999) and Kane (1999) suggest, they are best viewed as reasonable approximations of the public resources devoted to Fannie Mae and Freddie Mac and the allocation of those resources. The remainder of this section reviews two general criticisms of their approach. Following sections will review and evaluate the major empirical assumptions underlying the CBO and the Treasury estimates. Fannie Mae and Freddie Mac criticize the CBO-Treasury method because it does not include the value of the reduction in the rates on ROBERT S. SEILER JR. 17 single-family mortgages not financed by the GSEs. The enterprises argue that when all reductions in interest rates on conforming FRMs resulting from their activities are taken into account, the benefits that they provide the mortgage market clearly exceed their federal subsidies. There are two telling objections to this view. First, estimating how Fannie Mae and Freddie Mac allocate their federal subsidies should consider only the resources actually under GSE control. Second, any reduction in the interest rates on mortgages under the conforming-loan limit but not financed by the enterprises is offset by an equal reduction in the interest income earned by mortgage lenders that hold such loans—with a result of no net benefit for the housing finance system as a whole. Another criticism of the CBO-Treasury method is that the sim- plifying assumptions used in estimating the funding-cost advantage of Fannie Mae and Freddie Mac are unrealistic. As Feldman (1999) emphasizes, the GSE credit quality, market access, and liquidity are without parallel. By essentially assuming away the unique market position afforded by government sponsorship, the CBO-Treasury method understates the GSE funding-cost advantage and, thus, their federal subsidies.

The Funding-Cost Advantage

Any effort to measure the funding-cost advantage of Fannie Mae and Freddie Mac must contend with the limited number and dollar volume of securities issued by fully private firms that are comparable with the GSEs’ obligations. In estimating the enterprises’ debt- cost advantage, the CBO and the Treasury address this problem in different ways. The budget office relies on statistical estimates of the mean spreads between the yields on debt issued by the GSEs and financial firms of comparable risk. The Treasury relies on market data on the debt of federal agencies and fully private firms reported by Bloomberg Financial Services. To estimate the funding advantage of Fannie Mae and Freddie Mac in issuing MBSs, the CBO and the Treasury rely on previous studies and discussions with market participants.

Advantage in Issuing Debt. The CBO report relies on a study by Ambrose and Warga (1996), who estimate mean differences between the yields on bonds issued by Fannie Mae and Freddie Mac and on those issued by fully private firms.12 Ambrose and Warga use trader- 18 VALUE OF FEDERAL SUBSIDIES quoted data from 1985–1994 to create a series of matched portfolios of GSE bonds and similar bonds issued by fully private firms. The matching process controls for each bond’s credit rating, age, maturity, and nominal interest rate; separates bonds into those issued by financial firms and other corporations; and separates callable from noncallable bonds. The matching controls for factors that affect bond yields while constructing a large enough number of comparison portfolios to facilitate statistical analysis. Ambrose and Warga exclude any bond with less than twelve months remaining to maturity from the analysis. Further, they report average spreads between the yields of bonds issued by financial firms and by Fannie Mae and Freddie Mac for both the 1985–1994 period and for the 1985–1990 and 1991–1994 subperiods. As noted, the CBO report assumed that the risk that each GSE poses to the government is comparable with the risk borne by investors in the debt of a fully private firm with a double-A rating. For the 1985–1994 period, Ambrose and Warga report average spreads between the yields on double-A rated financial firms and the enterprises of 37 basis points for noncallable bonds and 63 basis points for callable bonds. For the 1991–1994 period, for which more data on private bonds are available, Ambrose and Warga report average spreads of 46 basis points for noncallable bonds and 105 basis points for callable bonds. The CBO report assumes the latter spreads in estimating the debt funding cost subsidies of Fannie Mae and Freddie Mac in 1995. For each GSE, the CBO weights the spreads by the amount of outstanding callable and noncallable debt held by each enterprise at the end of that year. The two GSEs, consultants hired by Freddie Mac, and academic economists criticize the methodology of the Ambrose and Warga study on several grounds (Cook 1996; Fannie Mae 1996; Freddie Mac 1996; Miller, Tollison, and Higgins 1996; Shilling 1996). First, the comparison portfolios of bonds do not adequately control for important bond-pricing factors—for example, call dates—that affect yield. This problem influences the study’s estimates of mean yield spreads, perhaps significantly. Second, a large variance is associated with the estimates of the mean yield spreads, especially the spreads for callable debt. Thus, one cannot be confident that the estimated spreads are close to their true values. If the markets for the GSEs’ callable and noncallable debt are efficient, the spreads for the two types of debt should be comparable. Third, Ambrose and Warga provide no com- ROBERT S. SEILER JR. 19 pelling reason for partitioning the 1985–1994 sample period as they do, with the resulting higher mean yield spreads in the 1991–1994 subperiod than in the 1985–1990 subperiod. Fourth, the inclusion of additional data from years since 1994 would lower the estimated spreads. Fifth, Ambrose and Warga do not address the serial corre- lation in the yield-spread time series or test for the stationarity of the series. Citing these criticisms, Fannie Mae and Freddie Mac argue that the mean yield spreads estimated by Ambrose and Warga are biased and cannot be used to estimate their debt-funding advantage. The Treasury report uses market data reported by Bloomberg Financial Services to estimate the debt-funding–cost advantage of Fannie Mae and Freddie Mac. Bloomberg reports the yields of bonds of various maturities issued by federal agencies and fully private firms. The yields are calculated using the Bloomberg Fair Value Model, which controls for imbedded options and other characteristics that affect yield and therefore compares bonds with different structures on an equal basis. The Treasury used this data source to compare the borrowing costs of the GSEs with those of banks and financial firms with an A credit rating. From January 1992 to April 30, 1996, the Bloomberg data showed average spreads between the yields on five-year agency bonds and A-rated bank bonds of 55 basis points. The Treasury report uses this yield spread to estimate the enterprises’ funding-cost advantage in issuing medium- and long- term debt in 1995. The Treasury chooses the five-year maturity because it is close to the average life of a thirty-year mortgage. To estimate the funding-cost advantage of Fannie Mae and Freddie Mac in issuing short-term debt (notes and bonds maturing in less than one year), Treasury uses Bloomberg data on the spreads between the yields on the GSEs’ discount notes and the three-month London interbank offered rate (LIBOR), which is a common measure of the short-term funding costs of highly rated banks. This spread averaged 18 basis points over the period for which Bloomberg data were available. Both approaches to estimating the debt-funding advantages of Fannie Mae and Freddie Mac have strengths and weaknesses. On the plus side of the Ambrose and Warga estimates adopted by the CBO, researchers use data on bonds issued by the GSEs and explicitly attempt to control for bond characteristics that affect yield. The few observations of finance industry bonds, especially callable bonds, limit 20 VALUE OF FEDERAL SUBSIDIES the number of comparison portfolios that can be constructed, so that the usability of the estimated mean yield spreads is not clear. Such small-sample problems are inherent in any effort to compare the debt- funding cost of the GSEs and private firms, however. The CBO’s report acknowledges the limitations of the data and, to minimize their effect, uses the Ambrose and Warga estimates for the shortest, most recent period, in which at least nine matched pairs of bonds are available for comparison in every month of the estimation period. The yield spreads estimated by the Bloomberg Fair Value Model and used by the Treasury probably reflect more accurately the effect of call options and other structural features that influence yield. The Treasury was also more accurate in using data on short-term debt yields to estimate the funding-cost advantage of Fannie Mae and Freddie Mac in issuing short-term debt. However, the Bloomberg data on federal agency bonds include debt issued by GSEs other than Fannie Mae and Freddie Mac as well as non-GSEs such as the Tennessee Valley Authority; the securities of those issuers may be less liquid and have higher yields than those of Fannie Mae and Freddie Mac. Further, it is not clear that the mix of callable and noncallable debt in the sample of bonds used by Bloomberg reflects the mix of outstanding GSE debt at any particular time. Fannie Mae and Freddie Mac and the latter’s consultants (Miller, Tollison, and Higgins 1996) suggest that the limited data on callable bonds issued by non-GSEs make it too difficult to estimate reliably the enterprises’ funding-cost advantage in issuing callable debt. However, ignoring such estimates altogether may introduce a bias into estimates of the GSEs’ overall debt-funding advantage.

Advantage in Securitizing Mortgages. Private conduits generally finance jumbo mortgages by issuing multiple classes of senior and subordinated securities structured as real estate mortgage investment conduits (REMICs), whereas the MBSs issued by Fannie Mae and Freddie Mac are single-class securities. The different financing structures prevent an easy comparison of data series on the yields on the GSEs’ MBSs and private-label securities. In the absence of such data, the CBO and the Treasury reports rely on previous estimates of the enterprises’ cost advantage in issuing MBSs made by Goodman and Passmore (1992), Ambrose and Warga (1996), and other researchers and on discussions with market participants. The two agencies estimate the GSEs’ funding advantage in issuing MBSs at 40 and 35 basis points, respectively. ROBERT S. SEILER JR. 21

The Differential between the Interest Rates on Jumbo and Conforming Fixed-Rate Mortgages

Fannie Mae and Freddie Mac rapidly expanded their financing of conforming, single-family, fixed-rate mortgages in the 1980s in response to the decline of the thrift industry (Hendershott 1992). The proportion of newly originated, conforming FRMs financed by the GSEs rose from about 4 percent before 1981 to more than 50 percent in 1986 through 1988, 69 percent in 1989, and 83 percent in 1993; the number has fluctuated between 47 percent and 59 percent in subsequent years.13 There is evidence that this expansion of the enterprises’ purchases has reduced the interest rates on conforming FRMs. Sev- eral studies have estimated the differential between the interest rates on jumbo and conforming FRMs. The CBO and the Treasury privatization reports make assumptions about this differential to estimate the portion of their federal subsidies that Fannie Mae and Freddie Mac pass through to the mortgage market.

Estimates of the Differential. Past research on the differential between the interest rates on jumbo and conforming FRMs uses data on closed loans from the Mortgage Interest Rate Survey (MIRS) conducted by the Federal Housing Finance Board (before 1989, the Fed- eral Home Loan Bank Board did the survey). To estimate that differential, the studies use econometric models that control for the impact of some variables that affect mortgage default risk, including the loan-to-value (LTV) ratio (the ratio between the loan amount and the home value), loan size, loan maturity, and whether loans finance new or existing homes.14 Two studies find a mean national spread between the interest rates on closed conforming and jumbo FRMs of 30 basis points in 1986–1987 (Hendershott and Shilling 1989) and 10–23 basis points in 1987 (ICF 1990). The most recent study finds that the spread ranged from 60 basis points in 1989 to 25 basis points in 1993 (Cotterman and Pearce 1996). Relying on the findings of the Cotterman and Pearce study, the CBO report assumes that Fannie Mae and Freddie Mac reduce the interest rate on conforming FRMs by 35 basis points on average. The Treasury takes 20–40 basis points as a reasonable range of estimates for the reduction in FRM rates. In the absence of data on the effect of GSE activities on the interest rates on adjustable-rate mortgages (ARMs) or multifamily mortgages, the two agencies assume, conservatively, that all such loans benefit from the same reduction in rates as FRMs. 22 VALUE OF FEDERAL SUBSIDIES

FIGURE 2–1 Spreads between Yields on Thirty-Year Fixed-Rate Jumbo and Conforming Mortgages, Fourth Quarter 1986 to First Quarter 1999

80 Mean of 45 basis points, 70 1986–1991

60 Mean 50 plus one Spread standard in points deviation 40

30 Spread in points 20 Mean minus one Mean of 25 10 standard basis points, deviation 1992–1999 0 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999

Source: Based on data from HSH Associates.

There is evidence that the jumbo–conforming rate differential has been lower in recent years than in the 1986–1993 period exam- ined by previous research. Figure 2–1 shows quarterly averages of the estimated spreads between the effective interest rates on jumbo and conforming FRMs from the first quarter of 1987 through the first quarter of 1999. The spreads were estimated from data collected by HSH Associates, a private publisher of financial information that weekly surveys lenders regarding the interest rate and points at which they offer to originate conforming and jumbo FRMs.15 The MIRS data on the terms of closed loans, which Cotterman and Pearce use, reflect market conditions about two months later than the HSH data on the terms offered by lenders. The method used to estimate the spreads in figure 2–1 adjusts for this difference in timing to make the estimates comparable to those of Cotterman and Pearce.16 For the 1989–1993 period, the mean and the standard deviations of the quar- ROBERT S. SEILER JR. 23 terly estimates of the jumbo-conforming spread estimated by Cotterman and Pearce and by the author using the HSH survey data are similar.17 Thus, the data displayed in figure 2–1 seem reasonably representative of conditions in the primary-mortgage market. Figure 2–1 suggests that the jumbo-conforming differential has averaged 26 basis points since 1991 after averaging 45 basis points in 1986–1991. As noted, Cotterman and Pearce estimated a mean spread of 60 basis points in 1989 and 25 basis points in 1993. The chart suggests that if the Cotterman and Pearce study were repli- cated with MIRS data for 1994 and subsequent years, the estimated jumbo-conforming spread would likely be significantly lower than for the 1989–1993 period. The chart also shows how the jumbo- conforming differential fluctuated in the 1994–1998 period.

The Effect of Geographic Diversification. Some of the funding-cost advantage of Fannie Mae and Freddie Mac in issuing MBSs, as well as some of the jumbo-conforming rate differential, may be due to the lower level of geographic diversity for the average pool of jumbo FRMs backing private-label MBSs than for the conforming FRMs securitized by the GSEs. As GAO (1996) observes, jumbo FRMs are generally concentrated in California and the Northeast. According to GAO, industry participants indicate that the rating agencies require additional overcollateralization of private MBSs backed by jumbo FRMs to compensate for the lower geographic diversification. Also, the studies of the differential between the interest rates on jumbo and conforming FRMs (Cotterman and Pearce 1996; Hendershott and Shilling 1989; ICF 1990) found the differential to be 5–10 basis points higher in California in 1986–1991. If research found that part of the differentials in the funding costs of the GSEs and private-label conduits in issuing MBSs and in the rates on jumbo and conforming FRMs are due to systematic differences in geographic diversification, then it would be appropriate to reduce estimates of the size of the enterprises’ federal subsidies and the amount that they pass through to the mortgage market.

Loan Size and the Default Risk of Single-Family Mortgages. The Treasury privatization report observes that the relationship between the size of a single-family mortgage and the interest rate on the loan is complex, both theoretically and empirically. Many factors affect the relationship between loan size and the mortgage rate.18 On one hand, origination and servicing costs are relatively fixed per dollar of loan balance, and larger loans provide more servicing income. 24 VALUE OF FEDERAL SUBSIDIES

These factors should tend to put downward pressure on mortgage rates as loan size increases. On the other hand, the value of the borrower’s option to prepay a mortgage increases with loan size because the financial benefits of refinancing a large mortgage exceed those of refinancing a small one. Also, the volatility of home prices tends to be greater in the market for high-priced properties in each geographic region because there tend to be fewer buyers in that market than in the market for low-priced homes. Greater price volatility is likely to mean greater default risk to lenders. By increasing the risks borne by lenders, these factors should tend to put up- ward pressure on mortgage rates. It is not clear whether Cotterman and Pearce (1996) properly specify the loan-size variable in their econometric models. Their models assume that as the size of a mortgage increases, the interest rate on the loan falls at a decreasing rate. This assumption may be unsatisfactory if the positive effects of loan size outweigh the negative effects beyond some loan level. For example, in the 1986 data reported by Hendershott and Shilling (1989), mortgage rates clearly fall as loan size increases below the conforming limit, whereas above the limit the relationship is either nonexistent or slightly positive. Data analyzed by Cotterman and Pearce for 1992 exhibit similar discontinuities in the relationship between loan size and rate. These discontinuities urge caution in interpreting the results of the three studies as unbiased estimates of the reduction in interest rates on conforming FRMs produced by Fannie Mae and Freddie Mac. Freddie Mac (1996) advocates adding 7 basis points to the Cotterman-Pearce estimates of the jumbo-conforming differential to reflect differences in origination costs per dollar of loan balance and other loan-size effects. If one assumes with the CBO and the Trea- sury that the GSEs reduce the interest rates of all mortgages that they finance by that differential, then making such an adjustment would significantly increase the estimated value of the enterprises’ federal subsidies that they pass through to the mortgage market. Given the uncertainty about the effect of loan size on mortgage rates, such adjustments would be premature.

How the GSEs Affect the Size and Allocation of Their Federal Subsidies

The CBO and the Treasury privatization reports each conclude that Fannie Mae and Freddie Mac retain about one-third of their federal subsidies. This conclusion should not be surprising. The GSEs have ROBERT S. SEILER JR. 25 shareholders and want to earn profits, and the retained portion of their federal subsidies enhances their profits. Moreover, they have broad discretion under the law to determine the volume of mortgages that they finance, the fees that they charge to guarantee MBSs, the prices that they pay for loans and other assets, and how they finance their portfolios. This discretion gives Fannie Mae and Freddie Mac substantial influence over the size and allocation of their federal subsidies. The CBO and the Treasury accept the conclusion of several previous studies that Fannie Mae and Freddie Mac are “tacitly colluding” in the market for financing conforming FRMs. Both agencies also assert that the most important way in which the GSEs retain federal subsidies is by financing mortgages and other assets with debt. The CBO report notes that the enterprises have expanded the retained proportion of their federal subsidies primarily by increasing the amount of mortgages and mortgage securities held in their retained portfolios.

The GSEs’ Guarantee Fee Pricing. The CBO and the Treasury privatization reports conclude that Fannie Mae and Freddie Mac each retain 5 basis points of the spread between the average interest rate paid to investors in MBSs and the average interest rate earned on mortgages backing such securities. This is equivalent to assuming that the fees that the GSEs charge for guaranteeing their MBSs are 5 basis points higher on average than if the enterprises fully passed through their cost advantage in issuing such securities. This conclusion is consistent with studies by Goodman and Passmore (1992) and Hermalin and Jaffee (1996), which conclude that Fannie Mae and Freddie Mac appear to be tacitly agreeing or colluding in setting the guarantee fees that they charge in the market for securitizing conforming FRMs above competitive levels.19 Whenever a few firms dominate a market, noncompetitive pricing among them is always a possibility. Economists have developed various theories of behavior in markets characterized by a few firms (Shepherd 1990; Viscusi, Vernon, and Harrington 1996). In some models, the oligopolists are assumed to agree tacitly or collude to maintain prices and profits above competitive levels.20 (Tacit collusion does not mean that the firms cooperate directly, which would be a violation of the Sherman Antitrust Act, but that the firms behave as if they were colluding directly by avoiding vigorous competition. The federal charters of Fannie Mae and Freddie Mac do not explicitly exempt the GSEs from that law.) These models assume 26 VALUE OF FEDERAL SUBSIDIES that the oligopolists assess the profitability of maintaining some form of tacit collusion about pricing against the profitability of breaking that agreement (“cheating”). The models imply that tacit collusion is more likely to be maintained if the number of firms in the industry is small, the firms are similar in product offerings and technology, the firms have frequent interactions through the process of attracting customers, and the firms can easily observe the actions of their competitors. The models also imply that when tacit collusion is sustained, prices and profits will be higher (and output lower) than under competition but that prices will not be as high (or output as low) as under monopoly. Three features that may facilitate tacit collusion characterize the market for securitizing conforming FRMs: two dominant firms finance similar loans, they use similar technology, and they have frequent interactions (Goodman and Passmore 1992; Hermalin and Jaffee 1996). However, neither Fannie Mae nor Freddie Mac can directly observe the guarantee fees and other major terms of the master commitments that the other GSE has negotiated with lenders. Fannie Mae and some other analysts infer that neither GSE can effectively monitor cheating by the other and, therefore, the enterprises cannot tacitly collude in the mortgage credit guarantee business (Fannie Mae 1996; Woodward 1996). The issue hinges on whether lenders have an incentive to provide each GSE with truthful information about the other’s behavior in a timely fashion. Each enterprise’s relationship with its largest customers is crucial in meeting its volume and market-share goals. Such lenders may be able to withhold information consistently and successfully. However, smaller customers have an incentive to provide such information to stay on relatively good terms with each enterprise. Thus, Fannie Mae and Freddie Mac probably can sustain tacit collusion in their deal- ings with smaller lenders, especially when market conditions increase the GSEs’ competitive advantage. The studies by Goodman and Passmore (1992) and Hermalin and Jaffee (1996) cite three pieces of evidence as indirectly supporting the view that Fannie Mae and Freddie Mac are pricing their mortgage guarantee services above competitive levels. First, there is econometric evidence of a negative relationship between the volume of mortgages available for securitization and the proportion of mortgages securitized by the GSEs. Goodman and Passmore argue that this finding is consistent with the hypothesis that the two enterprises are tacitly collusive duopolists that price procyclically as the supply ROBERT S. SEILER JR. 27 of mortgages varies over time.21 Second, Fannie Mae and Freddie Mac have earned after-tax returns on equity (ROEs) in recent years that are significantly higher than the 17 percent after-tax ROE that researchers have concluded is the minimum for investors in each GSE’s stock to earn for the enterprise to attract capital (Gatti and Spahr 1995; Lekkas and Van Order 1991).22 Third, the ratios between the market value of each enterprise’s equity and either market-to- market or book-value measures of its net worth have consistently been much greater than those of large depository institutions (Ugoletti 1997). The second and third pieces of evidence are of limited value because the after-tax ROEs and the market values of Fannie Mae and Freddie Mac reflect profits from both their retained portfolios and their off–balance-sheet MBSs. Data in Fannie Mae’s annual reports and investor/analyst reports provide more relevant evidence. The reports break down the GSE’s 1994–1998 financial performance into the mortgage credit guarantee and portfolio investment businesses. The breakdown cor- rectly treats all mortgages that the enterprise holds in portfolio as if they had been securitized. That is, Fannie Mae’s mortgage guarantee business is credited with all income and expense associated with bearing mortgage-credit risk (as well as “float” income earned by investing funds received from servicers for a short period before disbursing them to investors in MBSs), while the portfolio investment business is credited with all income and expense associated with bearing interest-rate risk (other than “float” income) and nonmortgage-credit risk. By making assumptions about the average amount of capital that Fannie Mae implicitly allocated to each business in each year, one can calculate its implicit ROE for that year. Using the Fannie Mae data, the author has calculated that the GSE earned an average after-tax ROE on the mortgage guarantee business exceeding 28 percent in 1994–1998 (table 2–3). The enterprise reported an average overall after-tax ROE (excluding the special contribution to the Fannie Mae Foundation in 1995) of about 24 percent for the five years. The author has also estimated the performance of the enterprise’s mortgage-guarantee business in 1989–1993 by decomposing available data from Fannie Mae into line-of-business results similar to those reported by the GSE for 1994–1998. The estimates suggest that the enterprise’s average after-tax ROE on the business was about 32 percent in 1989 and 1990, dropped to 25 percent in 1991, and rose to 28 percent in 1992 and 30 percent by 1993. 28 VALUE OF FEDERAL SUBSIDIES

TABLE 2–3 Fannie Mae’s Financial Performance by Line of Business, 1994–1998 (billions of dollars) 1994 1995 1996 1997 1998

Mortgage Guaranty Business Net interest income 349 407 412 466 650 Guaranty fees 1,656 1,716 1,926 2,019 2,052 On net MBS outstanding 1,083 1,086 1,196 1,274 1,229 On mortgages and mortgage securities retained in portfolio 573 630 730 746 823 Miscellaneous income, net 132 71 54 37 117 Credit-related expenses (378) (335) (409) (375) (261) Administrative expenses (419) (414) (431) (462) (524) Federal income taxes (463) (442) (490) (524) (494) Net income 877 1,003 1,063 1,162 1,541 Estimate capital allocationa 3,273 3,533 3,981 4,375 4,583 Estimated after-tax ROE 26.8% 28.4% 26.7% 26.6% 33.6% Portfolio Investment Business Net interest income 2,474 2,640 3,180 3,483 3,460 Guaranty fees on mortgages and mortgage securities in portfolio (573) (630) (730) (746) (823) Other 1,901 2,010 2,450 2,738 2,637 Miscellaneous income, net 11 22 32 88 157 Administrative expenses (106) (132) (130) (174) (184) Federal income taxes (542) (520) (661) (745) (707) Extraordinary item 0 0 (29) (13) (26) Net income 1,255 1,369 1,662 1,894 1,878 Estimated capital allocationb 5,524 6,717 7,885 8,908 9,740 Estimated after-tax ROE 22.7% 20.4% 21.1% 21.3% 19.3% MBS = mortgage-backed securities. ROE = return on equity. a. Average net MBS outstanding plus mortgages and MBS in portfolio times 0.45 percent times the ratio between average actual capital and the average amount required by the statutory leverage ratios. b. Average actual capital less capital allocated to the mortgage guaranty business. SOURCE: Fannie Mae and Freddie Mac Investor/Analyst Reports, fourth quarters of 1994–1998, and author’s estimates of capital allocations and returns on equity. ROBERT S. SEILER JR. 29

Profits declined in 1991 because Fannie Mae raised its capital allocation in anticipation of federal legislation and rose thereafter as the GSE lowered its charge-offs through loss-mitigation techniques. These findings suggest that Fannie Mae has consistently earned profits on its mortgage-credit guarantee business that are greater than those expected in a competitive market. In a competitive market, the after-tax ROE on book-value equity reported by the GSE would be quite close to the rate of return that investors require on the enterprise’s stock. As noted, previous research has estimated that rate of return at about 17 percent. While not definitive, these findings are consistent with the hypothesis that Fannie Mae and Freddie Mac are tacitly colluding in the market for securitizing conforming FRMs. They are also consistent with the decline in the average GSE guarantee fees in recent years. The existence of tacit collusion is consistent with the enterprises’ competing enough so that lenders and mortgage borrowers reap some, but not all, of the savings achieved through loss- mitigation techniques.

Growth of the GSE Portfolios of Retained Mortgages. In the 1980s the collapse of the thrift industry and the rapid appreciation in housing prices enabled Fannie Mae and Freddie Mac to increase their earnings at double-digit rates by focusing on increasing the volume of conforming single-family mortgage debt that they securitized. In the 1990s the GSEs have had to expand their retained portfolios of mortgages, mortgage securities, and other assets rapidly to sustain desired rates of profit growth. Slower appreciation in housing prices in the 1990s has limited the growth of the stock of outstanding single-family mortgages with balances below the conforming-loan limit. Conventional single- family mortgage debt outstanding, which grew at 11 percent annually between the end of 1980 and the end of 1990, has grown at 7 percent a year since 1990 (figure 2–2). Slower growth in the conventional single-family market has limited the ability of Fannie Mae and Freddie Mac to expand the overall volume of mortgages that they finance. Total single-family mortgage debt financed by the GSEs, which increased at an annual rate of 30 percent in 1981– 1993, has grown at about 8 percent a year since 1993. The proportion of outstanding conventional single-family mortgage debt financed by the enterprises has also risen much more slowly since 1993 (figure 2–3). 30 VALUE OF FEDERAL SUBSIDIES

FIGURE 2–2 Conventional, Single-Family Mortgage Debt Outstanding, 1980–1998 (billlions of dollars)

4,000

3,500

3,000

2,500

2,000

1,500

1,000

500

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 3Q1998 End of year

Source: President 1999.

As recently as 1991, Freddie Mac’s policy was to hold in portfolio no more than 5 percent of all mortgages that it financed. In 1992 the agency began to increase the proportion of loans that it financed with debt and gradually made its financing strategy more like Fannie Mae’s strategy, which had always emphasized debt financing.23 In 1993, following the lead of the Federal Home Loan Banks, both enterprises began adding huge amounts of mortgage securities to their portfolios. In most years recently, purchases of mortgage securities have accounted for more than 90 percent of the growth in each GSE’s retained portfolio (table 2–4). The retained mortgage portfolios of Fannie Mae and Freddie Mac averaged about 5 percent of outstanding conventional single-family mortgage debt throughout the 1980s ROBERT S. SEILER JR. 31

FIGURE 2–3 Share of Conventional, Single-Family Mortgage Debt Financed by Fannie Mae and Freddie Mac, 1980–Third Quarter 1998 (percent)

40 Total mortgages financed 35

30 Net MBS outstanding 25

10 Retained portfolios 5

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 3Q1998 End of year

Source: Fannie Mae and Freddie Mac Investor/Analyst Reports and President, 1999.

but had risen to 15 percent of such debt by the end of 1998. Since 1994, each GSE’s portfolio has accounted for more than one-half of its pretax revenues. The enterprises have grown their retained portfolios despite the fact that the government currently requires each to hold more than five times as much equity against a mortgage that it finances with debt as against a loan that it securitizes. On occasion, the supply of mortgage securities may increase rapidly, for example, if many non-GSE investors decide that they would prefer not to manage the prepayment risk associated with securities backed by FRMs. Fannie Mae and Freddie Mac often re- 32 VALUE OF FEDERAL SUBSIDIES portfolio. y Freddie Mac. y Freddie in the in fourth quarters, 1990–1998. Fannie Mae Freddie Mac Freddie Investor/Analyst Reports, TABLE 2–4 (billions of dollars) 0 2 6.4 16.3 31.9 63.7 91.3 116.0 197.9 11.8 16.7 20.5 24.2 44.0 69.7 102.6 125.2 197.4 21.9 26.8 27.3 39.1 41.0 43.8 46.5 48.5 57.1 1990–1998, and 1990 1991 1992 1993 1994 1995 1996 1997 1998 104.9 112.3 137.6 166.6 178.1 183.8 184.4 182.0 218.1 of Fannie Mae and Freddie Mac, 1990–1998 Mae and Freddie Mac, of Fannie Annual Reports, b c Growth and Changing Composition of Retained Mortgage Portfolios Mortgage and Changing Composition of Retained Growth a a : Fannie Mae and Freddie Mac : Fannie Mae and Freddie Mortgages/securities Mortgages Mortgages/securities Mortgages OURCE Mortgage securities’ share of portfolio growthMortgage securities’ share 0% unpaid principal balance less mortgage securities. gross a. Total b. Mortgage-backed securities (MBS) and Real Estate Mortgage Investment Conduit (REMIC) guaranteed by Fannie Mae. 29% REMIC securities guaranteed b bonds, Ginnie Mae mortgage securities, and repurchased c. In 1991–1993, includes mortgage revenue Mac. In 1995–1998, includes all MBS and REMIC securities In 1994, includes only MBS and REMIC securities guaranteed by Freddie 90%S 46% 89% 92% 91% 93% 90% Net guaranteed MBS outstanding of all mortgages financedPortfolio share of portfolioMortgage securities’ share of portfolio growthMortgage securities’ share 29% 1% 288.1 27% 10%Portfolio 40% 355.3 27% 13% 13% 424.4 29% 13% 471.3 11% 486.3 31% 13% 63% 513.2 33% 20% 82% 548.2 34% 28% 98% 579.1 35% 36% 94% 637.1 39% 41% 68% 48% Net guaranteed MBS outstanding of all mortgages financedPortfolio share of portfolioMortgage securities’ share 6% 316.4 0% 7% 359.2 407.6 7% 8% 438.6 19% 11% 459.5 29% 14% 451.4 44% 463.0 19% 454.0 59% 23% 448.5 66% 27% 71% 36% 78% Portfolio ROBERT S. SEILER JR. 33 spond to such market conditions by buying large amounts of mortgage securities. Large purchases of mortgage securities by the GSEs may temporarily support the prices of such assets. But there is no evidence that the enterprises pay above-market prices for mortgage securities. Thus, it is reasonable to assume that Fannie Mae and Freddie Mac retain all federal subsidies associated with the debt issued to finance mortgage securities. The same conclusion holds for the GSEs’ purchases of nonmortgage assets, such as corporate bonds or asset-backed securities.

Conclusion

The privatization studies submitted by the CBO and the Treasury in 1996 use a reasonable method to estimate the public resources devoted to Fannie Mae and Freddie Mac and the allocation of those resources. That method understates the value of the GSEs’ federal subsidies, however, because it assumes that government sponsorship benefits the enterprises only by lowering their borrowing costs. Future research could improve our understanding of the economic effects of the activities of Fannie Mae and Freddie Mac and could refine estimates of the size and allocation of the GSEs’ federal subsidies. Important topics include the size of the GSEs’ cost advantage in issuing debt and equity, the relationship between the size of single-family FRMs and the interest rates on those loans, the effect of geographic diversification of loan pools on the yields of MBSs, whether the enterprises are earning above-normal returns in setting their fees for guaranteeing MBSs backed by FRMs, and the effect of the GSEs’ purchases of mortgage securities on the interest rates on conforming FRMS.

Notes

1. Feldman (1999) and Kane (1999) provide excellent reviews of the literature that complements this chapter. 2. The CBO report was written by Marvin Phaup, the deputy assistant director of the agency’s Special Studies Division. 3. Fannie Mae and Freddie Mac are restricted to financing single-family mortgages that have balances less than certain prescribed levels, which are adjusted annually based on changes in U.S. housing prices. In 2000, this so- called conforming limit was $252,000. This chapter uses the term conforming 34 VALUE OF FEDERAL SUBSIDIES mortgage to refer to any single-family loan with a balance less than the limit. The Departments of Veterans Affairs and of Housing and Urban Develop- ment (HUD) and the Independent Agencies Appropriation Act for Fiscal Year 1999 repealed the limit on the size of multifamily mortgages eligible to be financed by Fannie Mae and Freddie Mac. 4. See Freddie Mac 1996, 32, and the statements by Robert B. Zoellick of Fannie Mae and Leland B. Brendsel of Freddie Mac in Congress 1996, 916– 39 and 951–67, respectively. 5. This statement ignores the cost to the government of exempting Fannie Mae and Freddie Mac from the requirement to register their securities with the SEC. 6. For the most recent OMB assessment of the expected budgetary cost of Fannie Mae and Freddie Mac, see OMB 1999, 202, table 8-1. 7. Woodward (1996) points out that Fannie Mae and Freddie Mac achieve an additional savings by suppressing information about the geographic lo- cation of mortgages in the pools that collateralize the MBSs that they guarantee. This process requires investors to treat each MBS as a generic asset for the purpose of evaluating its prepayment risk, which increases the liquidity of the securities. 8. Continually declining costs associated with becoming informed about the GSEs’ obligations arise from network externalities in the markets for the securities. Network externalities exist when the value of purchasing a good or service rises as the number of other people who have already purchased the good or service increases. For example, the more people who have phones or own computers that use a particular operating system, the more another person will benefit from buying a phone or a computer with that operating system. 9. Debt and other fixed-income obligations that are publicly traded are typically accompanied by one or more credit ratings prepared by private credit-rating agencies. Such ratings are ordinal indicators of the relative risk that issuers of debt or other obligations will default and of the amount of likely default losses. In assigning a credit rating to an issuer’s obligations, rating agencies must make assumptions about the consequences of a decline in its financial condition. At the request of a parent firm, the agencies will privately provide a rating that indicates the risk to which a wholly owned subsidiary exposes the firm. That type of rating is comparable with a rating that expresses the risk that a GSE poses to the government. In determining a risk-to-the-government rating for an enterprise, rating agencies assume that the government would not provide financial assistance to the GSE if its financial condition deteriorated. For a more detailed discussion, see CBO 1991, 50–55 and 58–59. ROBERT S. SEILER JR. 35

10. Relying on year-end 1996 data, Standard & Poor’s Corporation found that the risk to the government posed by the implicit federal guarantee of the obligations of Fannie Mae and Freddie Mac is equivalent, for each GSE, to the risk posed by the obligations of a private firm with a credit rating of AA- (S&P 1997). 11. In calculating that percentage, the CBO excludes from Fannie Mae’s pretax net income in 1995 the value of the GSE’s gift of stock to the Fannie Mae Foundation. 12. The yield on a bond is defined as the implicit annual rate of return earned on the security from the current month to its maturity date. 13. The estimates for 1989 and prior years are from Hendershott 1992. The author estimated the GSEs’ share for more recent years based on publicly available data and the percentages in table 1 on page 100, Cotterman and Pearce 1996. 14. The studies’ econometric models use both the contract and the effective mortgage rate as the dependent variable. The effective rate builds in discount points paid over a ten-year horizon. 15. The HSH data, which were purchased by the Congressional Budget Office and are copyrighted by HSH, include mortgage rates and points. In calculating the spread between the effective rate on jumbo and conforming FRMs, the author assumed that each point adds 20 basis points to the yield. Exhibit 16 on page 42 of Freddie Mac 1996 shows similar data from the GSE’s Primary Mortgage Market Survey and HSH Associates for July 1996 to March 1996, with effective rates calculated in the same way. 16. The method adjusts for the timing difference by using the HSH data from the first month of each quarter and the last two months of the previous quarter to estimate the jumbo-conforming differential for the quarter. 17. The mean and the standard deviations of the quarterly estimates of the jumbo-conforming spread estimated by Cotterman and Pearce for eleven states in 1989–1993 are 37 and 12 basis points, respectively (table 7, page 126). The mean and the standard deviations of the quarterly estimates of the spread calculated from the HSH data for the same period are 39 and 14 basis points, respectively. 18. This subsection draws on unpublished research by Treasury Depart- ment economists. GAO 1996, 58–62, also has a useful discussion of the effect of loan size on mortgage rates. The term mortgage rate refers to the effective rate, which includes the coupon rate and amortized discount points. 19. How Fannie Mae and Freddie Mac behave in setting their guarantee fees is linked to how they behave in buying MBSs and other mortgage securities. Each GSE’s demand for its own MBSs affects the market prices of those securities. The higher those prices, the higher the guarantee fees the 36 VALUE OF FEDERAL SUBSIDIES enterprise can charge, everything else being equal (the fees are paid from the spread between the average interest rate on the loans backing the MBSs and the interest rate paid to MBSs investors). Thus, to the extent that Fannie Mae and Freddie Mac tacitly collude in setting their guarantee fees, they also are tacitly colluding in setting the prices that they pay for MBSs and other mortgage securities. 20. Other models of duopoly imply that outcomes will be the same as under competition. The simplest such model, known as the Bertrand model, assumes that the two firms offer identical goods, have equal and constant marginal cost, and make simultaneous decisions about price. These assumptions imply that each firm chooses the competitive price, since all customers can easily shift to the lowest price firms. Price equals marginal cost, and the firms earn a normal profit. 21. Goodman and Passmore (1992) consider the pricing behavior of Fannie Mae and Freddie Mac in securitizing single-family mortgages in the context of models of the behavior of duopolists in markets where demand fluctu- ates cyclically. Other evidence of procyclical pricing may be cited; for example, the GSEs increased mortgage insurance coverage requirements in 1994 after interest rates increased and the demand for securitization declined. More precise tests of this hypothesis would have to consider cyclical fluc- tuations in mortgage credit quality, which declines on average when interest rates rise and origination volume falls. 22. Portfolio theory indicates that investors in each GSE must earn an after-tax ROE from holding the enterprise’s stock equal to the risk-free rate of interest plus a risk premium that reflects the covariance between the return on the stock and the return on the market as a whole. If Fannie Mae and Freddie Mac are behaving as they would in a competitive market, the average after-tax ROE that each GSE earns over time on its book- value equity should be close to that minimum. In fact, Fannie Mae and Freddie Mac consistently report book-value ROEs much greater than 20 percent a year. 23. In September 1989, Leland Brendsel, the chairman and chief executive officer of Freddie Mac, assured the chairman of the Oversight Subcom- mittee of the Ways and Means Committee of the U.S. House of Representatives (in an oral response to a question at a hearing) that Freddie Mac would not increase its debt-financed portfolio beyond 5 percent of all assets financed (Congress 1990, 99). However, as the subcommittee chairman observed, federal law did not bind the enterprise to that policy. After the Financial Institutions Reform, Recovery, and Enforcement Act of 1989 (FIRREA) allowed Freddie Mac to issue stock to the public and created a stockholder-elected board to run the GSE, the new board of directors re- versed the policy. ROBERT S. SEILER JR. 37

References

Alvarez, Aida. 1996. “Testimony.” In Oversight of the Federal National Mort- gage Association (Fannie Mae) and the Federal Home Loan Mortgage Corpora- tion (Freddie Mac) (Hearing before the Subcommittee on Capital Markets, Securities, and Government-Sponsored Enterprises), U.S. Congress, House of Representatives, Committee on Banking and Financial Services, 104th Congress, 2d session, April 17, June 12, July 24, 31, August 1. Serial 104- 55, 838–61. Ambrose, Brent, and Arthur Warga. 1996. “Implications of Privatization: The Costs to FNMA and FHLMC.” In Studies on Privatizing Fannie Mae and Freddie Mac, edited by U.S. Department of Housing and Urban Develop- ment, Office of Policy Development and Research, 169–204. Washington, D.C.: Government Printing Office. Bosworth, Barry P., Andrew S. Carron, and Elizabeth H. Rhyne. 1987. The Economics of Federal Credit Programs. Washington, D.C.: Brookings Institution. Cook, Douglas O. 1996. “Review of the Ambrose-Warga and Cotterman- Pearce Papers.” In Studies on Privatizing Fannie Mae and Freddie Mac, edited by U.S. Department of Housing and Urban Development, Office of Policy Development and Research, 211–17. Washington, D.C.: Govern- ment Printing Office. Cotterman, Robert F., and James E. Pearce. 1996. “The Effects of the Federal National Mortgage Association and the Federal Home Loan Mortgage Corporation on Conventional Fixed-Rate Mortgage Yields.” In Studies on Privatizing Fannie Mae and Freddie Mac, edited by U.S. Department of Housing and Urban Development, Office of Policy Development and Research, 97–168. Washington, D.C.: Government Printing Office. Downs, Anthony. 1980. “Comments on Silber and Black.” In Conference on the Economics of Federal Credit Activity. Part 1: Proceedings, edited by U.S. Congressional Budget Office, 57–60. Washington, D.C.: Government Print- ing Office. Federal Home Loan Mortgage Corporation (Freddie Mac). 1996. Financing America’s Housing: The Vital Role of Freddie Mac. McLean, Va.: Freddie Mac. Federal National Mortgage Association (Fannie Mae). 1996. “Fannie Mae Review of the Cotterman-Pearce and Ambrose-Warga Papers.” In Studies on Privatizing Fannie Mae and Freddie Mac, edited by U.S. Department of Housing and Urban Development, Office of Policy Development and Re- search, 218–21. Washington, D.C.: Government Printing Office. Feldman, Ron. 1999. “Estimating and Managing the Federal Subsidy of Fannie Mae and Freddie Mac: Is Either Task Possible?” Journal of Public Budgeting, Accounting, and Financial Management 11 (1) (spring): 81–116. 38 VALUE OF FEDERAL SUBSIDIES

Gatti, James F., and Ronald W. Spahr. 1995. Impact of Affordable Housing Goals: Capital Costs and Capital Requirements of FNMA and FHLMC. Report prepared for the U.S. Department of Housing and Urban Development. Goodman, John L., and S. Wayne Passmore. 1992. Market Power and the Pric- ing of Mortgage Securitization. Finance and Economics Discussion Series. Washington, D.C.: Federal Reserve Board. Hendershott, Patric H. 1992. “The Market for Home Mortgage Credit: Re- cent Changes and Future Prospects.” In The Changing Market in Finan- cial Services, edited by R. A. Gilbert, 99–123. Boston: Kluwer Academic Publishers. Hendershott, Patric H., and James D. Shilling. 1989. “The Impact of the Agen- cies on Conventional Fixed-Rate Mortgage Yields.” Journal of Real Estate Finance and Economics (2): 101–15. Hermalin, Benjamin E., and Dwight Jaffee. 1996. “The Privatization of Fannie Mae and Freddie Mac: Implications for Mortgage Industry Structure.” In Studies on Privatizing Fannie Mae and Freddie Mac, U.S. Department of Housing and Urban Development, Office of Policy Development and Research, 225–302. Washington, D.C.: Government Printing Office. ICF, Incorporated. 1990. Effects of the Conforming Loan Limit on Mortgage Mar- kets. Final report prepared for the U.S. Department of Housing and Ur- ban Development, Office of Policy Development and Research. Fairfax, Va.: ICF. Kane, Edward. J. 1999. “Housing Finance GSEs: Who Gets the Subsidy?” Journal of Financial Services Research 15 (3): 197–209. Lekkas, V., and R. Van Order. 1991. “Taking Stock.” Secondary Mortgage Mar- kets (fall): 24–27. Miller, Jones. C., III, Robert D. Tollison, and R. S. Higgins. 1996. “A Critical Review of the Ambrose and Wanga Privatization Study on Behalf of Freddie Mac.” In U.S. House Committee on Banking and Financial Ser- vices, pp. 888–914. Shepherd, William G. 1990. The Economics of Industrial Organization. 3rd ed. Englewood Cliffs, N.J.: Prentice Hall. Shilling, James D. 1996. “Comments on the Ambrose-Warga and Cotterman- Pearce Papers.” In Studies on Privatizing Fannie Mae and Freddie Mac, U.S. Department of Housing and Urban Development, Office of Policy Devel- opment and Research, 205–10. Washington, D.C.: Government Printing Office. Standard & Poor’s. 1997. Final Report of Standard & Poor’s to the Office of Fed- eral Housing Enterprise Oversight. Contract HE09602C. Ugoletti, Mario. 1997. “Issues in Measuring the Benefits of Government- Sponsored Enterprise Status and Costs to the Government.” Paper pre- ROBERT S. SEILER JR. 39

sented at the twentieth annual research conference of the Association for Public Policy Analysis and Management. U.S. Congress, House of Representatives, Committee on Banking and Fi- nancial Services. 1996. Oversight of the Federal National Mortgage Associa- tion (Fannie Mae) and the Federal Home Loan Mortgage Corporation (Freddie Mac) (Hearing before the Subcommittee on Capital Markets, Securities, and Government-Sponsored Enterprises), 104th Congress, 2d session, April 17, June 12, July 24, 31, August 1. Serial 104-55. U.S. Congressional Budget Office.1991. Controlling the Risks of Government- Sponsored Enterprises. Washington, D.C.: Government Printing Office. ———.1996. Assessing the Public Costs and Benefits of Fannie Mae and Freddie Mac. Washington, D.C.: Government Printing Office. U.S. Department of Housing and Urban Development. 1996. Studies on Priva- tizing Fannie Mae and Freddie Mac. Washington, D.C.: Government Print- ing Office. U.S. Department of the Treasury. 1996. Government Sponsorship of the Federal National Mortgage Association and the Federal Home Loan Mortgage Corpora- tion. Washington, D.C.: Government Printing Office. U.S. General Accounting Office. 1996. Housing Enterprises: Potential Impacts of Severing Government Sponsorship. Washington, D.C.: Government Print- ing Office. U.S. Office of Management and Budget. 1999. Analytical Perspectives, Budget of the United States Government, Fiscal Year 2000. Washington, D.C.: Gov- ernment Printing Office. U.S. President. 1999. Economic Report of the President, 1999. Washington, D.C.: Government Printing Office. Viscusi, W. K., Vernon, J. M., and Harrington, J. E. 1996. Economics of Regula- tion and Antitrust. 2nd ed. Cambridge, Mass.: MIT Press. Woodward, Susan E. 1996. Rechartering Fannie Mae and Freddie Mac: The Policy Issues. Working paper. Stanford Law School. Since the publication of the CBO and Treasury studies of the GSE subsidies, Fannie and Freddie have not attempted to demonstrate how or why the services that they furnish to the mortgage market are worth the cost to the federal government. They have largely confined themselves instead to denials of any subsidy and to hints of errors in the CBO and Treasury analyses. Robert Van Order, chief economist of Freddie Mac, has come forward with an analysis that offers a sophisticated economic rationale for continuing government support of Fannie and Freddie. He does not deny (or ad- mit) that Fannie and Freddie receive subsidies, but he contends that the GSEs are more efficient operators than the banks are; he characterizes both as receiving some form of government support to pursue the nation’s housing policies. Accordingly, whether or not they are optimal mechanisms, the GSEs contribute to society’s overall welfare by outcompeting the banks. By im- plication, if anyone wants to increase the efficiency with which government subsidies are used, the banks—rather than Fannie and Freddie—should be the first targets. Because this argument does not depend on the unsup- portable notion that Fannie and Freddie do not receive subsidies, it is the most sophisticated and credible position either company has advanced since the CBO and Treasury studies in 1996; it is published here for that reason.

40 3 The Economics of Fannie Mae and Freddie Mac

Robert Van Order

The central policy issue with government-sponsored enterprises (GSEs) such as the Federal National Mortgage Association (Fannie Mae) and the Federal Home Loan Mortgage Corporation (Freddie Mac) is embedded options. These options primarily take the form of conjectured guarantees and conjectures about conjectured guarantees. Depending on the point of view, they can be viewed as either subsidies or tools. Of course, they are both. The central focus of GSE regulation should probably be the control of the embedded options, so that their value matches public benefits, and that analysis should center on incentives for the GSEs and implications for economic welfare. However, GSE costs and benefits are complicated, both conceptually and empirically. My focus here is on methodology and research questions rather than on specific answers. I propose some models that, while consistent with history, are meant to be illuminating counterexamples rather than well-developed theories; what follows should be viewed in part as a research agenda. Research on Fannie Mae and Freddie Mac has not always been lucid, well grounded in economic theory, or serious about data. I want to do two things: (1) analyze what existing data can and cannot support and (2) analyze what standard neoclassical welfare economics might have to say about policy and the GSEs. I conclude that the data are too shaky to support much and that welfare economics does not imply unambiguous policies. Taking the

This essay does not necessarily represent the opinions of Freddie Mac.

41 42 THE ECONOMICS economics seriously limits what can be believed simultaneously. For instance, it is difficult to believe both that the GSEs convey distortionary subsidies to the mortgage markets, which make mortgage rates too low, and that they stifle competition and make rates too high.

Microfoundations

An analysis of the GSEs needs some sort of microfoundation.

The Industry. The relevant industry for the GSEs is not the secondary market, but rather the mortgage market as a whole. The distinction between the primary and secondary markets is largely irrelevant, economically. What matters is the different ways of getting money from capital markets as a whole to borrowers in the mortgage market. Two major routes for conventional mortgages exist: depositories and the GSEs. Both take advantage of some sort of federal backing. The fully “private” market is small and has been small for at least the past fifty years. The development of the mortgage market over the past twenty years has primarily been characterized by one type of federally sponsored institution (the GSEs) taking some of the market share of another (the depositories). The existence of some sort of guarantee has not really changed, and the value of the guarantee may have diminished. The competitive structure might best be characterized as that of dueling charters. There are two major charters in the industry: one for depositories and one for GSEs. Both involve imbedded options, and both have offsetting restrictions on lines of business, capital regulation, and mission regulation. Both types of charter promote different ways of connecting mortgage markets with the financial markets. What matters is the pluses and minuses of the two charters. Banks are no more private than GSEs. Competition primarily concerns leveraging these charters and the business lines allowed by them. Historically, the balance has been determined by cyclical movements in thrift capital levels and balancing the economies that the GSEs have in raising money (and their lower other costs) with the advantages that local depositories have in selecting the best loans, as well as the details of other aspects of their charters. Competitive balance is likely to be strongly affected by the ability to take advantage of information technology, as well as the ability to leverage the charters. ROBERT VAN ORDER 43

And there is competition. The GSE share of the conventional market (the percentage of conventional originations) was indeed high in the early 1990s, when the savings and loans were in trouble, although it has fallen since. The figure exceeded 60 percent in 1993, but it fell to below 30 percent for a while in 1994 and has fluctuated since, with a steady state around 45 percent (slightly more than half of conventional conforming). The recent strength of the depositories and innovations—such as a recent plan of the Chicago Federal Home Loan Bank to act as a capital market funding source for thrifts and the advent of depositories substituting uninsured 80 percent first mortgages combined with uninsured 15 percent second mortgages for 95 percent first mortgages with mortgage insurance—indicate that competition is far from trivial. The point is not that price equals marginal cost. Rather, it is not clear that the difference between price and marginal cost in the mortgage market varies from that in most other industries. With all competition taken into account, the gap is likely to be smaller than imagined. No empirical work supports any position. Economic theory says little more than that price should be somewhere between the monopoly price and marginal cost.

Benefits. Whether the GSEs get a subsidy has been a source of controversy. Clearly they borrow at lower rates than if they had the same portfolio but a different (nondepository) charter. The consensus (for example, recent stand-alone ratings by Standard & Poor’s) is that both companies would be in the (low) AA range on their own, but they would borrow at less than AAA rates. This difference looks like a subsidy. It is off budget; not all of it can be considered a “net” benefit; and it is not just a guarantee fee in the sense of measuring the expected costs to taxpayers of a bailout (for example, it includes other charter benefits, such as scale economies and uniqueness of the charter, which are benefits but do not increase the probability of bankruptcy—or maybe decrease it). Nonetheless, in terms of economics, the interest rate difference acts like a subsidy and affects resource allocation like a subsidy. But words can be confusing; so let us call it a benefit and distinguish between the gross benefit, which equals the spread just discussed, and the net benefit, which actually affects resource allocation. This distinction is discussed below. The size of the gross benefit has been overestimated. It is relatively easy to look up interest rate spreads. Using Bloomberg data 44 THE ECONOMICS like those used in an earlier Treasury study, the spread between “Agencies” and AA for comparable debt has ranged from about 15 to 36 basis points since 1992. In the last quarter of 1998, the spread was less than 20. For A, the range has been 20–75 basis points, with about 50 in the last quarter of 1998. A conservative range is 20–40 or 50, rather than the much higher levels quoted in the Treasury and the CBO studies. The difference between this and the Treasury’s earlier results is primarily that the Treasury looked at A rate spreads (rather than AA minus), and spreads have fallen since the study. Ex ante measures (quoted yields) are better than ex post returns because they are more like the market’s shadow prices for credit risk at a moment in time. Looking at callable debt can be tricky because it requires pricing callability, which is difficult (and exercising call options is not an important aspect of GSEs as credits). The large market in noncallable agency debt should reveal all that the market knows about the GSE benefit. Past estimates that did not do this and did not adjust well for callability had unreasonably large and incon- sistent results (for example, the results used by the Congressional Budget Office in 1996 had a wildly implausible 60 basis point difference between the subsidies for callable and noncallable bonds). In comparing benefits with the effects of mortgage rates, the same periods must be used. Also, constraints from the charter should, in principle, be netted out. Conversely, there are other advantages such as exemptions from state and local (but not federal or property) income taxes and registration with the Securities and Exchange Commission. These are probably worth 2 or 3 basis points and are swamped by varia- tions in borrowing spreads. Similar calculations can be made for depositories. This process is difficult because most are unrated, but it is doubtful that much of the industry is close to A or AA. Few banks have ratings of AA or better (none are AAA now); most are in the A to BBB range. Most thrift institutions are rated BBB or worse. Most of the thrift industry is unrated and is presumably worse than these ratings. Deposit insurance premiums should be netted out when they are paid (most banks do not pay premiums), but even then the depository gross benefit may exceed the GSE gross benefit. For instance, assume that the deposit insurance premium is 20 basis points and the GSE benefit is 30 basis points; then the benefit to depositories is greater than the benefit to GSEs if depositories have to pay 50 basis points or more than current deposit rates absent deposit ROBERT VAN ORDER 45 insurance. The Bloomberg data suggest that, as a crude benchmark, spreads between federal agency debt and BBB tend to be 40–100 basis points (1992–1998). At the end of 1998, the spread was high, slightly more than 100. Spreads from Treasuries have run between 50 and 160. If this benefit is a measure of social cost, and if the GSEs mainly substitute for depositories, then the GSEs could lower the social cost.

Why Have GSEs? GSEs are a compromise between government agencies, such as the Government National Mortgage Association (Ginnie Mae), and fully private businesses. The justification for such hybrids is that some policy goals can be accomplished better with private-sector incentives (for example, profit maximization, cost minimization). The policy problem is balancing these efficiencies with incentives to exploit implicit subsidies. The GSEs appear to have four major policy goals:

1. demonstration effects—from such tools as mortgage-backed securities (MBSs) and collateralized mortgage obligations (CMOs)—as well as the development of new (underserved) markets 2. the extension of homeownership (including specific underserved markets) 3. the maintenance of open markets and of liquidity 4. a more efficient way of supporting mortgage markets than through depositories.

The first goal may fade over time, although recent developments in automated underwriting by the GSEs, which are being imitated in the rest of the industry, do not suggest this. Many economists are skeptical of the second goal, and other tools do exist to accomplish this. Nonetheless, some recent research does suggest social benefits from homeownership. This effect could be important as an added benefit, especially if targeted to particular underserved groups for which homeownership’s social benefits are greatest; such groups might be hard to target with other policies. The third goal is a “traditional” one and includes aspects from smoothing short-term dis- turbances in the market to preventing all-out meltdowns. The last goal assumes that support for the mortgage market will continue and argues that the GSEs are a more efficient form of such support than other institutions. 46 THE ECONOMICS

Incentives. Incentives for the firms in the industry are complicated. Clearly GSEs and depositories can be expected to take advantage of their charter benefits. A firm with a completely unconstrained guarantee will tend to take on as much risk as it can to exploit the guarantee. But guarantees are never unconstrained. If an institution can be shut down—for example, by a regulator—it will have incentives to control risk so as to maintain access to future benefits, even in the absence of safety and soundness regulation. Hence, while a perceived guarantee might appear to provide incentives to take excessive risk, the GSEs also have a franchise—which can be interpreted as the present value of access to future charter benefits (because of the scarcity of their charters)—that induces them to be conservative. Indeed maintaining the franchise has probably been the most important factor so far in limiting the risk-taking of Fannie Mae and Freddie Mac. Of course, both institutions do take on some risk. The franchise is worthless if they only hold match-funded Treasuries. The major hazard of risk-taking is the option to take on more risk if the GSEs get into trouble—and if the regulators do not notice. Charters give implicit assets and liabilities, such as a franchise, to firms; one should look at these extended balance sheets in the context of value maximization. The embedded options that come from explicit or conjectured guarantees are of particular interest, but other implicit assets and liabilities also affect value maximization.

Policy. The key policy questions are best posed through standard welfare economics models, with an emphasis on consumer surplus, market failure, and so forth. Distributional consequences may not be important. In particular, an analysis of the costs of GSEs should focus on the effects on resource allocation. This process requires measuring the costs relative to what would happen in their absence; this line, in turn, requires an analysis of what the market structure would be in their absence. If the alternative is—as I am suggest- ing—dominance by the depository charter, then the costs should be relative to the costs under that charter rather than relative to a purely private market.

Modeling GSEs and Their Market

Behavior. GSEs have extended balance sheets with two implicit, but nonetheless real, assets: G, the conjectured guarantee, which gives them lower borrowing rates, and F, the franchise. Otherwise, they ROBERT VAN ORDER 47 have the usual collection of assets, liabilities, and shareholder equity (all measured at market values or estimates of shadow-market values). In the simplest case, F is the ability to get future guarantees; more broadly, however, F is the ability to get future benefits of the charter as well as the usual franchise elements accruing to all firms that act as ongoing businesses. Embedded options are complicated options on options with correspondingly complicated incentives. In the simplest competitive model, GSEs choose their asset- liability structure so as to maximize G + F. The optimization is complicated and depends critically on regulatory structure and closure rules. The supply curve for Fannie Mae and Freddie Mac comes from this. Depositories have a similar supply curve, although it might be argued that their franchise is worth less. The value of G + F will show up in the stock price and will contribute to market value exceeding book value, but it is difficult to go backward to infer the value of the benefits from the excess of market over book value. Consider the following simple model. It is straightforward to show that the present value over an infinite time horizon of future dividends is given by

V = daK/( r – (1 – d) a) (3–1) where a = return on equity (ROE) d = dividend rate r = risk-adjusted discount-rate value K = capital (also approximately book value)

I am ignoring taxes. When a = r, V = K. For most firms a is bigger than r, representing normal franchise value. Suppose that for a typical firm, a = .20 d = .2 r = .18 then V = 2K. So market is twice book value.

Suppose the benefit that GSEs get from their charter is only 5 percent of profits, so that they have an ROE of .21 instead of .20. Then the new solution is V = 3.5K. 48 THE ECONOMICS

Market value is now almost four times book value. Hence, going backward and inferring the amount of the benefit from looking at stock prices is likely to be deceptive. Large excesses of market over book value relative to those of other firms are consistent with relatively small subsidies. Indeed, in this model, retained subsidies as large as those contemplated by the CBO and Treasury would imply much larger shares of GSE profits than 5 percent and would drive V to infinity.

Welfare Effects. The welfare effects of GSEs inevitably involve second- best issues because the rationale for GSEs requires market imperfections in the first place. Economists’ central criticism of Fannie Mae and Freddie Mac has generally been that they distort mortgage rates and resource allocation by diverting funds toward housing and away from other types of investment. Proponents argue, for instance, that the GSEs produce benefits by stabilizing the market (ensuring that it is always open) and promoting homeownership. A major second- best issue is whether distortions caused by the GSEs are as great as those caused by deposit insurance for depositories.

A Dueling-Charter Model of Mortgage Markets

Model l. This model focuses on the fourth goal in the list. There are no externalities, the private demand for mortgage funds fully reflects social benefits, and general equilibrium effects are ignored. Supply and demand are as described, assuming normal profits and a supply curve that is flat. I begin with a fully private supply curve, which, by assumption, reflects social costs. Equilibrium is at point B in figure 3–1. It is a social optimum. Now introduce depositories, which are assumed to have costs greater than those of the fully private market, but a borrowing-cost advantage from their charter equal to EC, so that the actual supply curve in the market is one with the points C and G, while the unsubsidized curve has the points A and E. In this model, the depositories dominate the fully private market. I interpret this as approximately the structure of the mortgage market before the GSEs. The deadweight loss from the depositories is the area inside the triangle AEC. Now introduce the GSEs into the conforming market (currently mortgages below $275,000) with the same social costs as the private supply curve but with a subsidy such that in the conforming market the GSEs now dominate the depositories. Their supply curve is the ROBERT VAN ORDER 49

FIGURE 3–1 Model 1 of Mortgage Markets Mortgage rate

Social cost of depository A E

“Private” supply B F (= social cost)

Depository supply G C GSE supply D

$ Note: The GSE monopoly would increase welfare. 1. Area inside ACE = deadweight loss from depositories 2. Area inside BDF = deadweight loss from GSEs 3. B = social optimum 4. GD/FD = CBO measure of sponginess

lowest of the curves, with the point D on it. In this model (with flat supply curves), the market is perfectly segmented. No “private” market exists, and the conventional market is divided between the conforming market, which is dominated by the GSEs, and the nonconforming market, which is dominated by the depositories. Of course, that is a bit unrealistic; presumably, the supply curves have some slope, but that situation does not affect the basic results here. The deadweight loss in the conforming market is now the area inside the triangle BFD. This can be either bigger or smaller than the 50 THE ECONOMICS deadweight loss from the entirely depository-based system, depending on whether the gross benefit (the Treasury-CBO type of estimate) to the GSEs is bigger or smaller than the gross benefit to depositories (in the example above, whether depositories could attract deposits at 50 basis points or less, absent insurance). Intuitively, in this model, the GSEs have lower production costs than depositories (this situation tends to be welfare increasing), but the costs convey subsidies (this aspect is welfare reducing). If the gross benefit of the subsidies is equal to or less than that for depositories, then the subsidies are welfare increasing. The gross benefit to the GSEs can be broken up into two parts: FG, which is necessary to enter the industry, and GD, which is the net benefit and equals the extent to which the GSEs lower mortgage rates (in this model, that extent equals the difference between the conforming and nonconforming rates). Hence, in the context of the market structure described above (dueling charters), GSEs could be welfare improving (and could lower expected taxpayer cost), relative to the alternative.

Model 2. This model addresses goals one through three by intro- ducing social benefits in a simple way. Model 1 is a second-best model, with no particular reason for intervening in mortgage markets in the first place. Suppose that at least one of the above three reasons is valid (for example, Fannie Mae and Freddie Mac stabilize the market and thus lower long-run social costs by reducing disruptions that drive wedges in market rates and marginal products of capital) and that the goal can be represented by GSEs having a lower social cost than the purely private market. This situation is depicted in figure 3–2, which differs from figure 3–1 because of the lower social costs of GSEs given by the line with KHI and J. In this model the GSE market-supply curve is the same as in figure 3–1 (with D on it). Neither the fully private market, the depository-based market, nor the GSE-based market is first-best (the deadweight loss from the purely private market is the area inside the triangle KBH). If the GSE gross benefit were FJ and the depository benefit were less than EI, there would be a first-best optimum. But with the model as depicted in figure 3–2, there would be no first-best optimum, and the depository benefit is greater than EI. The deadweight loss from the GSEs (the area inside HJD) is now smaller, and the GSEs are more likely to be welfare increasing, even without distortions from deposit insurance. ROBERT VAN ORDER 51

FIGURE 3–2 Model 2 of Mortgage Markets Mortgage rate

Social cost of depository E A “Private” supply B F (= social cost)

H I J Social cost of GSE K G C Depository D supply

GSE supply

$ Note: Monopoly power may not increase welfare (if FD < BK). 1. Area inside HJD = deadweight loss from GSEs. 2. Area inside KBH = deadweight loss from free market. 3. Area inside ACE = deadweight loss from depositories. 4. Social optimum at H. 5. GSE may have lower deadweight loss than free market.

Considerations. How big is the benefit? In model 1 the gross benefit to Fannie Mae and Freddie Mac—the difference between its cost of funds and the purely private market’s cost—is the distance FD. How- ever, the distance FG is not a benefit to the GSEs because it is what is necessary to keep them in business along with the depositories. The net benefit is only GD. The benefit to depositories (gross and net are the same) is the distance EC, which may be bigger than the net benefit to the GSEs. Looking at the difference between bank commercial paper rates and deposit rates will give a number more like FG and 52 THE ECONOMICS will understate benefits because it does not account for the higher costs of depositories in the model. The guarantee component of the benefit is less than the gross benefit, but the amount is unknown. A measure of this would be the expected present value of costs of paying off debt-holders. This measure is not easy to model: it could be bigger or smaller than the net benefit. The gross benefit to GSEs can be measured, as discussed, but the net benefit can, at best, be inferred only indirectly. What does the benefit do? In model 1 the effect of the benefit is ambiguous: while, on the one hand, it extends the subsidy to homeownership and housing (this action has a negative effect in the model), the real resources used in producing mortgages, on the other hand, are less than those used with depositories (an action that off- sets some or all distortion). In model 2, where there are some fundamental benefits for the GSEs in the housing market, the distortion may be a net benefit. Who gets the benefit? The CBO measured the extent to which the GSE benefit is passed through to borrowers by the ratio of the extent to which Fannie Mae and Freddie Mac lower mortgage rates to the extent to which their cost of funds is lowered by their charter; that figure, in turn, is what I have defined as the gross benefit. In model 1 the GSEs lower mortgage rates by DG, which can be measured, as in the CBO study, by the difference between nonconforming and conforming rates. That study concludes that this ratio is only about 60 percent. There are two problems with the estimate:

1. The gross benefit is overestimated. As discussed, the best estimate of the funding advantage is more in the range of 30 or 35 basis points, which is similar to estimates of the benefits given by the difference between nonconforming and conforming rates. Of course, these numbers are imprecise, but they are sufficiently close that the data are consistent with no difference. Also, the rate spread between nonconforming and conforming loans will underestimate the effects of GSEs on conforming rates because conforming and nonconforming mortgages are substitutes (for example, in thrift portfolios); lowering one rate should lower the other as well (in figure 3–1 the supply curves are apparently not perfectly flat). 2. The calculation should use the net benefit. For instance, figure 3–1 is consistent with a CBO ratio of 60 percent. But by construction GSEs always earn only normal profits (the ROBERT VAN ORDER 53

model assumes perfect competition), so any benefits that they receive must be passed through. The difference is that some of the gross benefit simply cancels out the benefit to depositories. Hence, even in principle, the CBO number does not measure what it purports to measure. In principle the net benefit will be reflected in the stock price, but, as discussed, inferring that figure from actual data is likely to be difficult and almost certainly cannot tease out an implied guarantee fee.

Finally, the GSEs may keep some benefit in the form of a higher return on equity than normal. However, that amount would not go to current shareholders, because in an efficient market, like the stock market, it would be capitalized into stock price; current stockholders paid for it when they bought the stock and cannot be expected to earn more than normal returns. The benefits of the GSE charter have, indeed, been auctioned off. In short it is difficult to believe that distributional effects of Fannie Mae and Freddie Mac are important. Policy choices should probably focus on the resource allocation questions discussed above.

Regulatory Choices

The regulation of Fannie Mae and Freddie Mac has been suggested for three reasons: 1. safety and soundness, 2. mission, and 3. competition. These considerations lead me to the following six points.

1. Congress wants GSEs because whatever goal Congress has in mind can be accomplished better by a profit-oriented institution than by a government agency. 2. The charter acts as a constraint on the lines of business that the GSEs can enter. This limit is meant to ensure that they stick to their purpose. 3. But there are embedded options. The benefits of a GSE are supposed to come from the efficiency that results from maximizing profits by responding to demand and from minimiz- ing costs, rather than from exploiting options. Some conflict between the two is inevitable, particularly if a GSE gets into trouble. Hence, on grounds of efficiency alone, regulation for safety and soundness is needed. Given the depository charter in model 1, the second-best solution is a gross benefit to GSEs that slightly exceeds FG. In model 2, where there 54 THE ECONOMICS

are real reasons for intervention, the optimal gross benefit is still a bit more than FG, but even without the depositories, it is still positive and equal to FJ. One could imagine choosing capital requirements to keep the benefits at these levels. Other issues such as general equilibrium effects could affect the results. 4. Regulation to promote competition may be inefficient. In the first place, if the relevant industry is the mortgage industry as a whole, the conjectured level of competition is much greater than if the market were a duopoly. If the central criticism of Fannie Mae and Freddie Mac is that they distort resource allocation and cause the overproduction of housing (model 1), then more competition further distorts resource allocation. Furthermore, monopoly power, if it exists, provides incentives to take less risk to keep the franchise. Surely one of the major factors in the thrift debacle in the 1980s was the lack of franchise value, which left savings and loans nothing to lose by gambling. Those who believe in model 2 can worry about competi- tiveness, but then they have to acknowledge the GSE benefits. One cannot believe both model 1 and the absence of enough competition. 5. I have focused on efficiency. For reasons discussed, I see no real equity issue. One cannot believe both that the stock market is efficient and that current shareholders make excess profits systematically. This is not to say that franchise value is insignificant, but franchise value also provides incentives to limit risk-taking. 6. Mission issues, though interesting and controversial, are beyond the scope of my comments.

Given the mission and the charter, the focus of regulation should be on safety and soundness, with an eye to keeping embedded options under control and equal in value across charters. That objective is not likely to be easy. But once this is done, regulation can focus on mission issues in terms of costs and benefits. Such was the intention of the 1992 regulatory legislation that established the Of- fice of Federal Housing Enterprise Oversight and defined mission goals for Fannie Mae and Freddie Mac; this thinking is consistent with recent approaches to bank regulation, which focus on capital adequacy and the control of risk, rather than on micromanagement. Statements contending that the GSEs and national banks are similar institutions—first articulated in Robert Van Order’s essay—have become a staple of the GSEs’ defense against criticism. Spokespersons for Fannie Mae and Freddie Mac have argued that the GSEs’ high levels of profitability are the result of efficiency of operation and have noted that they have lower operating costs than banks. The Van Order argument went somewhat further in an attempt to view GSEs and banks as two kinds of government-supported entities that were established to carry out national housing policy and are ultimately in competition with one another. He sees the mortgage markets, in a colorful phrase, as a case of “dueling charters.” In part the lower costs cited by Fannie and Freddie spokespersons result because Fannie and Freddie operate in the wholesale secondary market and have none of the costs of maintaining branches, originating loans, or servicing individual accounts. But the distinctions between national banks and GSEs go much further than this substantial difference in their operations; national banks function under an entirely different statutory and regulatory structure than do the GSEs. In the following essay Professor Carnell details the significant differences between the statutory and regulatory regimes under which GSEs and national banks function. These differences, as Professor Carnell shows, offer substantially greater financial benefits to the GSEs—with significantly less direct or indirect cost—than what is made available to banks. Professor Carnell notes the strong argument that banks receive no net subsidy from the federal government, after taking account of the cost of deposit insurance and the regulatory system in which they function. All these differences complicate any assertion that GSEs and banks are in any way comparable entities.

55 4 Federal Deposit Insurance and Federal Sponsorship of Fannie Mae and Freddie Mac: The Structure of Subsidy

Richard Scott Carnell

“When facing attack, seek safety in numbers”: good advice in war- fare, diplomacy, politics, and public relations. Small wonder that the Federal National Mortgage Association (Fannie Mae) and the Federal Home Loan Mortgage Corporation (Freddie Mac)—each a master of those venerable arts—would look around for company when critics argue that the two firms receive overly generous benefits from the federal government.1 Accordingly, Fannie and Freddie have sought to characterize federally insured depository institutions as government-sponsored enterprises receiving federal benefits no less significant than those provided to Fannie and Freddie. I describe the federal government’s sponsorship of Fannie Mae and Freddie Mac and recount the two firms’ recent attempts to de- pict depository institutions as GSEs. I then outline six structural reasons why—contrary to what one might expect—federal sponsorship of Fannie and Freddie, with its perceived implicit backing tends to impart a greater net subsidy than explicit federal deposit insurance.

Fannie Mae and Freddie Mac as GSEs

Government-sponsored enterprises2 are a particular type of federal instrumentality (Moe and Stanton 1989, 323–24). Although privately owned, profit-oriented corporations, GSEs have congressional charters and public missions and receive an array of governmental benefits not available to businesses generally (Treasury 1996, 25–27; CBO

56 RICHARD SCOTT CARNELL 57

1996, ix, 1, 9–10). GSEs also derive great advantage from the capital market participants’ perception that the federal government implicitly backs each GSE, which enables GSEs to borrow large sums at rates below those available to even the most creditworthy fully private borrowers (CBO 1996, 7, 9–11, 14–16; Treasury 1996, 27–28). This perceived implicit guarantee is the GSEs’ most important and most distinctive characteristic (CBO 1996, ix). Congress chartered Fannie Mae and Freddie Mac to develop and improve the secondary market for residential mortgage loans. To help the two GSEs carry out that mission, federal statutes confer special benefits:

• authorizing the Treasury to loan each firm up to $2.25 billion (12 U.S. Code §§ 1455(c), 1719(c)) • exempting each firm from most state and local taxes (§§ 1452(e), 1723a(c)(2)) • exempting each firm from registering its securities with the Se- curities and Exchange Commission (§§ 1455(g), 1723c, 15 U.S. Code § 77c(a)(2)) • imposing no limits on investments of federally chartered depository institutions’ investments in each firm’s securities (12 U.S. Code §§ 24 (Seventh), 1464(c)(1)(D)-(F), 1757(7)(E)) • making each firm’s debt securities and mortgage-backed securities eligible for open-market purchase by the Federal Reserve banks (§ 355(2)), eligible collateral for deposits of public funds and for advances from Federal Reserve and Federal Home Loan banks (§§ 347, 1430(a)(2), 1452(g), 1723c), and lawful investments for fiduciaries (15 U.S. Code § 77r-1(a); 12 U.S. Code §§ 1452(g), 1455(e)) • permitting each firm to issue and transfer securities through the Federal Reserve’s electronic book-entry system, the system used for issuing and transferring U.S. Treasury securities (12 U.S. Code §§ 1452(d), 1723a(g)).

Some court decisions—Mendrala v. Crown Mortgage Co., 955 F.2d 1132 (7th Cir 1992), and McCauley v. Thygerson, 732 F.2d 978 (DC Cir 1984)3 —have also given Freddie Mac the benefit of two judicial limitations on the federal government’s liability: sovereign immunity and the analogous rule prohibiting claimants from using promis- sory estoppel offensively against the government. These various specific benefits bring us to the most important benefit of all: the market participants’ belief that the government 58 DEPOSIT INSURANCE AND SPONSORSHIP implicitly backs Fannie Mae and Freddie Mac (Stigum 1990, 357– 58),4 notwithstanding disclaimers by statute (12 U.S. Code §§ 4501(4), 4503)5 and in each firm’s securities (§§ 1455(h)(1), 1719(b), (d)–(e)).6 In the words of one observer, GSEs “are regarded by most people who lend them money as the government in disguise” (Stigum 1990, 358). This perceived implicit backing enables the two GSEs to borrow more than $800 billion7 at rates only slightly above those on Treasury securities (Treasury 1996, 27–28; see generally GAO 1990, 83–88). It also increases the value of having Fannie or Freddie guarantee mortgage-backed securities: investors will accept lower returns on such securities than on mortgage-backed securities lacking such guarantees (GAO 1996, 17). The two firms have nearly $1.2 trillion in net outstanding mortgage-backed securities (in addition to the nearly $500 billion in mortgage-backed securities in their portfolios).8 Four other entities fit the above definition of government-sponsored enterprise: the Federal Home Loan Bank System, the Farm Credit System, the Federal Agricultural Mortgage Corporation (Farmer Mac), and the Student Loan Marketing Association (Sallie Mae).9 But Fannie and Freddie dwarf all but the Home Loan Bank System (see figure 4–1). They stick out like skyscrapers in a potato patch. Moreover, Fannie and Freddie may find other GSEs meager company qualitatively as well as quantitatively. Sallie Mae is being fully privatized; it will relinquish any claim to government sponsorship (20 U.S. Code § 1087-3).10 Farmer Mac has never made a success of its basic business (“Interest Rate Farming” 1998). The Home Loan Bank System has long outlived its original purpose and, in tandem with its captive regulator, is scrambling to justify its continued existence (Carnell 1998). Fannie and Freddie seem to regard the system as an eccentric and slightly disreputable cousin whose an- tics pose a nagging threat to the family honor. On balance, associat- ing with the other four GSEs offers Fannie and Freddie little comfort or cover.

Characterizing Federally Insured Depository Institutions as GSEs

But what if Fannie and Freddie could define themselves as part of a much larger set of financial institutions—not a set of two or three or six but a set numbering in the thousands? Not surprisingly, Fannie and Freddie have increasingly sought to do just that by attempting to characterize depository institutions as government-sponsored RICHARD SCOTT CARNELL 59

FIGURE 4–1 GSE Obligations (dollars)

700 Debt Net asset-backed securities outstanding 600

500

400

300

200

100

0 Fannie Freddie FHLBS FCS Farmer Sallie

enterprises. Robert Van Order, chief economist of Freddie Mac, cat- egorically stated, “Banks are GSEs” (see chapter 3 of this volume).11 Adolfo Marzol (1999, 1–2), executive vice president and chief credit officer of Fannie Mae, likened Fannie to national banks:

[National banks] were created by government charter to ensure the flow of credit to all regions of the country. Fannie Mae was created to ensure the flow of mortgage credit to lenders in all regions of the country. Today, both national banks and Fannie Mae operate under federal charter and are important fixtures of a healthy financial market.

Marzol acknowledged that “we do have a government charter,” but he stressed that “so do Norwest, Bank of America and thousands of national banks and thrifts” (p. 2). He depicted Fannie as a federal 60 DEPOSIT INSURANCE AND SPONSORSHIP instrumentality like those federally chartered depository institutions, only more efficient.12 Similarly Fannie’s CEO, Franklin D. Raines, stressed that FDIC-insured depository institutions “also enjoy special benefits,” including reduced borrowing costs (Raines 1999a; see also Raines 1999b). These advocates suggest that the government gives depository institutions benefits comparable to, or even greater than, those provided to Fannie and Freddie—and that any greater success shown by the two GSEs reflects their greater efficiency.13 Classifying depository institutions as GSEs could expand the roster of GSEs far beyond the conventional list of six companies, which have aggregate assets of nearly $1.5 trillion. One could include all federally chartered FDIC-insured depository institutions: the 2,410 national banks, with aggregate assets of $3.2 trillion (FDIC 1999b, 18), and the 934 federal savings associations, with aggregate assets of $812 billion (Office of Thrift Supervision 1999, 1–2). But one need not stop there: one could look beyond federal charters to federal deposit insurance. Including all state-chartered FDIC-insured depository institutions would bring the total to 10,350 FDIC-insured depository institutions, with aggregate assets of $6.6 trillion (FDIC 1999b, 18). Including the 10,841 federally insured credit unions, which have aggregate assets of $407 billion (National Credit Union Ad- ministration 1999), would push the total to 21,256 federally insured depository institutions, with aggregate assets of nearly $7 trillion. In any event, amalgamating depository institutions with GSEs would yield a set so numerous, diverse, and asset-rich as to provide ample cover even for fast-growing financial gorillas (Ely 1999, 12). FDIC-insured depository institutions do share some important characteristics with Fannie Mae and Freddie Mac. Both types of entities function as financial intermediaries. Both operate in the residential mortgage market. Both are privately owned and profit-oriented. Both are federal instrumentalities. Both benefit from federal statutes that preempt some state laws (Moe and Stanton 1989, 325). Both reap valuable benefits from their relationship to the government, including the potential for lower borrowing costs. Both also face moral hazard arising from that same relationship. But these common characteristics should not obscure the pro- found ways in which depository institutions differ from Fannie and Freddie. To begin with, depository institutions as a group do a predominantly retail business. All but a few deal directly with individuals and small firms, at least to some extent. They structure themselves to compete for that business; they maintain more offices RICHARD SCOTT CARNELL 61 and larger staffs than a wholesale-oriented firm would. By contrast, Fannie and Freddie do a wholesale business, which enables them to operate with lower overhead than depository institutions. Hence the comparison is misplaced, and advocates like Marzol err in denying the existence and importance of subsidy and in then attribut- ing Fannie’s success solely to greater efficiency.14 In fact, as explained next, Fannie and Freddie enjoy a larger net subsidy than federally insured depository institutions.

The Richer Net Subsidy of Government Sponsorship

One might well expect federally insured depository institutions to derive a greater net subsidy15 from explicit deposit insurance than Fannie Mae and Freddie Mac derive from their perceived implicit federal backing. Deposit insurance carries the force of law,16 and customers value such insurance. The perceived backing of GSEs by contrast stands officially disclaimed and amounts to “conjectured guarantees and conjectures about conjectured guarantees” (chapter 3 of this volume). Yet the government’s perceived implicit backing of Fannie Mae and Freddie Mac tends—paradoxically—to impart a greater net subsidy than explicit deposit insurance. I see six structural reasons for this tendency.17

1. Federal deposit insurance applies only to depositors and only up to a specified dollar amount. By contrast the government’s perceived implicit backing of GSEs has no limits: it applies to all of a GSE’s liabilities, with no dollar ceiling. 2. The FDIC routinely uses receivership to extinguish the claims of failed depository institutions’ shareholders and to require such institutions’ uninsured depositors and nondeposit creditors to bear part of the institutions’ losses. But no credible, workable receivership mechanism exists for Fannie and Freddie—an omission reinforcing the privileged position of the two firms’ creditors. 3. Federal deposit insurance involves strong indirect cross- guarantees among insured depository institutions—a product of safeguards designed to ensure that depository institutions, rather than the taxpayers, bear any losses incurred in protecting insured depositors. No similar cross-guarantees reduce the government’s risk exposure to Fannie and Freddie. 62 DEPOSIT INSURANCE AND SPONSORSHIP

Indeed the two GSEs do not even cross-guarantee each other. 4. To a much greater degree than depository institutions, Fannie and Freddie reap the benefits of special company-specific laws and avoid the discipline of generic law. 5. Banking law affords depository institutions little protection against competition. Restrictions against geographic expansion have greatly eroded, and federal and state regulators routinely issue depository institution charters to persons having the requisite qualifications. If depository institutions do receive a net federal subsidy, they should generally face enough actual or potential competition to compel them to transmit the subsidy to their customers. But Fannie and Freddie enjoy significant protection against competition. Their government sponsorship reduces their borrowing costs and increases the value of their guarantees to such an extent that fully private firms have difficulty competing with them. GSE charters are available only through an act of Congress. By impeding competition with Fannie and Freddie, these constraints on entry increase the potential for the two GSEs’ government benefits to end up in the hands of their shareholders rather than their customers. 6. The government does not require Fannie and Freddie to pay any explicit fee for their government sponsorship or to provide public benefits that impose significant costs on their shareholders.

No Limitation. The government’s perceived implicit backing of GSEs has no limit. Indeed it tends to defy quantitative limitation, and it applies to all of a GSE’s liabilities. How can the government explicitly limit a contingent liability whose existence it officially denies? Under the circumstances a quantitative limit on an implicit guarantee would be a contradiction in terms. Thus no GSE analogue exists—or realistically can exist—to three key statutes limiting the FDIC’s exposure. The first limits deposit insurance coverage to $100,000 (12 U.S. Code § 1821(a)(1)(B)). The second generally permits the FDIC to protect a depository institution’s uninsured depositors and nondeposit creditors only if doing so is the “least costly to the deposit insurance fund of all possible methods” for meeting the FDIC’s obligation to insured depositors (that is, the least-cost resolution method; § 1823(c)(4)). The third gives a failed institution’s depositors and the FDIC priority over RICHARD SCOTT CARNELL 63 nondeposit creditors (§ 1821(d)(11)(A)(ii)).18 Each of these statutes presupposes that a failed institution’s general creditors may get nothing and commonly will suffer some loss. By contrast capital market participants expect the government to protect all of a GSE’s creditors, including its general creditors (Stigum 1990, 42). The least-cost resolution statute illustrates how one can limit an explicit guarantee not only quantitatively but through process constraints. The statute contains a narrow systemic-risk exception (12 U.S. Code § 1823(c)(4)(G)) “for those rare instances in which the failure of an institution could threaten the entire financial system” (Con- gress 1991, 45). The exception becomes an option only if recommended to the secretary of the Treasury by two-thirds majori- ties of both the Federal Reserve Board and the FDIC board of directors. The secretary of the Treasury can make the exception only if the secretary determines “in consultation with the President” that least-cost resolution of a given institution “would have serious adverse effects on economic conditions or financial stability.” The secretary must document the determination. The General Accounting Office must review and report on the exception, including the potential for it to increase moral hazard. To recoup the additional cost of deviating from least-cost resolution, the FDIC must levy a special assessment on insured depository institutions. Congress designed these rules to promote accountability and make the process sufficiently unpleasant that systemic-risk exceptions would be granted rarely (if at all) and never lightly (Carnell 1993, 368). The constraints on granting exceptions underscore the difference between the government’s carefully circumscribed support for FDIC-insured depository institutions and its hazier and more open-ended sponsorship of GSEs. Establishing or expanding a GSE can be likened to making a major purchase on the black market. If all goes well, a buyer not only obtains the desired product or service but saves money and avoids red tape. If anything goes wrong, however, he will probably not wish to bring the transaction—or himself—to the attention of the police or the courts. At least as a practical matter, the buyer may find himself bereft of legal remedies against whatever injuries he may suffer—whether from breach of contract, fraud, robbery, or mayhem. By operating outside the law, a person may effectively relinquish much of the law’s protection. Explicit deposit insurance parallels transactions in the lawful market. The government establishes and publicizes the $100,000 limit 64 DEPOSIT INSURANCE AND SPONSORSHIP on its liability. It also requires least-cost resolution lest the FDIC un- wisely use the deposit insurance fund to protect a failed institution’s uninsured depositors or nondeposit creditors. If disgruntled stakeholders complain, the FDIC can simply point to the letter and policy of the law. Stakeholders will have difficulty substantiating why they should receive any more than what the law allows. Implicit backing of GSEs parallels black-market transactions. If all goes well, the government achieves a given policy objective while avoiding such constraints as scarcity of appropriated funds, political resistance to overt subsidies, and bureaucratic torpor. But if something goes wrong, the government may well have less effective limits on its risk-exposure than it would under explicit statutory rules such as those applicable to the FDIC. The government could in theory adhere to the letter of the law, disclaim any implicit guarantee, and confound the expectations of capital market participants. But that approach risks being seen as a sharp practice lacking legitimacy. Moreover, in deciding whether to honor a perceived implicit governmental guarantee, policymakers face additional pressures if a financial crisis looms. However much policymakers may believe that the government is not and should not be liable, they must confront the risk that seriously upsetting established expectations might create contagious uncertainty about the government’s willingness to meet other expectations. A crisis is thus a particularly inopportune time for attempting to reeducate, or bargain with, market participants about the scope of the government’s undertakings.

No Credible, Workable Receivership Mechanism. Unlike FDIC- insured depository institutions, Fannie and Freddie face no credible, workable receivership mechanism. Depository institution receivership laws facilitate rapid and relatively efficient resolution of claims against a failed or failing institution (see generally Kaufman 1994, 138–39). Working with Congress, the FDIC and others have developed an array of mechanisms for taking control of such an institution, cutting off contingent claims, giving insured depositors ready access to their money, and preserving any going-concern value—see, for example, 12 U.S. Code §§ 1821(d)–(f), (i), (n), 1823(c)– (d). The FDIC can, for example, use a pass-through receivership to sell a troubled institution as a going concern to new purchasers able to resolve its capital and other deficiencies. If an institution is insolvent, its shareholders lose their ownership interest, its general creditors normally incur a partial or total loss on their claims, and its RICHARD SCOTT CARNELL 65 uninsured depositors often incur some loss (§ 1821(d)(11)(A), (i)(2); see generally FDIC 1999a). Even as Congress has given the FDIC new receivership tools, it has given regulators new grounds for placing FDIC-insured depository institutions in receivership (and strengthened incentives to so). The Federal Deposit Insurance Act now contains more than fifteen grounds for receivership (12 U.S. Code § 1821(c)(5)).19 Some grounds have long existed, such as insolvency, inability to pay creditors’ claims as they become due, dissipating assets, and concealing records. Congress added other grounds in 1989 and 1991, such as having no reasonable prospect of becoming adequately capitalized; being undercapitalized and failing to submit an acceptable capital restoration plan; and being critically undercapitalized. These new grounds typically facilitate earlier action by regulators, aimed at taking control of an institution before its problems grow even worse. More- over, the prompt corrective action statute creates a presumption in favor of placing certain critically undercapitalized institutions in con- servatorship or receivership (§ 1831o(h)(3)). No comparable receivership regime applies to Fannie and Freddie. Neither their charters nor the 1992 statute creating the Of- fice of Federal Housing Enterprise Oversight (OFHEO) contain any provision for receivership, and the bankruptcy code does not permit a federal instrumentality to become a debtor in a bankruptcy proceeding.20 The lack of a receivership statute reinforces the market perception that the government would assure full payment to Fannie’s and Freddie’s creditors. After all, even if policymakers wanted to write down creditors’ claims or require creditors to accept stock in partial satisfaction of those claims, how could policymakers effectuate such a reorganization without additional legislation? Freddie has suggested that a federal district court “could appoint a receiver for Freddie Mac under common law practice” (GAO 1990, 93n), but such an approach would entail considerable legal and practical uncertainty. The 1992 statute authorizes OFHEO to appoint only a conservator (see 12 U.S. Code §§ 4616(b)(6), 4617(a)(1), 4619(a)(1)). The conservator would generally have all the powers of the GSE’s shareholders, directors, and officers (§ 4620(a)). The conservator could thus dilute existing shareholders’ interests by authorizing the issuance of new stock but could not impose a reorganization on creditors. Likewise, although OFHEO itself could require a significantly or critically undercapitalized GSE to “acquire new capital,” reduce its obligations or assets, or terminate or curtail 66 DEPOSIT INSURANCE AND SPONSORSHIP any excessively risky activity (§ 4616(b)(1)–(4)), the statute does not authorize OFHEO to impair creditors’ claims.21 Thus the omission of receivership seems to signal that a GSE’s creditors would have nothing to fear from the law, no matter how egregious the GSE’s problems. In any event the omission holds the potential for considerable forbearance.

No Cross-Guarantee Protection for Government. The Federal De- posit Insurance Act contains safeguards designed to ensure that depository institutions—rather than the taxpayers—bear any losses incurred in protecting failed institutions’ insured depositors. First, the FDIC has a duty to manage each deposit insurance fund so as to maintain at least $1.25 in reserves for each $100 of estimated insured deposits (12 U.S. Code § 1817(b)(2)(A)(i), (iv)). Second, if the reserve ratio falls below that target, the FDIC must set risk-based premiums high enough to raise an average of some $0.23 per $100 of domestic deposits (§ 1817(b)(2)(E))—a historically high rate. Third, the law imposes no upper limit on premium rates and rate increases (see § 1817(b)(2)(A)). Fourth, the law sets a fifteen-year deadline for re- capitalizing a depleted fund (§ 1817(b)(3)). This system involves strong indirect cross-guarantees among insured depository institutions. It in effect gives the relevant FDIC fund a call on the capital and earnings of all institutions insured by that fund.22 Though lacking the four specific safeguards just enumerated, the Farm Credit Act establishes the Farm Credit Insurance Corpora- tion to ensure the timely payment of principal and interest on Farm Credit System banks’ obligations, using a fund built up through premiums from such banks (12 U.S. Code §§ 2277a-1, -4, -9). No equivalent protections exist in the government’s sponsorship of Fannie Mae and Freddie Mac. No other entity cross- guarantees them, nor do they cross-guarantee any other entity—not even each other.

Regulation under Company-Specific Law, Not Generic Law. To a much larger degree than depository institutions, Fannie and Freddie reap the benefits of special company-specific laws and avoid the discipline of generic law. Moe and Stanton (1989, 324) have explained how generic law promotes accountability. Without such law the government tends to deal ad hoc with “each institution and set of circumstances” in ways favorable to the narrow interests of the institution’s insiders and clientele. Generic law by contrast RICHARD SCOTT CARNELL 67

can help shift the political debate to questions about whether exceptions to general rules are warranted. In a debate over exceptions, advantage tends to shift to the agencies and the oversight committees in Congress . . . most concerned with overall government management and away from more particular interest group constituencies.

Freddie Mac’s charter contains a striking example of company- specific legislation: a provision easily construed as immunizing Freddie from a wide range of general legislation (12 U.S. Code § 1456(a)):

All rights and remedies of the Corporation . . . shall be im- mune from impairment, limitation, or restriction by or under (1) any law (except laws enacted by the Congress expressly in limitation of this sentence) which becomes effective after the acquisition by the Corporation of the subject or property on, under, or with respect to which such right or remedy arises or exists or would so arise or exist in the absence of such law, or (2) any administrative or other action which becomes effective after such acquisition.23

This immunity statute expressly excludes from its reach only “laws enacted by the Congress expressly in limitation of” the immunity statute itself). During the 1990s the government improved its safety-and- soundness regulation of Fannie Mae and Freddie Mac. The Federal Housing Enterprises Financial Safety and Soundness Act of 1992 created a new safety-and-soundness regulator, OFHEO (12 U.S. Code §§ 4511–4513); required a new set of capital standards (§§ 4611–4612); and included regulatory enforcement authority (§§ 4631–4636) and prompt corrective action (§§ 4614–4619, 4622). But the 1992 act, though modeled on depository institution regulation, tends to begrudge OFHEO authority possessed by depository institution regulators. Such regulators have broad authority to prescribe capital standards, including authority to impose new standards or toughen existing standards in light of experience (12 U.S. Code §§ 1831o(c)(1), 3907(a)). OFHEO by contrast faces major constraints on the form and structure of capital standards (§§ 4611–4612). Its authority to take enforcement action against Fannie and Freddie (§§ 4631–4636) is conspicuously weaker than that of its depository 68 DEPOSIT INSURANCE AND SPONSORSHIP institution counterparts (§ 1818). Similar weaknesses exist in the statutes governing prompt corrective action (compare § 1831o with §§ 4614–4619, 4622). For example, an undercapitalized depository institution cannot increase its total assets unless (1) the institution has an acceptable capital restoration plan, (2) the asset growth com- ports with the plan, and (3) the institution’s capital ratio increases at a rate sufficient to enable the institution to become adequately capitalized within a reasonable time (§ 1831o(e)(3)).24 Yet no statute bars Fannie and Freddie from continuing to grow while undercapitalized even if they have no capital restoration plan or if the growth conflicts with such a plan (§ 4615). The statute regarding prompt corrective action authorizes growth restrictions only against a significantly or critically undercapitalized GSE, and makes such sanc- tions purely discretionary (§§ 4616(b)(2), 4617(b), (c)(2)). These and other triumphs of hobnailed lobbying have led Stanton (1999, 14) to describe OFHEO’s statutory authority as “a sort of parody of the authority of the federal bank regulators.”

Protection from Competition. Constraints on chartering GSEs increase the potential for Fannie and Freddie’s government benefits to end up in the hands of their shareholders rather than their customers. Banking law affords depository institutions scant protection against competition. Restrictions against geographic expansion have greatly eroded, and federal and state regulators routinely issue depository institution charters to persons having the requisite qualifications. Insofar as depository institutions may receive a net federal subsidy, they should also face enough actual or potential competition to compel them to transmit that subsidy to their customers. Fannie and Freddie by contrast enjoy significant protection against competition. Their government sponsorship reduces their borrowing costs and increases the value of their guarantees to such an extent that fully private firms have difficulty competing with them. Moreover, one cannot apply for a GSE charter at OFHEO or any other regulatory agency. Only Congress can create a GSE. By impeding competition with Fannie and Freddie, these constraints on entry increase the potential for the two GSEs’ government benefits to end up in the hands of their shareholders rather than their customers.

No Explicit Fee or Costly Public Benefits. The government does not require Fannie and Freddie to pay any explicit fee for their RICHARD SCOTT CARNELL 69 government sponsorship or to provide public benefits that impose significant costs on their shareholders. Few people understand the value of government sponsorship to the GSEs—or the current and potential costs of that sponsorship to the government and the taxpayers. In popular usage subsidy connotes an explicit governmental payment and commonly fails to encompass implicit benefits. The GSEs use such blind spots to good advantage. Fannie can buffalo the unsophisticated by denying that it receives any subsidy. Insofar as these assertions are accepted, any public benefits provided by Fannie and Freddie may be regarded as a windfall. Federal budget law abets such misperceptions by failing to require the budget to recognize the costs of sponsoring GSEs (see generally Shadow Fi- nancial Regulatory Committee 1996). The public benefits required of Fannie and Freddie do not impose significant costs on their shareholders. During congressional deliberations over the 1992 act, Fannie and Freddie thwarted a proposed requirement that they fund affordable-housing grants equal to 20 percent of their dividend payments (Feldman and Koppell 1999, 13)—a requirement that would have made a real contribution toward expanding home ownership by the working poor.25 Instead Fannie and Freddie must comply with weak (but impressive- sounding) housing goals prescribed by the secretary of Housing and Urban Development: a low- and moderate-income housing goal, a special affordable housing goal, and a goal for central cities, rural areas, and other underserved areas (12 U.S. Code §§ 4562–4564). The goals are so weak that Fannie and Freddie meet them while actually doing less for affordable housing than depository institutions. Ac- cording to the U.S. Treasury (1996, 62; see generally 54–67), “Fannie Mae and Freddie Mac have a lower relative share of loans to targeted borrower groups than do most of the other mortgage market participants.” The Treasury noted that “the majority of affordable housing loans purchased by the GSEs meet standard underwriting criteria” and that it was “unclear” whether Fannie and Freddie needed government sponsorship to continue their current affordable housing efforts: “Knowing that would require more information on whether the GSEs’ affordable housing efforts generate returns significantly below those of comparable lines of business” (p. 76). When the Treasury requested such information of Fannie and Freddie for use in a congressionally mandated study, the two GSEs responded differently. Freddie replied that it “purchases most single- family and multifamily mortgages in support of affordable housing 70 DEPOSIT INSURANCE AND SPONSORSHIP through its standard mortgage purchase programs and under the same credit standards as its other mortgage purchases” (Treasury 1996, 76). The response suggests that affordable housing goals do not impose significant costs on Freddie’s shareholders: if “most” affordable housing loans meet Freddie’s usual credit standards, then they presumably provide something approximating a normal return. Fannie by contrast called the information proprietary and refused to provide it unless the Treasury signed a written agreement constraining the Trea- sury from making public use of the information, which would have defeated the purpose of obtaining the information (Zigas 1996, 4). Yet in using affordable housing loans to justify continued government sponsorship, Fannie faces a logical bind: if those loans provide anything approximating a normal return, then Fannie has incentives to make those loans in any event; and if those loans do not provide such a return, then why does Fannie refuse to say so and to share information about its experience? One can reasonably infer that Fannie with- held the information lest it undercut the firm’s own characterization of its affordable housing efforts. Fannie and Freddie use the affordable housing goals as policy bait for the unwary. The goals appear to yield meaningful public benefits and yet evidently impose no significant costs on the GSEs’ shareholders. The GSEs can satisfy the goals by doing business that they would have a profit motivation to do in any event. Ed Kane (1999, 10) has aptly summarized the result: “Incentive-conflicted GSE managers get to frame their obligations to taxpayers as affordable housing goals and to decide more or less on their own what other behaviors discharge the ethical obligation that their federal status creates for them.”

Conclusion

As financial institutions, government-sponsored enterprises naturally tend to have more in common with FDIC-insured depository institutions than with nonfinancial companies. But Fannie Mae and Freddie Mac differ significantly from depository institutions—not least in the extent to which the two GSEs’ federal sponsorship im- parts a greater net subsidy than explicit federal deposit insurance. To protect the taxpayers from the risks associated with insuring deposits, Congress has imposed rigorous safeguards on depository institutions, their regulators, and the FDIC as deposit insurer. Con- gress has limited the scope of deposit insurance. It has generally required failed depository institutions’ shareholders, nondeposit RICHARD SCOTT CARNELL 71 creditors, and uninsured depositors to absorb losses. It has built extensive cross-guarantees into each deposit insurance fund. Reforms implemented since the late 1980s have forced depository institutions to internalize more of the costs of their own risk-taking and have thus reduced the insurance funds’ uncompensated risk-bearing—to such an extent that the FDIC and many independent economists believe that well-capitalized depository institutions currently receive little or no net marginal subsidy.26 Moreover, insofar as depository institutions do receive such a subsidy, the pressure of competition should force them to transmit it to their customers. By contrast Fannie Mae and Freddie Mac enjoy a market perception of implicit government backing that has no dollar limit and applies to all their liabilities. That perception amounts to a $2 trillion wink at the letter of the law, a wink made with confidence drawn from the numerous special benefits conferred on Fannie and Freddie (including the statutes exempting their securities from registration and equating them with U.S. Treasury securities for various purposes) and from the lack of any credible receivership mechanism. But for Fannie and Freddie the deal is sweeter still. The government charges no fee for its sponsorship. It requires no payments to an insurance fund or affordable housing fund. It requires no direct or indirect cross-guarantee of anyone. Nor does it require public benefits that impose significant costs on Fannie and Freddie’s shareholders. All in all, Fannie and Freddie receive such formidable government-conferred advantages that no firm but a GSE can compete against them effectively. The lack of competition in turn increases the potential for Fannie and Freddie’s government benefits to end up in the hands of their shareholders rather than their customers. With the federal government giving Fannie Mae and Freddie Mac such valuable benefits and requiring so little in return, Fannie and Freddie understandably seek to deflect critical scrutiny. They assert or imply that FDIC-insured depository institutions receive a similar, or even greater, government subsidy—and attribute their own growth and profitability to greater efficiency. We should not let them obscure our vision.

Appendix Progress in Curtailing Subsidy to FDIC-Insured Depository Institutions

The government provides FDIC-insured depository institutions27 with three key benefits not available to financial institutions generally: 72 DEPOSIT INSURANCE AND SPONSORSHIP deposit insurance, access to the Federal Reserve discount window, and Fedwire daylight overdrafts. The government does not recover the full value of those benefits through explicit prices.28 But depository institutions can receive those benefits only by incurring substantial costs not imposed on nondepository financial institutions, including (1) assessments to cover interest payments on the so-called FICO (Financing Corporation) bonds, which funded part of the thrift cleanup; (2) forgone interest earnings on required reserves; and (3) most important, the cost of complying with banking regulations (Jones and Kolatch 1999, 8–9), including regulations regarding safety and soundness, consumer protection, entry and exit, and geographic and product expansion, as well as the cost of undergoing peri- odic safety and soundness, compliance, fiduciary, and information- system examinations (Madjid-Sadjadi and Vencill 1999, 9; see generally FFIEC 1992). One can usefully distinguish between gross and net subsidies to depository institutions. The gross subsidy represents the total value of the special benefits that the federal government provides. The net subsidy represents the difference between the gross subsidy and the offsetting costs depository institutions must incur. The net marginal subsidy represents the difference between the special benefits and the offsetting costs associated with acquiring an additional dollar of insured deposits. The FDIC and many independent economists (for example, Shadow Financial Regulatory Committee 1997; Shull and White 1998, 474–75) believe that well-capitalized depository institutions currently receive little or no net marginal subsidy. The FDIC has concluded that any such subsidy is competitively insignificant. “For well- capitalized banks,” Chairman Donna Tanoue has testified, “the evidence shows that if a net marginal funding advantage exists at all, it is very small” (Senate 1998, 412). The slight size of any net marginal subsidy reflects progress made since the late 1980s in protecting the deposit insurance funds and reducing the taxpayers’ potential exposure. Bank regulators prescribed risk-based capital standards (for example, 12 C.F.R. part 3, app. A) and intensified their scrutiny of banks’ asset quality. The Federal Reserve tightened limits on daylight overdrafts and then began charging for them (Federal Reserve Board 1989, 1992). Con- gress enacted an extensive set of reforms, chiefly in the Federal De- posit Insurance Corporation Improvement Act of 1991 (commonly known as FDICIA), which sought to bring the incentives of deposi- RICHARD SCOTT CARNELL 73 tory institutions’ owners, managers, and regulators more closely into line with the interests of the insurance funds (see generally Carnell 1993). These reforms include

• risk-based deposit-insurance premiums (12 U.S. Code § 1817(b)(1)); • prompt corrective action, under which regulators must impose increasingly stringent restrictions and requirements on a depository institution as its capital declines below required levels—in the interest of resolving the institution’s problems at no cost or minimal cost to the FDIC (§ 1831o); • least-cost resolution, under which the FDIC can generally protect a depository institution’s uninsured depositors and nondeposit creditors only if doing so is the “least costly to the deposit insurance fund of all possible methods” for meeting the FDIC’s obligation to insured depositors (§ 1823(c)(4)(A)); • other measures designed to curtail too-big-to-fail practices, including • limiting the use of the Federal Reserve discount window to prop up sick institutions (§ 347b(b)); • limiting each depository institution’s exposure to other depository institutions (§ 371b-2); • ensuring that netting provisions in clearing and settlement contracts are enforceable and thereby reducing the risk that a large institution’s failure would disrupt the payment system (§§ 4401-4407); and • giving a failed institution’s depositors and the FDIC priority over nondeposit creditors (ibid. § 1821(d)(11)(A)(ii)).

These and other reforms have forced depository institutions to internalize more of the costs of their own risk-taking and have thus reduced the insurance funds’ uncompensated risk-bearing (FDIC Chairman Helfer in House 1997, 209–14; see also Hovakimian and Kane 2000). The key post-FDICIA trends were consistent with progress toward those objectives: bank profitability soared; bank capital rose to the highest levels in half a century; the FDIC’s problem bank list shrank; and bank failures fell to a fraction of their level at the beginning of the decade. The Bank Insurance Fund’s reserve ratio, which had fallen to negative 0.36 percent in 1991, now stands at 1.40 percent (FDIC 1999b, 17),29 surpassing the statutory target of 1.25 percent (12 U.S. Code § 1817(b)(2)(iv)). 74 DEPOSIT INSURANCE AND SPONSORSHIP

Notes

1. Critics acknowledge that Fannie and Freddie have played an important role in developing the secondary mortgage market but contend that the two government-sponsored enterprises now provide a poor return on their federal benefits. Critics make four key arguments. First, deep, liquid, and highly efficient secondary markets would exist even without continued government sponsorship of Fannie and Freddie. Such sponsorship benefits the public chiefly by lowering mortgage interest rates somewhat, insofar as the two GSEs pass through the lower borrowing costs made possible by such sponsorship (CBO 1996, 18–19, 26, 31; Treasury 1996, 69–70). The General Accounting Office (GAO 1996, 56–57) estimated that Fannie and Freddie’s government sponsorship lowered fixed-rate, single- family mortgage interest rates by 0.15–0.35 percentage point. Second, such a general diminution in mortgage rates does little to increase homeownership at the margin (for example, by the lower middle class, the working poor, or members of certain minority groups). Housing markets capitalize lower mortgage rates into higher housing prices so that lower rates benefit existing homeowners at the expense of those seeking to become homeowners. Nor do lower mortgage rates benefit persons for whom a lack of a down payment represents the greatest barrier to homeownership (CBO 1996, 29–31; Calomiris 2001; Savage 1999, 1, 5–6). For example, a Census Bureau study concluded that lowering mortgage interest rates by less than three percentage points from the 8.67 percent rate prevailing in 1995 “had no significant effect on the number of renters who would have qualified for a mortgage on a modestly priced house” (Savage 1999, 5). Third, Fannie and Freddie are a “spongy conduit” for federal subsidies to housing, “soaking up nearly $1 for every $2 delivered” to homeowners (CBO 1996, xiv; see also 19–21, 24; Treasury 1996, 34–35; Kane 1999, 8–9). That is, Fannie and Freddie absorb roughly $1 of each $3 in government benefits that they receive. Fourth, the government could provide greater public benefits at lower cost by restructuring its support for housing—for example., by directly subsidizing homebuyers or by instituting competitive bidding for the government sponsorship currently accorded to Fannie and Freddie (Calomiris 2001; CBO 1996, 31–32, 39–41; see also Canner and Passmore 1995). 2. GSEs are not government agencies (see 5 U.S. Code § 105, defining executive agency as “an Executive Department, a Government corporation, and an independent establishment”), even though market participants commonly refer to GSEs’ securities as “agency securities” (Stigum 1990, 41–42). Nor are GSEs “government corporations” (see 5 U.S. Code § 103, defining RICHARD SCOTT CARNELL 75 government corporations as those “owned or controlled” by the federal government). For a scholarly analysis of government corporations and GSEs, see Froomkin 1995. 3. Both Mendrala and McCauley arose when Freddie Mac’s charter gave Freddie “all immunities and priorities, including . . . all immunities and priorities . . . to which it would be entitled if it were the United States or . . . an unincorporated agency of the United States” (12 U.S. Code § 1456(a) (1988)). The omission of this language in 1989 (Financial Institutions Reform, Recov- ery, and Enforcement Act of 1989, § 731(j)(1), 103 Stat. 435), as part of a nominal privatization of Freddie Mac, should undercut both decisions. 4. Some statutes relating to Fannie and Freddie obliquely abet this perception. The Federal National Mortgage Association Charter Act calls Fannie “a Government-sponsored private corporation” (12 U.S. Code § 1716b). The Secondary Mortgage Market Enhancement Act of 1984 generally permits fiduciaries to invest in securities issued or guaranteed by Fannie or Freddie “to the same extent” as in U.S. Treasury securities and specifies that for purposes of any applicable investment limits Fannie and Freddie securities “shall be considered to be obligations issued by the United States” (15 U.S. Code. § 77r-1(a)(1)–(2)). 5. The phrasing of these statutory disclaimers may help explain why they fail to impress capital market participants. They do not literally disclaim implicit backing, much less attempt to signal that market participants err in per- ceiving such backing. The first disclaimer is simply a congressional finding that “neither the [two] enterprises . . . , nor any securities or obligations issued by the enterprises . . . , are backed by the full faith and credit of the United States” (12 U.S. Code § 4501(4)). The second and third disclaimers adopt a studiously neutral tone and merely preclude construing the Federal Housing Enterprises Financial Safety and Soundness Act of 1992 as creating or evi- dencing government backing: the act “may not be construed as obligating the Federal Government . . . to honor, reimburse, or otherwise guarantee any obligation or liability” of Fannie, Freddie, or the Federal Home Loan banks or “as implying that any such [entity is] backed by the full faith and credit of the United States” (§ 4503). One may question the utility of such narrow, neutral language, especially given the strength and long standing of the market perception. “Indeed, the disclaimer itself hints at a special federal relationship; completely private firms do not need to disclaim federal backing because no one believes such backing exists” (Moe and Stanton 1989, 323). 6. Congress requires each firm’s securities to include “appropriate language . . . clearly indicating” that the securities “are not guaranteed by the United States and do not constitute a debt or obligation of the United States or of any agency or instrumentality thereof” other than the GSE in question (12 U.S. Code §§ 1455(h)(1), 1719(b), (d)–(e)). 76 DEPOSIT INSURANCE AND SPONSORSHIP

7. As of June 30, 1999, Fannie Mae had $500 billion in net total debt securities (Fannie Mae 1999a, 1), and Freddie Mac had $314 billion (Freddie Mac 1999, 2). 8. As of June 30, 1999, Fannie Mae had $911 billion in outstanding mortgage-backed securities, of which it held $250 billion in portfolio, leaving it (after rounding) with $662 billion in net outstanding mortgage-backed securities (Fannie Mae 1999a, 2). Freddie Mac had $718 billion in outstanding mortgage-backed securities, of which it held $202 billion in portfolio, leaving it with $516 billion in net outstanding mortgage-backed securities (Freddie Mac 1999, 2). 9. Specifically, each of these four entities is privately owned and profit- oriented, has a congressional charter and a public mission, receives a characteristic set of statutory privileges, and benefits from market participants’ perception of implicit government backing. Statutory benefits vary somewhat from one GSE to another. For example, the Federal Agricultural Mortgage Corporation, unlike the other GSEs, has no exemption from securities registration (12 U.S. Code § 2279aa-12(a)(1)). Similarly, although state and local governments can tax interest on most GSEs’ debt securities, they cannot tax interest on securities issued by the Federal Home Loan Bank System or Farm Credit System (§§ 1433, 2023, 2077, 2098, 2134, 2278b-10(b)). 10. The Student Loan Marketing Association Reorganization Act of 1996 requires Sallie Mae to form a holding company and transfer its business to that company, which has no government sponsorship (20 U.S. Code § 1087- 3). The old GSE is gradually winding up its business, and must dissolve itself by September 30, 2008. The Sallie Mae data in figure 4–1 are for the old GSE. 11. Van Order (2001) contended that when assessing Fannie and Freddie’s government sponsorship, one should view the “relevant industry not [as] the secondary market, but rather the mortgage market as a whole.” Within that single industry, two “dueling charters” exist—one for depository institutions and one for GSEs proper: Both involve imbedded options, and both have offsetting restrictions on lines of business, capital regulation and mission regulation. Both types of charter promote different ways of connecting mortgage markets with the financial markets. What matters is the pluses and minuses of the two charters. Competition is largely about leveraging these charters and the business lines allowed by them. 12. Assailing “the myth that Fannie Mae thrives because of an implied government guarantee that lowers our cost of funds,” Marzol (1999, 2) RICHARD SCOTT CARNELL 77 pointed to the disclaimer in Fannie’s debt securities, asserted that the purchasers of those securities understood the absence of a government guarantee, and stated that “our cost of funds is significantly higher than the U.S. Treasury’s, and national banks earn far more from the funds that they lend than we do.” According to Marzol, “Fannie Mae is profitable because we have relatively low overhead and only 4,000 employees.” 13. Fannie Mae (1999b) later asserted that it “operates with entirely private capital” and that the term GSE is “a misnomer which fails to recognize that [such a company’s] activities . . . are entirely supported by the revenue of the company and are not in any way guaranteed by the federal government.” Fannie characterized national banks and federal thrifts as “government-guaranteed enterprises” whose “business activities . . . are derived largely from the ability to hold consumer deposit accounts which carry the explicit guarantee of U.S. taxpayers in the event of default.” 14. Similar emphasis on Fannie’s efficiency appears in Raines 1999b: Our operating costs are the lowest in the industry, especially when you compare our business model to money center or regional banks. Banks begin at portfolio margins almost four times ours, from 411 to 430 basis points. But their cost structures eat up almost 300 of those basis points, which reduces their return on assets to 117 and 143 basis points. On the other hand, Fannie Mae begins with a much lower initial margin and our business model gives up only 40 basis points to non-interest expense. 15. The gross subsidy represents the total value of the special benefits provided by the federal government—benefits not available to businesses generally or even financial institutions generally. The net subsidy represents the difference between the gross subsidy and the offsetting costs that the entity must incur as a depository institution or GSE—costs not imposed on financial institutions generally. 16. Although no statute expressly places the government’s full faith and credit behind federal deposit insurance, the Justice Department’s Office of Legal Counsel has ruled that “when Congress authorizes a federal agency or officer to incur obligations, those obligations are supported by the full faith and credit of the United States, unless the authorizing statute specifically provides otherwise” (Justice 1987, 6: 264). Moreover, the Federal De- posit Insurance Act requires FDIC-insured savings associations, and permits FDIC-insured banks, to display a logo including a “statement that insured deposits are backed by the full faith and credit of the United States Govern- ment” (12 U.S. Code § 1828(a)). 17. Additional reasons may exist, as well as some potential overlap among the six reasons listed here. 78 DEPOSIT INSURANCE AND SPONSORSHIP

18. Depositor preference gives depositors priority over general creditors: depositors must get paid in full before general creditors can get paid at all. When the FDIC pays a failed institution’s insured depositors, it takes over their claims against the institution (in legal terms, the FDIC becomes subro- gated to those claims; 12 U.S. Code § 1821(g)(1)). Depositor preference facili- tates resolving a failed institution by cutting off contingent liabilities (for example, potential lawsuits). When protecting such an institution’s insured depositors, the FDIC typically incurs some loss, which arises because the institution’s deposits exceed the value of the assets remaining after satisfy- ing secured claims and senior unsecured claims. Under such circumstances no assets remain to pay general claims, including contingent claims. 19. This statute, section 11(c)(5) of the Federal Deposit Insurance Act, contains thirteen subparagraphs, each setting forth a separate ground for receivership. Moreover, each clause of subparagraphs (K) and (L) sets forth a conceptually distinct ground. However one counts the number of grounds, they are plentiful. 20. The bankruptcy code permits only a person to become a debtor (11 U.S. Code § 109(a)), and generally defines person so as to exclude a governmental unit (§ 101(41)). A governmental unit in turn includes any “instrumentality of the United States” (§ 101(27)). Thus Stanton (1991, 206) concludes that “a set of interlocking definitions” in the bankruptcy code precludes a bankruptcy filing by a GSE by “suggest[ing] that a federal instrumentality, except when expressly authorized by federal law, cannot claim bankruptcy under the code.” Similarly the General Accounting Office (1990, 92–93) concluded that the bankruptcy code “does not appear to apply either to Fannie Mae or to Freddie Mac.” 21. A conservator may, however, “avoid any security interest taken by a creditor with the intent to hinder, delay, or defraud” the GSE or its creditors (12 U.S. Code § 4620(b)). 22. The Federal Credit Union Act achieves a similar result more directly. Each federally insured credit union must maintain in the National Credit Union Share Insurance Fund a deposit equal to 1 percent of its own deposits (12 U.S. Code § 1782(c)(1)(A)(1)). Such deposits constitute reserves of the fund, as do any interest earnings and insurance premiums. The National Credit Union Administration sets the fund’s normal operating level of reserves at 1.2–1.5 percent of insured deposits (§ 1782(h)(4)). If the fund’s reserves fall below 1 percent, each credit union must write down its deposit in the fund and by a specified date replenish that deposit so that it once again equals 1 percent of the credit union’s own deposits (§ 1782(c)(1)(B)(iv); 12 C.F.R. § 741.4(c)). The fund’s reserve ratio has remained above 1 percent since the 1 percent deposit requirement became effective in 1985; that stabil- RICHARD SCOTT CARNELL 79 ity suggests that credit union deposit insurance has not posed significant risks to the taxpayers. Credit unions have paid an explicit premium only once since 1985. The de facto premium consists of forgone earnings on the 1 percent deposit, that is, the difference between what the credit union could have earned on the money and any dividends received from the fund. 23. This statute exhibits the heavy-handed legislative drafting characteristic of the thrift industry and its regulator, the Federal Home Loan Bank Board, at the height of their impunity. The board served until 1989 as Freddie Mac’s board of directors. 24. These rules seek to ensure that asset growth strengthens, rather than weakens, a capital-deficient institution. They also promote market discipline insofar as they compel the institution to raise new capital: the institution must persuade investors that it is viable and that its growth plans make sense. 25. “[T]oday’s critical housing priorities center not on the general operation of the nation’s credit markets but on the needs of certain borrowers that continue to find homeownership beyond their grasp.” The Department of Housing and Urban Development has “identifie[d] a lack of funds for downpayments and insufficient income—not credit availability—as the two main financial barriers to homeownership among the targeted borrower groups” (Treasury 1996, 75). 26. See appendix below. 27. In the interest of simplicity, the text does not deal specifically with federally insured credit unions, but the same basic principles that apply to FDIC-insured depository institutions hold true for such credit unions. Like FDIC-insured institutions, credit unions receive deposit insurance, with the same $100,000 limit as applies to the FDIC, and can obtain access to the Federal Reserve discount window, Fedwire daylight overdrafts, and Fed- eral Home Loan bank membership (12 U.S. Code §§ 461(b)(1)(A)(iv), (2), (7), 1422(12)(B), 1424(a)(1), 1787(k)(1)). (Unlike other depository institutions, credit unions enjoy exemption from the federal corporate income tax; 26 U.S. Code § 501(c)(14)(A); 12 U.S. Code § 1768.) Like FDIC-insured institutions, credit unions must also incur costs not imposed on other businesses. These costs include forgone earnings on the required 1 percent deposit in the National Credit Union Share Insurance Fund and the cost of complying with credit union regulations. Like the FDIC’s deposit insurance funds, the National Credit Union Share Insurance Fund involves strong indirect cross- guarantees among insured institutions (see note 22) and other safeguards designed to reduce risk to the taxpayers (for example, 12 U.S. Code §§ 1786, 1787, 1790d). 80 DEPOSIT INSURANCE AND SPONSORSHIP

28. Most FDIC-insured depository institutions pay no premiums for their deposit insurance. Under misguided legislation enacted in 1996, if a deposit insurance fund exceeds its reserve target, a well-capitalized institution pays no premiums unless regulators find it to have significant financial, operational, or compliance weaknesses (12 U.S. Code § 1817(b)(2)(A)(iii), (v)). This exemption from premiums applies to 9,709 of the 10,350 FDIC- insured institutions (94 percent), and those institutions hold some 97 percent of domestic deposits at FDIC-insured institutions (FDIC 1999b, 19). Likewise, discount window access—particularly the potential for a sol- vent institution to liquefy its assets during a financial emergency—has real value not reflected in any explicit charge (Walter 1998, 5). Fedwire daylight overdrafts incur charges at an annual rate of only 0.27 percent, far below short-term loan rates (ibid., 6–7). 29. The fund has $1.40 in reserves for each $100 of estimated insured deposits. These improvements, though greatly facilitated by macroeconomic pros- perity, also reflect the beneficial effects of the FDICIA-era reforms enumerated above. Those inclined to disagree should recall (1) that the improvements manifested themselves more rapidly than the economic recovery and (2) that the U.S. economy also prospered during much of the 1980s but with very different results for depository institutions and the federal deposit insurance funds.

References

Calomiris, Charles W. 2001. “An Economist’s Case for GSE Reform.” In Serv- ing Two Masters, Yet Out of Control: Fannie Mae and Freddie Mac, edited by Peter J. Wallison. Washington, D.C.: AEI Press. Canner, Glenn B., and Wayne Passmore. 1995. “Credit Risk and the Provi- sion of Mortgages to Lower-Income and Minority Homebuyers.” Federal Reserve Bulletin 81: 989–1016. Carnell, Richard Scott. 1993. “A Partial Antidote to Perverse Incentives: The FDIC Improvement Act of 1991.” Annual Review of Banking Law 12: 317-71. ———. 1998. “Twelve Banks in Search of a Purpose.” Remarks at conference, “The Future of Federal Home Loan Banks,” December 2, American Enterprise Institute, Washington, D.C. Available at . Comprehensive Deposit Insurance Reform and Taxpayer Protection Act of 1991. 102d Cong., 1st sess., S. Rept. 167. Ely, Bert. 1999. “Fannie Mae and Freddie Mac: What’s at Risk.” Paper presented at conference, “Fannie Mae and Freddie Mac: Public Purposes and RICHARD SCOTT CARNELL 81

Private Interests,” September 8, American Enterprise Institute, Washing- ton, D.C. Federal Deposit Insurance Corporation. 1999a. Historical Statistics on Bank- ing—Bank and Thrift Failures. Updated August 5, 1999. Found September 5, 1999, at . ———. 1999b. Quarterly Banking Profile—Second Quarter, 1999. Washington, D.C.: FDIC. Federal Deposit Insurance Corporation Improvement Act of 1991. U.S. Law 102- 242, 105 Stat. 2236-393 (1991). Federal Financial Institutions Examination Council. 1992. Study on Regula- tory Burden. Washington, D.C.: FFIEC. Federal Home Loan Mortgage Corporation. 1999. Supplement (August 13) to information statement of March 31, 1999. Found September 5, 1999, at . Federal Housing Enterprises Financial Safety and Soundness Act of 1992. U.S. Law 102-550, U.S. Statutes at Large 106: 3941–4012. Federal National Mortgage Association. 1999a. 1999 Second Quarter Earn- ings: Selected Financial Information. Found September 5, 1999, at . ———. 1999b. FM Watch Observer—Glossary of Terms. Found December 6, 1999, at . Federal Reserve Board. 1989. “Payments System Risk Reduction Program.” Federal Register 54: 26,094. ———. 1992. “Daylight Overdraft Pricing.” Federal Register 57: 47,084. Feldman, Ron, and Jonathan G. S. Koppell. 1999. “Congressional Oversight of Fannie Mae and Freddie Mac: Does Their GSE Structure Matter?” Pa- per presented at conference, “Fannie Mae and Freddie Mac: Public Pur- poses and Private Interests,” September 8, American Enterprise Institute, Washington, D.C. Financial Institutions Reform, Recovery, and Enforcement Act of 1989. U.S. Law 101-73, U.S. Statutes at Large 103: 183–553. Froomkin, A. Michael. 1995. “Reinventing the Government Corporation.” Illinois Law Review 543–634. Hovakimian, Armen, and Edward J. Kane. 2000. “Effectiveness of Capital Regulation at U.S. Commercial Banks, 1985 to 1994.” Journal of Finance 55: 451–68. “Interest Rate Farming.” 1998. Grant’s Interest Rate Observer, July 3, pp. 10–11. Jones, Kenneth, and Barry Kolatch. 1999. “The Federal Safety Net, Banking Subsidies, and Implications for Financial Modernization.” FDIC Banking Review 12 (1): 1–17. 82 DEPOSIT INSURANCE AND SPONSORSHIP

Kane, Edward J. 1999. “Housing Finance GSEs: Who Gets the Subsidy?” Paper presented at conference, “Fannie Mae and Freddie Mac: Public Purposes and Private Interests,” September 8, American Enterprise Insti- tute, Washington, D.C. Kaufman, George G. 1994. “Bank Contagion: A Review of the Theory and Evidence.” Journal of Financial Services Research : 123–50. Madjid-Sadjadi, Zagros, and C. Daniel Vencill. 1999. “Fact, Fiction and Fuzzy Logic: Is There a U.S. Commercial Bank ‘Safety Net Subsidy’? Evidence from Canadian Banking.” Unpublished manuscript. Marzol, Adolfo. 1999. “Fannie Mae’s Response.” Paper presented at conference, “Fannie Mae and Freddie Mac: Public Purposes and Private Inter- ests,” September 8, American Enterprise Institute, Washington, D.C. Moe, Ronald C., and Thomas H. Stanton. 1989. “Government-Sponsored Enterprises as Federal Instrumentalities: Reconciling Private Management with Public Accountability.” Public Administration Review 49: 321–29. National Credit Union Administration. 1999. “Credit Union Data—Call Report Data Files Under FOIA—Consolidated Balance Sheets, June 1999.” Found September 5, 1999, at . Office of Thrift Supervision. 1999. Thrift Industry Quarterly Highlights: Finan- cial Data for Eight Quarters Ending June 1999. Washington, D.C.: Office of Thrift Supervision. Raines, Franklin D. 1999a. Remarks at New York Financial Writers’ Associa- tion Annual Awards Dinner, June 17, New York. Found September 5, 1999, at . ———. 1999b. Remarks at Merrill Lynch Investor Conference, September 14. Found October 6, 1999, at Savage, Howard A. 1999. Who Could Afford to Buy a House in 1995? U.S. Cen- sus Bureau Current Housing Reports. H121/99-1. Washington, D.C.: Government Printing Office. Shadow Financial Regulatory Committee. 1996. Extending the Credit Reform Act to GSEs. Statement 131, February 12. Washington, D.C.: SFRC. ———. 1997. Bank Activities and the Extension of Bank Subsidies. Statement 137, May 5. Found September 1999 at . Shull, Bernard, and Lawrence J. White. 1998. “The Right Corporate Struc- ture for Expanded Bank Activities.” Banking Law Journal 115: 446–76. Stanton, Thomas H. 1991. A State of Risk: Will Government-Sponsored Enter- prises Be the Next Financial Crisis? New York: HarperCollins. ———. 1999. “Devising an Effective Legal Framework for Supervising the Public Benefits and Public Costs of Government-Sponsored Enterprise.” RICHARD SCOTT CARNELL 83

Paper presented at conference, “Fannie Mae and Freddie Mac: Public Pur- poses and Private Interests,” September 8, American Enterprise Institute, Washington, D.C. Stigum, Marcia. 1990. The Money Market. 3d ed. Homewood, Ill.: Dow Jones- Irwin. U.S. Congressional Budget Office. 1996. Assessing the Public Costs and Ben- efits of Fannie Mae and Freddie Mac. Washington, D.C.: Government Print- ing Office. U.S. Department of Justice, Office of Legal Counsel. 1987. “Debt Obligations of the National Credit Union Administration.” Opinions of the Office of Legal Counsel 6: 262–72. U.S. Department of the Treasury. 1996. Government Sponsorship of the Federal National Mortgage Association and the Federal Home Loan Mortgage Corpora- tion. Washington, D.C.: Government Printing Office. U.S. General Accounting Office. 1990. Government-Sponsored Enterprises: The Government’s Exposure to Risks. GAO/GGD-90-97. Washington, D.C.: Gov- ernment Printing Office. ———. 1996. Housing Enterprises: Potential Impacts of Severing Govern- ment Sponsorship. GAO/GGD-96-120. Washington, D.C.: Government Printing Office. U.S. House of Representatives. 1997. Committee on Banking and Financial Services. Examination of Financial Modernization within the Jurisdiction of the Committee on Banking and Financial Services: Hearings before the Subcom- mittee on Capital Markets, Securities and Government Sponsored Enterprises. 105th Cong., 1st sess., H. Hrg. 5. U.S. Senate. 1998. Committee on Banking, Housing, and Urban Affairs. The Financial Services Act of 1998—H.R. 10: Hearing. 105th Cong., 2d sess., S. Hrg. 1017. Van Order, Robert. 2001. “Devising an Effective Legal Framework for Su- pervising the Public Benefits and Public Costs of GSEs.” In Fannie Mae and Freddie Mac: Public Purposes and Private Interests, edited by Peter J. Wallison. Washington, D.C.: AEI Press. Walter, John R. 1998. “Can a Safety Net Subsidy Be Contained?” Federal Re- serve Bank of Richmond Economic Quarterly 84 (winter): 1–20. Zigas, Barry. 1996. Letter to Richard S. Carnell. June 28. Much debate concerning Fannie Mae and Freddie Mac revolves around economic analysis. As Robert Van Order contends in chapter 3, one can make a highly sophisticated argument in support of the GSEs without denying that they receive implicit subsidies from the federal government. According to supporters of Fannie and Freddie, the additional benefits that the GSEs offer outweigh whatever cost can be reasonably attributable to their subsidy. The additional benefits include advancing the government policy of expanding homeownership, bringing technological progress into the residential mortgage market, and maintaining mortgage market liquidity. Overall, they contend, these benefits contribute more to societal welfare than they cost as subsidies. Charles W. Calomiris notes that Van Order bases his argument on the mistaken view that national banks use any subsidy only for the mortgage markets. He points out that if the GSEs were successful in driving the banks out of the residential mortgage markets because they used their subsidy more efficiently, the banks would then simply use their subsidy in some other area of the economy, and no increase in welfare would ensue. In this chapter Calomiris analyzes these claims from the point of view of an economist. He argues that Fannie and Freddie do not meet the economist’s standard for an optimal mechanism—one that maximizes the potential difference between realized benefits and realized costs, taking account of the opportunity costs associated with forgoing other means of realizing the same gross benefits. Indeed, he concludes, “It is more plausible to argue that these institutions do significant harm by burdening taxpayers, raising housing prices, discouraging homeownership, and retarding technological progress and competition in the mortgage market.”

84 5 An Economist’s Case for GSE Reform

Charles W. Calomiris

An optimal mechanism is one that not only produces some gross benefits (as virtually all mechanisms do), but produces the most net benefits among conceivable alternatives—a mechanism that achieves bona fide objectives at least cost. Are the government- sponsored enterprises that have come to dominate the mortgage market—specifically Fannie Mae and Freddie Mac—optimal mechanisms for accomplishing legitimate policy objectives? Is their mixing of private interests and public purposes—that is, the combining of government subsidies and the conferring of monopoly status on a privately owned duopoly1—the best way to achieve some set of desirable outcomes? In this essay I address that question and point to some logical problems in the positions advocated by the mortgage GSEs. One cannot reasonably argue that these GSEs are useful mechanisms in the current environment. The perspective of the 1930s that considered Fannie and Freddie as beneficial no longer holds. I begin by reviewing the economic arguments for and against the GSEs and finish by considering various options for reform.

Criteria for Appraising Privileges and Subsidies

The possible economic advantages or disadvantages produced by any financial institution created for the public interest divide into five categories: (1) subsidizing or taxing an activity that society regards as undervalued or overvalued in the market; (2) improving or worsen- ing static productive efficiency through economies or diseconomies of scale or scope; (3) speeding or retarding productivity growth by

85 86 GSE REFORM influencing technological change; (4) reorganizing market transactions in a way that produces more or less liquidity, perhaps by encouraging or discouraging product homogeneity or by standing ready to make markets in certain assets; and (5) altering the political economy that underlies public finance for the better or worse. Advocates and opponents of the special chartering of private financial corporations have argued the merits of such special privileges on one or more of those five grounds. The debate began with the chartering of monopolistic banks under the European empires of the seventeenth century and was revisited in the United States during the struggles over the chartering of the First and Second Banks of the United States and in the hundreds of state-level conflicts over the special chartering of state banks in the antebellum period. In the postbellum era the chartering of the Federal Reserve Banks and the creation of the GSEs saw similar debates. In that sense the current disagreement over the GSEs is nothing new. Historically, financial institutions have often been chartered and granted special privileges due to some combination of expected social gains and the political influence of their founders; such entities have sometimes been abolished when their missions became outdated or the relative political influence of their owners waned. In judging the merits of continuing special privileges and subsidies to a GSE, by any of the five criteria, one must focus on net benefits. In economics net benefits are not defined simply as the difference between gross benefits and realized accounting costs, but rather the difference between gross benefits and the full range of economic costs, including opportunity costs. An optimal mechanism maximizes the potential difference between realized benefits and realized costs, and only the optimal mechanism has a positive net benefit when costs are defined inclusive of opportunity costs, which take account of all forgone means to realizing gross benefits.

Subsidizing Mortgages. On behalf of Fannie Mae, Adolfo Marzol (1999) has argued against the view that the GSEs’ redeeming social value was rooted in overcoming market failures. He described as a snipe hunt the search for market failures (inefficiencies) to justify GSE activities. In his view the central issue is the extent to which the GSEs subsidize mortgages and thereby make homeownership more possible and more affordable. It is often argued in this connection (see Weicher 1999) that homeownership benefits more than the homeowner alone: because homeowners, some believe, are better citizens, gov- CHARLES W. CALOMIRIS 87 ernment should subsidize homeownership. In the jargon of economics homeownership creates positive externalities. Demonstrating that the subsidization of homeownership is a bona fide objective worthy of subsidy (and for present purposes I concur) is not enough; rather, advocates of the GSEs must argue that the best mechanism for subsidizing homeownership is the granting of privileges and subsidies to the GSEs. How much do the GSEs receive from the government, and how does that government subsidy to the GSEs ultimately affect the homeownership decisions of American families? Is there a more powerful and less costly way to encourage homeownership? The Congressional Budget Office has estimated that Fannie Mae and Freddie Mac receive more than $6 billion per year in off–balance- sheet taxpayer subsidies from the implicit protection that their debt receives by the government (see chapter 2 in this volume).2 Robert Van Order (chapter 3 in this volume) of Freddie Mac contested the size of that subsidy but not the basic logic that CBO and others use to argue that a large subsidy exists.3 The size and importance of the GSEs and the role of the government in GSE governance and in a credit line at the Treasury account for the extraordinarily low interest that these entities pay on their debt. The CBO estimates that roughly one-third of that $6 billion is transferred as profit to the stockholders of the GSEs, while two-thirds is passed on to mortgage sellers in the form of higher prices (lower interest rates) in the secondary mortgage market. The higher prices paid by subsidized GSEs in the secondary mortgage market are passed on to homeowners as interest rate savings in the competitive primary mortgage market. But is the most effective means of promoting homeownership a GSE mechanism that operates through lower mortgage interest rates and suffers a one-third leakage when transferring taxpayer-financed subsidies to homebuyers? Can’t the government find a better method to transfer subsidies in support of homeownership? The easiest approach to those questions is to consider a tan- gible alternative mechanism, namely, the use of down payment subsidies administered by the government. Would that mechanism be a better vehicle for mortgage subsidization? Would it likely be more or less costly to administer? An answer to the first question requires an in-depth analysis of how current interest rate subsidies are chan- neled. How different an effect would a down payment subsidy have? An answer to the second question calls for determining whether the transaction costs of administering a down payment assistance 88 GSE REFORM program would exceed one-third of the subsidies transferred (the leakage from GSE subsidies). In a determination of the more powerful vehicle for producing benefits, down payment assistance would be much more effective in accomplishing the central social objective of increasing homeownership. GSE subsidies are too small and too spread out to affect homeownership affordability; they do not target enough assistance to those for whom it would matter the most, that is, low-income individuals without sufficient accumulated wealth to qualify for a mortgage. Under the current GSE mechanism, all qualifying homeowners enjoy a small interest rate reduction (which in the aggregate they as taxpayers repay 150 percent to the government). In the aggregate, therefore, the GSEs tax rather than subsidize the public. Middle- class homeowners are the recipients of the largest gross transfers from government support for the GSEs and possibly the largest net transfer. Upper-income taxpayers do not often qualify for GSE- purchased mortgages (because of limits on the size of qualifying mortgages), and lower-income taxpayers are often renters (sometimes because they lack sufficient accumulated wealth to meet the requisite down payment on a home). For a low-income taxpayer, ten or fifteen basis points less in annual interest cost translates to a $50 or $75 reduction in the annual cost of debt service (assuming a $50,000 mortgage), which probably does not matter much in the decision to buy a home. Fannie Mae and Freddie Mac often contend that they make special efforts to assist low-income households to become eligible for mortgages. Although the GSEs like to advertise their special programs for lower-income homebuyers, in fact they do not channel significant subsidies to the low-income. Channeling such subsidies would mean absorbing a significant share of their credit risk at below-market cost. Evidence suggests that Fannie Mae and Freddie Mac limit their exposure to credit risk on low-income mortgages either by requiring high down payments (Bunce and Scheessele 1996) or by relying on external credit enhancements.4 Fannie Mae and Freddie Mac tend to avoid areas with significant credit risk (especially in multifamily mortgages, where Fannie Mae is the sole lender at risk for only 11 percent of its portfolio) (Fannie Mae 1999, 28). The GSEs lay off risk partly because of statutory limits on their powers (which require private mortgage insurance for mortgages with loan- to-value ratios exceeding 80 percent) and partly because of their profit-maximizing choices in risk management. Laying off credit risk CHARLES W. CALOMIRIS 89 is rewarded with lower regulatory requirements regarding capital and thus a higher return on equity. Both Fannie Mae and Freddie Mac have rapidly increased their use of external credit enhancement. In 1998 nearly one-third of all mortgage-backed securities issued by Fannie Mae involved some form of third-party credit enhancement (Fannie Mae 1999, 38), and 40 percent of Freddie Mac’s mortgage purchases were credit enhanced (Freddie Mac 1999, 47). Credit-enhanced mortgages accounted for 15 percent of all outstanding Fannie Mae mortgages held or securitized at year-end 1998 (p. 58). For Freddie Mac, 27 percent of its outstanding mortgage portfolio at year-end 1998 had some form of external credit enhancement (Freddie Mac 1999, 27). To a first approximation, the GSE subsidies are best thought of not as a means of promoting homeownership, but rather as a way to transfer a small amount of income (per household) from the rich primarily to the middle class.5 If the implicit tax and transfer of income were the sole objective of the GSEs, then the tax system could accomplish it more effectively and at much lower cost (that is, without the one-third leakage to GSE stockholders). Consider how much greater an effect on homeownership a down payment assistance program would have. Suppose the government made available up to $10,000 in matching down payment subsidies to all homebuyers. Specifically, individuals could receive (once in a lifetime) up to 50 percent of their down payments (up to a maximum of $10,000) from the federal government, limited for use in purchasing homes that cost less than a given amount (say, $150,000). The transfer of the funds to the homebuyer, once made, would be irrevocable.6 If this program could be administered at neglible cost, the current off–balance sheet cost to the government of more than $6 billion would permit more than 600,000 low- and middle-income families per year to qualify for substantial assistance toward buying a home. Many poorer families who cannot take advantage of current GSE subsidies because of minimum wealth constraints would become homebuyers because of the larger up-front subsidization of down payments.7 A means-tested system of matching funds for down payments offers another important advantage: in contrast to interest rate subsidies from the GSEs, the impact on the value of homes would not undermine the program’s effectiveness. By reducing the effective discount rate for the flow of nonpecuniary benefits of homeownership, interest rate subsidies from GSEs raise home prices in a way 90 GSE REFORM that eliminates much or all benefit of the reduced interest rate on the mortgage. Thus the effective net tax rate on homeowner-taxpayers from GSE subsidies may be 250 percent rather than 150 percent of the interest rate subsidy (as the homeowner-taxpayer must pay for the subsidy twice: once as a homeowner and once as a taxpayer). Down payment assistance would raise home prices somewhat, but because it would relax the constraints of minimum wealth on homeownership, the program’s net effect on recipients’ welfare would be unambiguously positive—the higher price of the home would be more than offset by the effect of the relaxation of the wealth constraint. Furthermore, targeting assistance to low-income individuals would generate additional positive externalities. By boosting homeownership and home values in predominantly low-income areas, means-tested down payment assistance would contribute to the renewal of America’s cities in ways that existing GSE subsidies do not. What about the costs of administering this program? Even if the down payment assistance program suffered administrative expenses equal to the enormous GSE leakage rate of one-third, it would still permit more than 400,000 low- and middle-income families per year to qualify. But the administrative costs of the proposed assistance program should be close to zero. No government discretion would be needed to determine who qualifies for assistance. Quali- fying for assistance would be easy to verify. The only information needed—the identity of the applicant and the price of the house— would be easy-to-observe facts that could be warranted by the mortgage originator. If the objective of GSE subsidies is to promote homeownership by eliminating the financial barriers faced by low- and middle- income families, then down payment assistance would be a much more powerful and cost-effective means to that end. The subsidization of the GSEs cannot be reasonably justified on the basis of the objective of promoting homeownership.

Static Efficiency. Another dimension of possible social gain from subsidizing the GSEs is the promotion of static efficiency in mortgage intermediation. To avoid confusion, three arguments must be clearly distinguished: (1) the argument in defense of federal chartering of nationwide mortgage intermediaries (without government guarantees or monopoly rights); (2) the argument supporting the chartering of monopolistic nationwide mortgage intermediaries; and (3) the argument for government subsidization of those intermedi- CHARLES W. CALOMIRIS 91 aries. None of these arguments—for chartering GSEs, for granting them a monopoly, or for subsidizing their cost of funds—can be justified in the U.S. mortgage market today. For the first type of argument—federal chartering of mortgage banks—it once was possible to contend that because of branching restrictions on banks and no practical means for securitizing mortgage portfolios, the chartering of national mortgage intermediaries in the 1930s offered large potential efficiency gains from economies of scale. Economies of scale of such intermediaries relative to small, geographically isolated banks result from the ability to spread fixed costs over a larger portfolio and to achieve superior portfolio diversification by holding a national mortgage portfolio. That historical argument, though reasonable, cannot justify perpetuating the current GSEs. First, such reasoning does not imply any benefit from a limited number of monopolistic GSEs, that is, the current Fannie-Freddie duopoly; on the contrary, the argument actually supports chartering competitive nationwide banks—to avoid the costs of limits on branching—and applies to all areas of banking, not just mortgages. Second, the chartering argument suggests no need for subsidizing such nationwide mortgage banks because their economies of scale would be realized without government help. Third, the historical argument holds little practical relevance to the U.S. mortgage market today. Branching restrictions on banks have been removed, and private commercial banks can operate competitive, nationwide networks. Furthermore, the revolution in mortgage securitization (and more generally the securitization of virtually all bank loan products) now allows originating institutions with limited capital to participate in the creation of enormous portfolios of assets (tranches of those are sold to a variety of international investors in the marketplace). Any need for a concentration of capital in the area of origination to permit the creation of large portfolios of mortgages has greatly diminished. A second type of argument defends the creation of a monopolistic mortgage bank on the grounds of efficiency. Limits to competition among GSEs may have promoted greater standardization of mortgages, which may have facilitated the development of a national mortgage market. That argument, though somewhat plausible, is also a historical one with scant relevance to today’s marketplace or to the debate over eliminating the current GSE duopoly. Economies of standardization involve the creation of focal points that help standards to develop. If those standards are beneficial— 92 GSE REFORM and possibly even if they are not —they will be self-reinforcing; any need to limit the creation of competitive standards no longer exists. Furthermore, limiting markets to a single standard incurs costs— we return to that issue in the discussion of technological progress and efficient pricing. Securities markets provide interesting examples of the creation and perpetuation of benchmarks and product standards in the absence of monopoly rights. The freedom to adopt a standard or to differentiate one’s product produces an interesting mix of imita- tion and innovation. For example, the Standard & Poor’s 500 remains a powerful focal point of portfolio evaluation because of economies of standardization. At the same time new benchmarks with particular purposes—to track small stocks or Internet stocks, for example—have developed as new needs arise. Reaping the gains of product standardization does not require the creation of institutional monopolies. The third class of argument supports the subsidization of mortgage intermediaries on the grounds of static efficiency. Robert Van Order (1999) has put forth the only such argument applied to the GSEs. He contends that the GSEs are more efficient mortgage portfolio creators than banks (due to purported economies of scale and superior technology) and therefore allocative efficiency requires that the GSEs (rather than banks) manage mortgage portfolios. Van Or- der argues that government subsidies to banks—typically thought to include underpriced deposit insurance and access to the discount window and Fedwire—may artificially reduce banks’ costs of funds and thus permit banks to outbid the GSEs for mortgages. He charges that by subsidizing the GSEs, the government can reduce the deadweight loss from permitting inefficient banks to displace GSEs as mortgage intermediaries. Though a clever argument for subsidizing the GSEs, it is not a convincing one for four reasons. First, as already stated, the notion of significant technological superiority of the GSEs as mortgage portfolio managers is no longer plausible in the current era of large-scale nationwide banking and private securitization of mortgages. Second, empirical studies of banks do not support Van Order’s claim that banks receive large net safety net subsidies (through deposit insurance or access to the discount window) that allegedly reduce their cost of funds. That situation may have been true of banks in the 1980s, but much has changed since then. Those changes include reforms to thrift capital standards in the Financial Institutions Reform, Recovery, and Enforcement Act of 1989 (FIRREA), rules for CHARLES W. CALOMIRIS 93 enforcing capital requirements for financial institutions and limits placed on the too-big-to-fail doctrine in the Federal Deposit Insur- ance Corporation Improvement Act of 1991 (FDICIA), additional reforms enacted in 1991 that limited Fed lending to distressed banks, depositor preference legislation in 1993 that reduced FDIC liabilities for failed banks, and the high capital ratios of banks in recent years. Empirical studies estimating the implicit safety net subsidy received by commercial banks indicate little or no existing average subsidy; for many banks the safety net rules entail a net tax (Hovakimian and Kane 1999).8 Third, even if banks were being granted a safety net subsidy that allowed them to compete inefficiently in the mortgage market, surely creating new GSE subsidies to counteract the bank safety net subsidy would not remedy the problem. If banks receive a distortionary safety net subsidy, then bank capital standards, prudential regulation, deposit insurance pricing, and access to the discount window should be further reformed. Indeed, even though banks enjoy little or no current average safety net subsidy, I have advocated several further reforms to bank regulation that would provide additional safeguards against such abuse (Calomiris 1994, 1997, 1999; SFRC 1999). Fourth, Van Order bases his welfare analysis on an incomplete partial-equilibrium model of bank lending that focuses only on the mortgage market. But suppose that a bank enjoys a safety net subsidy that permits inefficient entry into any lending market. Even if a countervailing GSE subsidy blocks that bank from entering the mortgage market, the institution can still enter other markets: those for consumer loans, credit card receivables, and commercial loans and leases. Contrary to Van Order’s analysis, preventing the subsidized bank from entering the mortgage market by subsidizing the GSEs worsens welfare from a general-equilibrium perspective. The GSE subsidy does not prevent the bank from inefficiently accessing other loan markets. The GSE subsidy in general equilibrium adds a new distortion and inefficiency to an already distorted financial system.

Technological Progress. According to another argument sometimes made for preserving the mortgage GSE duopoly—though not an argument for subsidizing that duopoly—the GSEs can perhaps better capture the benefits of technological improvements, and that situation might spur them to be more innovative than competing private banks. That highly controversial argument has questionable relevance in today’s mortgage market. 94 GSE REFORM

In a general theoretical and empirical context, countervailing arguments and evidence (beyond the scope of this chapter to review) suggest that monopolies might be less innovative than competitive systems. Some observers of the computer industry, for example, fault Microsoft (which manages the dominant operating system for personal computer users) for slowing technological progress by gobbling up new software products and either keeping them off the market or delivering them within its operating system in an undesirable way. Other arguments suggest that the social costs of monopoly may outweigh the benefits of promoting greater technological progress. In any case, singling out financial innovation as an area where technological progress requires a monopoly is strange. The revolution in financial engineering, derivatives, and securitization has pro- ceeded apace in all financial markets, and many observers argue that competition has spurred those developments (Calomiris 1998). Scant evidence in the mortgage market of recent years points to GSE technological superiority. Indeed Fannie Mae and Freddie Mac have significantly lagged behind in some of the most important new mortgage product areas. For example, the move toward credit scoring in the pricing of mortgages and the related development of the new high-loan-to-value and subprime mortgage markets occurred because of private initiatives and initially relied entirely on private financing sources (Calomiris and Mason 1998). The pioneers in the field were new finance companies that, unlike the GSEs and insured commercial banks, lacked any access to government safety nets. Arguably the GSEs’ policy of pooling mortgage risks (that is, purchasing mortgages in the secondary market as pools rather than on the basis of their estimated individual credit risk and prepayment risk) has slowed the development of credit risk and prepayment risk-pricing models. Even though the GSEs are only secondary market purchasers, if they eschew the efficient pricing of individual credit and prepayment risk, then originators have little incentive to adopt risk-pricing models in the primary market given the GSEs’ dominant position in that market. Credit-risk models based on Fair Isaac Co. (FICO) scores have been used for pricing so-called nonconforming market products (high-loan-to-value and subprime mortgages), precisely the market niches where the GSEs have been least dominant. The GSEs do sort mortgages according to their risks to some extent, as indicated by the varying credit-enhancement requirements CHARLES W. CALOMIRIS 95 for different mortgages, and use credit scoring as part of their internal risk-management process, but they still pool risks to a significant degree in their purchases of mortgages. Originators would face stronger incentives to track and report individual characteristics related to prepayment and default risk if individual mortgage prices reflected all observable individual risks, as would be the case in a competitive secondary market.9 Inefficient risk pooling may have partially spurred the recent entry into the secondary market by the new Mortgage Partnership Finance (MPF) program, pioneered by the Federal Home Loan Bank (FHLB) of Chicago. In essence the MPF program allows banks to retain the credit risk associated with the mortgages that they originate, while passing on the market risk to the participating FHLB. Splitting credit risk from market risks allows originating banks to benefit directly from measuring, controlling, and retaining default risk and may encourage better pricing of risk in the primary mortgage market.10 Furthermore, the potential gains to competing financial firms from investing in financial innovation may be even higher in the future. Recent legal precedents encouraging the patenting of financial products—which will ease protecting the rights of inventive financial engineers—point toward an increasing ability of competing private market participants to capture the gains of their inventions (White and Case 1999). In summary there seems to be no legitimate basis to argue for GSE monopolization of the mortgage market as a means to spurring technological progress. Competitive firms have managed to innovate successfully—particularly in financial markets including the mortgage market—and will continue to do so.

Liquidity. In August 1998 the Russian debt default and its spillover effects buffeted mortgage security markets and other markets for publicly traded debt. Those spillover effects reflected the behavior of financial institutions and other investors in the wake of large declines in asset values. The declines reduced the capital of investing institutions and forced them to curtail overall asset risk to limit the rising default risk on their own debts. The sudden sales of risky assets generated an abnormal spike in risk and liquidity premiums (the yield spreads on illiquid, risky securities) as investors scrambled for low-risk securities and cash. Fannie Mae and Freddie Mac like to point to the accelerated growth of their portfolios during the August 1998 crisis as evidence 96 GSE REFORM of their important role as providers of liquidity during financial cri- ses. Freddie Mac (1999, p.10) boasts that “when . . . spreads wid- ened, we seized the opportunity to grow our retained portfolio by a record $91 billion, more than three times the growth we generated in 1997.” Fannie Mae (1999, p. 10) also took advantage of the crisis: “During the height of the turmoil in financial markets, Fannie Mae purchased nearly $50 billion of mortgages.” Access to implicit government support and a credit line at the Treasury underlay the willingness of the GSEs to continue absorbing risk while other institutions were scrambling for liquidity. The liquidity argument for the GSE subsidies merits serious consideration. As with the other arguments reviewed, the liquidity argument is incomplete; it emphasizes gross advantages, not net gains, and does not consider alternative, possibly superior means for achieving the same ends. That the GSEs are the best means of achieving net gains from liquidity protection in the mortgage market is far from obvious. First, the willingness of Fannie Mae and Freddie Mac to absorb risk may be viewed negatively (as symptomatic of inefficient subsidization of risk) as well as positively (as symptomatic of greater access to liquidity protection). The willingness of thrifts, GSEs, and large money-center banks to absorb interest rate risk and default risk in real estate markets and debt markets in the 1980s resulted in enormous losses, which in large part have been attributed to the incentives that they faced as protected institutions (Barth and Bartholomew 1992; Boyd and Gertler 1994; Brewer 1995; Stanton 1991). Not only did those excessive risk-loving decisions lead to large losses for taxpayers, they also distorted resource allocations and wasted funds that could have been used productively elsewhere. Furthermore, by making the financial system more vulnerable, that risk-taking binge set the stage for the capital crunch of the late 1980s and early 1990s and the protracted recession that accompanied it. The willingness to absorb risk does not always contribute to the financial system. A desirable mechanism for liquidity protection combats mo- ments of liquidity crisis (generally measured in days or weeks as in June–July 1970, October 1987, and August–September 1998) without promoting a long-run tendency to undertake excessive risks. Fortunately such a mechanism exists: the Federal Reserve System’s discount window. The Fed’s mission is to provide liquidity to the financial system. Most of the time that purpose means targeting the overall sup- CHARLES W. CALOMIRIS 97 ply of money or credit in the economy to be consistent with stable, noninflationary growth. But sometimes the Fed’s role as a central bank involves targeting specific markets or intermediaries for assistance. The discount window is the appropriate tool for that purpose (Calomiris 1994). For example, the Fed made it clear to banks that it was concerned about the upheaval in the commercial paper market in June 1970 and about the liquidity needs of the securities and fu- tures markets in 1987. The Federal Reserve did not offer to protect banks from the credit risks of lending to commercial paper issuers or securities houses, but it did make it clear that banks could access the discount window (where borrowing rates are lower than in the private market) without fear of the implicit nonpecuniary regulatory penalties that normally accompany large borrowings from that subsidized source. The Fed’s policy regarding the discount window provides a state-contingent source of protection against temporary squeezes in particular markets. If banks (or reformed GSEs without access to government subsidies but with access to the discount window) became the primary holders and securitizers of mortgages, then the Fed could assume the role of protecting the mortgage market against sudden disruptions to liquidity, much as it has done in other markets. And the Federal Reserve would do so with minimal moral- hazard costs from encouraging long-run excessive risk taking. Thus a more effective, less costly means for providing liquidity protection for the mortgage market than the current GSE subsidies exists.

Political Economy. The first Bank of the United States was an effective fiscal agent for the federal government and was widely viewed as a uniquely valuable payments intermediary for the nascent nation. Yet its recognized efficiencies did not prevent opponents from block- ing the renewal of its charter in 1811. President Andrew Jackson’s veto of the rechartering of the second Bank of the United States in 1832 also ended a quite efficient intermediary. The second bank seems to have played a unique role as the only interstate bank during the 1820s to serve as a conduit for interstate trade finance (through the intermediation of bankers’ acceptances) and as a disciplinarian that prevented other banks from issuing excessive amounts of liabilities by monitoring bank note issues and channeling excessive amounts back to the banks that issued them (Temin 1969; Calomiris 1993). Those decisions—which many economists and historians (including myself) have criticized—still are not generally regarded as 98 GSE REFORM entirely unwarranted.11 Jefferson opposed the first Bank of the United States and Jackson opposed the second bank primarily because they saw in them a great risk to the political economy of the republic. Too great a concentration of political power, they believed, threatened the democratic process. Some critics have portrayed opponents of these banks as unsophisticated, and surely some of them were. But such concerns cannot be so easily dismissed. In a republic legislators enact policy, and their initiatives and votes reflect the political power of existing constituencies. Creating new institutions reallo- cates economic and therefore political power. Institutions that may produce net benefits in some static sense may have unforeseen and undesirable consequences. The political economy case against the GSEs is even stronger than the case opposing the first and second Banks of the United States. Observers of the current GSEs often note that they spend an enormous amount of resources, time, and effort lobbying the federal government to influence economic policy. Fannie Mae and Freddie Mac’s senior executives often seem to be hired more for their political connections than for their knowledge of the mortgage market (Ullmann 1999). Their reach is broad and their power is great. They exert more control over the markets in which they participate economically and politically than any other financial institution in U.S. history. Although an economist has difficulty in quantifying the political costs to society of creating such entities, economic research into how government reaches its decisions suggests that the creation of concentrated vested interests in general entails significant costs borne by average citizens that should be considered (Olson 1965; Stigler 1988). Political economy considerations reinforce the conclusion of the previous analysis that no sustainable economic argument exists for net social gains from maintaining the GSEs as a subsidized duopoly.

Implications for Reforming the GSEs

Institutions have a way of perpetuating themselves partly because of the political influence that they maintain and partly because of the difficulty of orderly dissolution that logic suggests would be desirable. With the GSEs, however, fortuitous circumstances make the way out relatively easy. The Office of Federal Housing Enterprise Oversight has established prudential standards for Fannie Mae and Freddie Mac (see HUD 1999). Because the amount (per dollar of debt) that these GSEs collect CHARLES W. CALOMIRIS 99 as a government subsidy is proportional to the riskiness of their debt, establishing credible risk-based capital requirements that target low default risk for their debt could rein in their subsidy. Although OFHEO’s task may seem rather prosaic, in fact the proper measurement of risk and the maintenance of adequate capital commensurate with that risk could save the U.S. taxpayers billions of dollars annually. Congress should also establish a carefully crafted subordinated debt requirement for the GSEs to provide additional market discipline to supplement regulatory discipline.12 Even if capital guidelines could be established and enforced adequately by some combination of market and regulatory discipline, limiting the GSEs’ subsidy is not enough; the power that Fannie Mae and Freddie Mac wield in markets and in politics must be reduced. Because Fannie Mae and Freddie Mac both specialize in mortgage intermediation, that structural homogeneity makes it easier by simple division to create several competing entities out of the two. The following five steps offer a fruitful approach to reforming the GSEs and substituting in their place a better means of encouraging homeownership.

• Step 1. Congress and the administration should decide how much the federal government spends on supporting homeownership. That amount should be placed in the budget and should be targeted to homeowners through a down payment assistance program like the one summarized above.13 • Step 2. During the transition to full privatization of the GSEs, OFHEO should continue to develop risk-based capital standards for them that account for the full range of risks undertaken and that effectively target a nearly zero default rate on GSE debt. The required capital maintained by the GSEs should include a mandatory minimum subordinated debt requirement, as described by Shadow Financial Regulatory Committee (1999). Removing OFHEO from HUD and establishing its true independence by insulating its operating budget from annual congressional review would help. Both the Office of the Comptroller of the Currency and the Federal Reserve are insulated from the annual budgetary process. OFHEO should be granted that same independence. • Step 3. Fannie Mae and Freddie Mac either should be phased out entirely (the preferable course) or should be divided into 100 GSE REFORM

several new entities that would not be too big or too politically influential to be protected from failure by the U.S. government. Such new entities should either be subject to the strict capital guidelines described in step 2 or should be invited to apply for bank charters. • Step 4. None of the new entities should have access to credit from the Treasury or should have government-appointed directors. Nor should they receive any other special tax treatment. If the new entities are not reconstituted as banks, they should still be given access to the Fed’s discount window on the same terms as member banks. • Step 5. Once Fannie and Freddie have been truly privatized, the Federal Home Loan Banks should be either privatized or closed. The only possibly legitimate rationale for maintaining the FHLBs in their current subsidized and concentrated form is to provide some competition for Fannie Mae and Freddie Mac (to reduce the one-third leakage of their taxpayer-financed subsidy, which currently flows to their stockholders). Once the special privileges of the other mortgage GSEs have been eliminated, the argument for retaining the FHLBs in their current form dis- appears.14

Conclusion

This chapter has considered possible justifications for perpetuating the monopoly rights and government subsidies enjoyed by Fannie Mae and Freddie Mac. None of the arguments suggests that the mortgage GSEs (Fannie Mae, Freddie Mac, and the Federal Home Loan Banks) meet the economic criterion of an optimal mechanism. Indeed it is more plausible to argue that these institutions do significant harm by burdening taxpayers, raising housing prices, discouraging homeownership, and retarding technological progress and competition in the mortgage market. The promotion of homeownership, static efficiency in the mortgage market, technological progress in mortgage products and pricing, and mortgage market liquidity could be achieved best by abolishing the subsidies and the economic and political power enjoyed by the GSEs and by pursuing alternative, superior means to the legitimate ends described above. Doing so not only would improve the efficiency of the financial system and make homeownership more achievable for millions of CHARLES W. CALOMIRIS 101

Americans, it would also improve the quality of our political institutions by removing from the political arena a powerful voice for special interests at public expense.

Notes

1. The characterization of Fannie Mae and Freddie Mac as a duopoly is apt. They often coordinate their actions and do not compete away the value of the subsidy that each receives from the government. To meet their revenue projections, they must have purchased (as soon as the year 2000) 100 percent of all eligible U.S. mortgages originated (all mortgages with face value less than $240,000), and in subsequent years they must purchase all new eligible mortgages as well as a growing proportion of preexisting mortgages held by others. For growth projections, see Wallison and Ely 1999. 2. The Congressional Budget Office estimates a subsidy to the GSEs of $6.5 billion. In 1995 the U.S. Treasury estimated the subsidy at $4.6–6.9 billion (Smalhout 1999). 3. Van Order (this volume, chap. 3) also argues that the existence of a gross subsidy to the GSEs does not necessarily imply an undesirable distortion from that subsidy. We return to that question in the section on static efficiency. 4. Canner, Passmore, and Surette (1996) show that Fannie Mae and Freddie Mac bear little of the total credit risk from loans to low-income borrowers in the United States. For 1995 Fannie Mae and Freddie Mac extended 14 percent of all lower-income, FHA-eligible mortgages (in dollar terms) but absorbed only 4 percent of the risk in the market for FHA-eligible, low-income mortgages (p. 1089). The GSEs are required to obtain private mortgage insurance from outsider providers on low–down payment mortgages and thus cannot subsidize credit risk in that way. But even where they could subsidize credit risk, for example, to high-risk, low-income homeowners, they often appear to choose to raise down payment requirements or obtain credit enhancement from originators or third parties. While the details of that risk sharing are not laid out in detail in the annual reports of Fannie Mae (1999) and Freddie Mac (1999), their discussion of risk sharing in those reports indicates that they obtain credit enhancements from outside investors on loans with high credit risk. Thus while the GSEs as profit-maximizing firms intent on earning high returns on equity could in principle target assistance to high-risk, low-income mortgages by absorbing credit risk of those mortgages at below market interest costs, they choose not to do so to any significant extent. 102 GSE REFORM

5. Under the current GSE system, the poorest families (say, those with annual income less than $20,000) receive small positive net transfers from the GSEs since the amount that they receive is small (often zero) and their share of federal tax burden (other than payroll tax) is essentially zero. Middle-income families (say, those earning roughly $20,000–50,000) probably receive the biggest net transfers from the GSEs. Those citizens can qualify for homeownership on the basis of their income and wealth and thus can gain access to the GSE subsidy but pay little federal income tax (other than payroll tax) to finance the GSEs. Middle-upper-income families (say, those with incomes of $50,000–100,000) receive and pay for the GSE subsidies, and for them the net transfers (as a fraction of their income) are smaller. The wealthiest families receive little or no GSE benefits (because their mortgages typically exceed the $240,000 size limit for GSE purchases), and as taxpayers they pay a significant fraction of the cost of financing the GSE subsidies. Another complicating factor in calculating the net subsidies received by households is the effect of the GSE subsidies on the prices of houses purchased. Reduced mortgage rates—by reducing the effective discount rate on housing services—lead to the bidding up of the value of homes, and thus new home purchasers may receive little or no gross benefit from subsidized mortgage rates. 6. From an incentive standpoint, down payments are useful for ensuring that homeowners have a vested interest in their homes. Making the transfer from the government to the homeowner irrevocable (rather than a loan), preserves that incentive benefit. 7. For a discussion of the economics behind credit constraints based on minimum borrower wealth, see Stiglitz and Weiss 1981 and Calomiris and Hubbard 1990. 8. Kwast and Passmore (1997) argue that commercial banks continue to enjoy a safety net subsidy. Their evidence is that commercial banks have lower capital ratios than most other financial institutions. Calomiris and Mason (1999) show, however, that risk-adjusted capital ratios (which vary according to the riskiness of the assets of an intermediary) are similar across intermediaries; banks’ lower capital ratios reflect the lower risk of their assets. For additional criticisms of the view that banks enjoy a large safety net subsidy, see Ely 1999. 9. The pooling of risks in the pricing of mortgages is inefficient because it entails a cross-subsidy from low-risk individuals to high-risk individuals. Those cross-subsidies are not only inefficient in a static sense, they also fail to reward good behavior (that can produce a high credit rating), which can raise overall risk in the market. Nevertheless, it is possible to argue that pooling serves a socially desirable outcome since it provides some insur- CHARLES W. CALOMIRIS 103 ance to individuals against random occurrences that affect credit and prepayment risks. For that reason pooling could be preferred (from the standpoint of aggregate welfare) to sorting. In any case many alternative means exist other than the GSEs to achieving pooling, including regulations requiring the uniform pricing of conventional mortgages by originators. 10. It is not yet clear whether the FHLBs will be successful in their challenge to the Fannie-Freddie duopoly. The FHLBs themselves are GSEs and can thus compete with Fannie Mae and Freddie Mac in ways that private intermediaries cannot. Some observers—notably Alex Pollock (1998), president of the Chicago FHLB—have argued that the MPF program would add competition to the mortgage market and thus might reduce the rents that Fannie and Freddie can extract. That may be, but this sort of competition presents a danger. Increased competition among protected entities (whose subsidies rise with the amount of risk they undertake) can encourage greater risk taking at taxpayers’ expense. For that reason the creation of new subsidized GSEs or enhanced competition among GSEs is not a substitute for eliminating GSE subsidies. 11. A better solution—one that would have retained the social gains from having these institutions but also would have addressed concerns about their excessive power—would have been the free chartering of banks with nationwide branching power (the feature that made the first and second Banks of the United States so valuable). 12. For a detailed explanation of the advantages of such a requirement and the pitfalls to be avoided when constructing a credible tranche of junior, risky debt offerings by protected financial institutions, see Calomiris 1997, 1999 and Shadow Financial Regulatory Committee 1999. 13. More generally it would improve government use of resources if annual budget calculations included all off–balance-sheet items with all implicit or explicit government guarantees. Doing so would discourage the perpetuation of inferior policy mechanisms (like GSEs) simply because they facilitate politically convenient, but dishonest, budgetary accounting. 14. The survival and expansion of the Federal Home Loan Banks exemplify the difficulty in eliminating existing institutions in an environment where political entrepreneurship offers high returns. The FHLBs were threatened with extinction in 1989 as the thrift industry suffered a collapse and the functions of thrifts became largely absorbed within the commercial banking sector. The FHLBs were retained initially in 1989 largely to help fund the S&L bailout, for which they were given some explicit liability. To avoid extinction, the Federal Home Loan Banks pursued a two-pronged “regulatory arbitrage” strategy for survival aimed at attracting a new constituency of members to defend them in Congress: commercial banks. The two prongs 104 GSE REFORM of the strategy were (1) offering subsidized credit to mitigate the costs of CRA compliance by banks and (2) offering a new inexpensive general funding source (advances) to substitute for the curtailment of bank access to the discount window. That strategy boosted FHLB membership from less than 3,000 at year-end 1990 to roughly 6,500 by 1998. The Federal Home Loan Bank’s subsidized loans (advances) became particularly attractive to small weak banks in the early 1990s. Those banks were finding it harder to access the Fed’s discount window because of new restrictions on discount window lending enacted in 1991 in the wake of congressional criticism of Fed lending to insolvent banks during the 1980s. For a discussion of Fed lending policies to insolvent banks, see Gilbert 1994, 1995. For a discussion of Fed discount window policy changes, see those articles and Calomiris 1994.

References

Barth, James R., and Philip F. Bartholomew. 1992. “The Thrift Industry Cri- sis: Revealed Weaknesses in the Federal Deposit Insurance System.” In The Reform of Federal Deposit Insurance, edited by J. R. Barth and R. D. Brumbaugh, Jr., pages 36–116. New York: Harper Business. Boyd, John, and Mark Gertler. 1994. “The Role of Large Banks in the Recent U.S. Banking Crisis.” Federal Reserve Bank of Minneapolis Quarterly Review, winter: 2–21. Brewer, Elijah III. 1995. “The Impact of the Current Deposit Insurance Sys- tem on S&L Shareholders’ Risk/Return Tradeoffs.” Journal of Financial Services Research 9: 65–69. Bunce, H. L., and R . M. Scheessele. 1996. “The GSEs’ Funding of Affordable Loans.” Working paper HF-001. Department of Housing and Urban De- velopment, Office of Policy Development and Research. Calomiris, Charles W. 1993. “Regulation, Industrial Structure, and Instabil- ity in U.S. Banking: An Historical Perspective.” In Structural Change in Banking, edited by Michael Klausner and Lawrence J. White, pp. 19–119. Homewood, Ill.: Business One-Irwin. ———. 1994. “Is the Discount Window Necessary? A Penn Central Perspec- tive.” Federal Reserve Bank of St. Louis Review, May–June: 31–56. ———. 1997. The Postmodern Bank Safety Net. Washington, D.C.: AEI Press. ———. 1998. “Universal Banking ‘American-Style’.” Journal of Institutional and Theoretical Economics 154 (March): 44–60. ———. 1999. “Building an Incentive-Compatible Bank Safety Net.” Journal of Banking and Finance, September. Calomiris, Charles W., and R. Glenn Hubbard. 1990. “Firm Heterogeneity, In- ternal Finance, and Credit Rationing.” Economic Journal 100 (March): 80–104. CHARLES W. CALOMIRIS 105

Calomiris, Charles W., and Joseph R. Mason. 1998. High-Loan-to-Value Mort- gage Lending: Problem or Cure. Washington, D.C.: AEI Press. ———. 1999. “Is There a Bank Safety Net Subsidy?” Working paper. Colum- bia Business School, New York. Canner, Glenn B., Wayne Passmore, and Brian J. Surette. 1996. “Distribution of Credit Risk among Providers of Mortgages to Lower-Income and Mi- nority Homebuyers.” Federal Reserve Bulletin 82 (December): 1078–1102. Ely, Bert. 1999. “Banks Do Not Receive a Federal Safety Net Subsidy.” Work- ing paper. Ely & Co., Alexandria, Virginia. Federal Home Loan Mortgage Corporation. 1999. 1998 Annual Report. Wash- ington, D.C.: Freddie Mac. Federal National Mortgage Association. 1999. 1998 Annual Report. Washing- ton, D.C.: Fannie Mae. Gilbert, R. Alton. 1994. “Federal Reserve Lending to Banks That Failed: Im- plications for the Bank Insurance Fund.” Federal Reserve Bank of St. Louis Review 76 (January–February): 3–18. ———. 1995. “Determinants of Federal Reserve Lending to Failed Banks.” Journal of Economics and Business 47: 397–408. Hovakimian, Armen, and Edward J. Kane. 1999. “Effectiveness of Capital Regulation at U.S. Commercial Banks, 1985 to 1994.” Working paper. Bos- ton College. Kwast, Myron, and Wayne Passmore. 1997. “The Subsidy Provided by the Federal Safety Net: Theory, Measurement, and Containment.” Working paper 1997-58, Federal Reserve Board Finance and Economics Discus- sion Series. Marzol, Adolfo. 1999. “Remarks.” Presentation at conference, “Fannie Mae and Freddie Mac: Public Purposes and Private Interests,” American En- terprise Institute, Washington, D.C. Olson, Mancur. 1965. The Logic of Collective Action. Cambridge: Harvard University Press. Pollock, Alex J. 1998. “Mortgage Partnership Finance: A New Option for Depository Institutions, a More Efficient Mortgage Finance System.” Pre- sentation notes, Federal Home Loan Bank of Chicago. September. Seiler, Robert S. Jr. 1999. “Estimating the Value and Allocation of Federal Subsidies to Fannie Mae and Freddie Mac.” Paper presented at conference, “Fannie Mae and Freddie Mac: Public Purposes and Private Inter- ests,” American Enterprise Institute, Washington, D.C. Shadow Financial Regulatory Committee. 1999. International Bank Capital Regulation: What Next? Washington, D.C.: SFRC. Smalhout, James. 1999. “Freddie and Fannie Aren’t Sovereign.” Euromoney, July. 106 GSE REFORM

Stanton, Thomas H. 1991. A State of Risk. New York: Harper Business. Stigler, George, ed. 1988. Chicago Studies in Political Economy. Chicago: Uni- versity of Chicago Press. Stiglitz, Joseph E., and Andrew Weiss. 1981. “Credit Rationing in Markets with Imperfect Information.” American Economic Review 71 ( June): 393–410. Temin, Peter. 1969. The Jacksonian Economy. New York: W. W. Norton. Ullmann, Owen. 1999. “Crony Capitalism: American Style.” International Economy (July–August): 6–11. U.S. Department of Housing and Urban Development, Office of Federal Housing Enterprise Oversight. 1999. Risk-Based Capital Regulation: Second Notice of Proposed Rule Making. Washington, D.C.: OFHEO. Van Order, Robert. 1999. “Notes on the Economics of Fannie Mae and Freddie Mac.” Paper presented at conference, “Fannie Mae and Freddie Mac: Public Purposes and Private Interests,” American Enterprise Institute, Washington, D.C. Wallison, Peter J., and Bert Ely. 1999. “Growth Projections for Fannie and Freddie.” Paper presented at conference, “Fannie Mae and Freddie Mac: Public Purposes and Private Interests,” American Enterprise Institute, Washington, D.C. Weicher, John C. 1999. “The Development of the Housing GSEs.” Paper presented at conference, “Fannie Mae and Freddie Mac: Public Purposes and Private Interests,” American Enterprise Institute, Washington, D.C. White and Case LLP. 1999. “Court Ruling Means That Patents Are a Poten- tial New Source of Competitive Advantage in the Financial Services In- dustry.” New York. PART TWO Controls and Risks

Any discussion of how Fannie Mae and Freddie Mac can be controlled—or their risks reduced—must take account of the politics involved. In this context the two companies may be in a class by themselves. Not only have they linked themselves to the nation’s housing policies—Fannie says it is in ”the American Dream business”—but they have also developed powerful constituencies in the housing and securities industries, which immediately come to the GSEs’ defense when they are challenged in Washington. All this would not be particularly unusual for a business, but the intensity with which the GSEs manage what Fannie Mae calls its political risks is unusual and troublesome. Although a substantial part of their revenues and profits derives directly from their government backing, Fannie and Freddie have no reluctance to return the favor by making large political contributions of soft money to the major political parties and through their executives to individual lawmakers who can help or hurt them; the two GSEs hold fundraisers attended by the major players in their housing and securities constituencies; they open “partnership” offices in the dis- tricts and states of important congressmen and senators; and through their foundations they make grants to community groups that in turn support Fannie and Freddie’s friends in Congress. They play a brand of hardball in the lobbying game that inspires genuine fear among those who oppose them. Because of their extraordinary power and the effectiveness of their public relations, no one usually details those facts. In this chapter Ralph Nader pulls together an indictment that should long ago have gained the attention of the media. The combination of Fannie and Freddie’s wealth, their Washington connections, and their willingness to use both without appar- ent restraint has led some to wonder whether they can ever be controlled. As Nader notes, “The overwhelming political power of these GSEs raises serious questions about the willingness or the ability of the federal government to regulate them.”

109 6 How Fannie and Freddie Influence the Political Process

Ralph Nader

Fannie Mae and Freddie Mac are fast learners. Born of the federal bureaucracy, these enterprises have swiftly and skillfully managed to pick up the roughshod tactics of the private corporate world and at the same time cling tightly to one of the federal government’s deepest and most lucrative welfare troughs. The combination has produced two government-sponsored enterprises that are not only too big to be allowed to fail but perhaps too influential and too politically connected to be regulated or shaped effectively in the public interest. Any suggestion that their power be limited or that subsidies be reduced triggers an immediate no-holds-barred counterattack from Fannie and Freddie. As John Buckley, Fannie Mae’s vice president for communications, bluntly told the Wall Street Journal, “We’re not casual about managing our political risk.” What makes it difficult to deal in any rational manner with Fannie and Freddie’s power is the fact that these GSEs are wrapped around a product—housing—that is right up there in the American psyche with apple pie and motherhood. Unquestionably the GSEs’ secondary market operations have significantly contributed to housing by providing stability and increasing access to the housing finance markets. To suggest otherwise would be inaccurate. Nor do the entities miss a beat in promoting and at times exaggerating their contributions, particularly in the critical area of affordable housing. Fannie and Freddie are extremely skilled at playing the heart strings to the fullest about America’s love affair with homeownership. Rarely does a Fannie Mae newspaper advertisement appear

110 RALPH NADER 111 or a television spot air without emphasis on the corporations’ role in “helping families achieve the American Dream.” Fannie and Freddie are not the first to make good public relations use of the American dream of homeownership. The Federal Home Loan Bank System, created in the New Deal—and the savings and loan industry that it served—used the same phrase to describe its role in housing the nation. For years the link between housing and the savings and loans muted serious questions about the industry. Even as the industry was collapsing in the 1980s, many members of Congress—along with the media—continued to cling to the image of the American Dream and raised no serious questions about where and why the savings and loans had wandered from their original role. When they did look behind the dream, it was too late for oversight, too late to ask the hard questions that clever public relations, weak regulation, and a passive, if not corrupt, Congress and Reagan administration had hidden too long. Letting any industry—exalted missions notwithstanding—climb on a self-constructed pedestal and remain exempt from serious independent inquiry is foreign to a market economy and a democratic system of government. Such exemptions from normal checks and balances can also be quite expensive for the taxpayers. But Fannie and Freddie do not depend solely on the goodwill generated by a housing finance role to shield their corporations from scrutiny. Far from it. These GSEs, particularly Fannie Mae, have built an almost impenetrable moat that essentially seals off their operations from serious ongoing examination by Congress, the media, or even many advocacy groups engaged in promoting affordable housing. The mentality of “see no evil, hear no evil, and speak no evil” that pervades official Washington’s approach to the GSEs is the product of an influence machine that is oiled by revolving doors, the care and feeding of key politicians across the nation, a quick-strike take- no-prisoners public relations operation, and targeted contributions to advocacy organizations—activities financed by slush funds created by generous forms of corporate welfare. Such corporate welfare includes exemptions from D.C. and some state and local income taxes worth $300 million annually while their foundations donate only a fraction of that amount locally. All this is aided and abetted by national media that infre- quently use their vast investigative resources to look beyond the handouts and the smiling faces of young couples and children that 112 INFLUENCING THE POLITICAL PROCESS invariably adorn the slick brochures selling the GSEs’ housing role to the American public. As a result, most citizens have only a vague idea of what lies behind the term government-sponsored enterprises. And certainly Freddie and Fannie do nothing to enlighten homebuyers about the corporate welfare that fuels their operations. But the financial markets are well aware of the benefits that flow to the GSEs from their congressional charters and ongoing links to the federal government. These links provide good reasons for the markets’ perception that the U.S. Treasury and taxpayers would be the fall guys in the event of a default. Here are some of the links to the federal government that lead to this market perception: • Fannie and Freddie each have a contingency line of credit of $2.25 billion that can be drawn from the Treasury. • Their securities are government securities for the purpose of the Securities and Exchange Act of 1934. • The securities are exempt from registration under the Securi- ties Act of 1933. • The securities serve as eligible collateral for Federal Reserve Banks’ discount loans. • The secretary of the Treasury approves the issuance of their securities. • The Federal Reserve is the fiscal agent for those issuances. • Their obligations are eligible for unlimited investments by national banks and state bank members of the Federal Reserve as well as by federally insured thrifts. The Congressional Budget Office estimates that the credit enhancement stemming from the government links is at least $6.5 billion annually. Only part of that subsidy is passed on to the homebuyers—about $4.4 billion—with the remainder of the subsidy pocketed by private shareholders and the corporations’ executives and lobbyists. CBO estimates that about 40 percent of Fannie and Freddie’s earnings in 1995 could be traced to the benefits of their government-sponsored status. Obviously a significant part of the subsidy derived from GSE status is being used, not to assist homebuyers, but to increase corporate power and control over all facets of the mortgage business. As Tom Stanton, a Washington attorney and long-time observer of GSEs, says, “Thanks to their special federal charters, Fannie and Freddie form a duopoly in the mortgage market . . . and they decide for RALPH NADER 113 themselves how much of the government subsidy they will give to shareholders and how much they will pass on in service to public purposes.” One of the prime rationales offered for the continuing subsidy is the need to have Fannie and Freddie provide support for affordable housing, one of the nation’s critical needs. But questions are increasing about how well Fannie and Freddie are performing this part of their mission. As the Urban Institute said in a 1999 study, “The GSEs appear to be lagging the market in servicing low and moderate income and minority borrowers.” William Apgar, assistant secretary of housing and urban development in the Clinton administration, put this issue succinctly: “There are two clear and indisputable facts. . . . One, the number of Americans in need of affordable housing stands at an all-time high. Two, Fannie and Freddie are making record profits.” In defending their special status, Fannie and Freddie have drawn heavily from the political ranks to bolster their lobbying and top executive jobs. In recent years the roster has looked like a virtual who’s who of Washington’s power elite. Nowhere is this more ap- parent than in the top job at Fannie Mae. The present occupant of the CEO’s chair at Fannie is Franklin Raines, who has moved smoothly through the revolving doors between high-level government jobs and Fannie Mae. From 1991 to 1996 Raines was vice- chairman of Fannie Mae. Then he decided to move over to a key job in the Clinton administration as the director of the Office of Man- agement and Budget. Now he has pushed the revolving doors again and has moved back to Fannie’s Wisconsin Avenue offices, this time as the chief executive officer. Raines succeeded James Johnson, an old hand in Democratic politics in Minnesota as well as nationally. He was the executive assistant to Vice President Walter Mondale and was the campaign manager for Mondale’s failed bid for the presidency in 1984. In the event that Raines or Johnson lacks the name of a key Democratic operative on their Rolodex, they can always turn to Jamie Gorelick, who joined Fannie in 1997 as vice-chairman after a stint as the deputy attorney general in the Clinton administration’s Justice Department. Staffing at the top levels of Fannie could leave the impression that the corporation is a hangout for Democrats. Not true. The tough bipartisan power game puts a high priority on finding former political officeholders and Congressional staff of both parties to fill executive jobs as well as staff, lobbying, and board positions. Charles Lewis, the executive director of the Center for Public Integrity, says 114 INFLUENCING THE POLITICAL PROCESS that Freddie and Fannie’s hiring of high-level government officials blurs the line between public service and private enterprise. The hirings are an example, Lewis says, “of how incestuous these quasi- governmental entities are with the people who ostensibly regulate or have oversight over them.” Fannie Mae’s board of directors has always included notable Washington insiders from both political parties who know the political power game well. The current board includes Kenneth Duberstein, the chief of staff for President Ronald Reagan and a long- time lobbyist for Fannie; Ann McLaughlin, the labor secretary in the Reagan administration; and Jack Quinn, the chief of staff to Vice President Al Gore and later the counsel to President Bill Clinton. In recent years Fannie’s staff has included Robert Zoellick, an executive vice president, who had been the deputy chief of staff in the George Bush White House and later a deputy assistant secretary of Treasury under James Baker; John Buckley, the senior vice president for Fannie Mae’s communications, who had been a top aide in Senator Robert Dole’s presidential campaign; Ellen Seidman, the senior vice president of Fannie, who later became an economic policy advisor to President Clinton and then headed the Office of Thrift Supervision; Mary Cannon, a Fannie vice president, who moved to the GSE after serving as an assistant secretary of the Department of Housing and Urban Development under Secretary Jack Kemp; Thomas Donilon, the Fannie general counsel, who was the chief of staff to Secretary of State Warren Christopher; William Maloni, a senior vice president, who had been a top aide to a senior Demo- cratic member of the House Banking Committee and the head of congressional liaison for the Federal Reserve Board; Gerald McMurray, a senior vice president and formerly the staff director of the House Banking Committee’s Housing Subcommittee; and Arnie Christensen, vice president for regulatory affairs, who was a top aide to Speaker of the House Newt Gingrich. Such in-house political expertise is extremely valuable, but both Fannie and Freddie also maintain cadres of full-time lobbyists who watch every move on Capitol Hill and in the executive branch that might have even a minuscule effect on the fortunes of the GSEs. The Center for Responsive Politics tabulated lobbying expenses of $5,550,000 for Fannie Mae in 1998. Freddie, according to the center, incurred lobbying expenditures of $2,160,000 that year. Fannie hired more than forty outside lobbyists to augment its staff of nine full-time lobbyists. Freddie had six in-house full-time RALPH NADER 115 lobbyists—including Ira Paull from the staff of the Senate Banking Committee—plus a dozen outside lobbyists drawn from law and consulting firms. The biggest chunk of Freddie’s lobbying fund— $320,000—went to Clark & Weinstock, a firm that includes former Republican congressman Vin Weber from Minnesota. Freddie also had on its lobbying tab the former Comptroller of the Currency and assistant secretary of the Treasury, Jim Smith, from the Nixon administration, and former Democratic congressman Richard Lehman of California, an alumnus of the House Banking Committee. Among the political luminaries on Fannie’s lobbying team were former senators John C. Culver of Iowa and Steve Symms of Idaho along with former congressmen Tom Downey and Rod Chandler. The list included Nicholas Calio, President George H. W. Bush’s congressional liaison; Michael Boland, former aide to Senator Trent Lott; and Ken Duberstein, President Bush’s chief of staff who now doubles as a member of Fannie’s board. The 1998 lobbying numbers may pale alongside lobbying to- tals for 1999. The formation of FM Watch, the financial industry group trying to curb the GSEs, has brought a flurry of new lobbyist hires by Fannie Mae. Among them are Williams and Jensen, as well as Griffin, Johnson Dover and Stewart, two big Washington lobbying firms. Freddie Mac has recently added former Speaker of the House Newt Gingrich and Kimberly Kranys Cobb, former counsel to the Republican Conference and aide to Senator Connie Mack, as consultants.

Playing the Tough Lobbying Game

As another indicator of how tough Fannie plays the lobbying game, the Wall Street Journal reported that Fannie Mae has been pressuring lobbying firms and trade groups to steer clear of FM Watch. Accord- ing to the Journal, Fannie’s communications director, John Buckley, conceded that the GSE had approached lobbying firms with this message: “We said to a number of our friends: You don’t want to be on the side of the bad guys.” Pressure tactics are nothing new in Fannie’s lobbying manual. In 1999 Legal Times reporter Sam Loewenberg quoted three unnamed former and current public officials as saying that Fannie has gone so far as to try to get uncooperative Hill staffers and regulators fired. “They will try to hurt an individual—a member of Congress, a staff member—to make them regret their actions,” one official said. 116 INFLUENCING THE POLITICAL PROCESS

Fannie and Freddie back up their lobbying with significant campaign donations, particularly in the form of soft money to both major political parties. While the top executives at Fannie Mae have extensive links to the Democratic Party and to the Clinton administration, the soft money donations show a decided leaning to the Republican campaign apparatus. Compilations by the Center for Responsive Politics show $501,350 in soft money donations by Fannie Mae in the 1997–1998 election cycle: $312,250 (62 percent) to the Re- publicans and $189,100 (38 percent) to the Democrats. Freddie Mac contributed $875,000 to the two parties in the form of soft money: $650,500 (74 percent) to the Republicans and $225,000 (26 percent) to the Democrats. In addition to those contributions, top executives and employees of Fannie and Freddie are individual donors to various members of Congress. For example, seven different contributions from employees of Fannie and Freddie ranging from $250 to the individual maximum of $1,000 appear on the 1997–1998 campaign reports of Representative John LaFalce, the ranking Democrat on the House Banking Committee, and a dozen different employees of the two corporations made similar contributions to Representative Rick Lazio, the Republican chairman of the Banking Committee’s Hous- ing Subcommittee. The Fannie Mae and Freddie Mac foundations—which award grants and loans for a variety of purposes—are clearly part of the public relations framework of the two corporations. The foundations are tightly linked to the corporations with the chief executive officers—Leland Brendsel of Freddie Mac and Frank Raines of Fannie Mae—serving as chairmen of their respective foundations. The Freddie Mac Foundation intended to distribute $17 million in grants in 1999, part of the $50 million since the foundation was formed in 1991. Much of the foundation’s outlay targeted programs serving children and youth. The Fannie Mae Foundation in 1998 listed $33 million in grants to 1,100 national and local organizations that it says are “working to transform communities.” Many grants over the past several years have gone to advocacy and research groups involved in housing and related areas—groups such as the Center for Community Change, Enterprise Foundation, ACORN, Greenlining, and National Low Income Housing Coalition. But the list contains a wide variety including conservative organizations with close corporate ties such as the Heritage Foundation, Alexis de Tocqueville Institution, and Manhattan Institute. RALPH NADER 117

The degree to which these foundation grants are part of a corporate strategy to enhance Fannie’s image or to gain political support or to mute criticism is difficult to pin down. But in some cases an organization receiving a grant has been asked or signaled to lobby for Fannie’s interests. We need to know more about the quid pro quo expected from the grantees. In one such incident in 1999 the Office of Management and Bud- get proposed that Fannie Mae pay fees for the registration of their securities—a cost borne by other publicly traded corporations. A Fannie Mae executive contacted a major recipient of Fannie Foun- dation money and insisted that the organization lobby the Clinton administration to drop the fee and even provided a script to guide the lobbying effort. In that same lobbying effort, according to the Wall Street Journal, Fannie Mae “enlisted more than two dozen congressmen and big-city mayors” to protest the fee. In the end the administration caved in to Fannie’s pressure. A Fannie executive bragged that it had taken the corporation only fifty hours to scuttle the proposal.

Fannie Mae Road Shows

Fannie Mae invests heavily in road shows with elaborate press conferences to announce its local housing efforts. Front and center in these shows are the local congressman, the mayor, and other assorted dignitaries. Not only does Fannie gets its name out front, but local politicians are overjoyed with their front-page pictures and television coverage. The shows are an ongoing effort that brings feel-good messages about Fannie to cities across the nation. A typical ceremony in Indianapolis in March 1998 paraded Mayor Steve Goldsmith before cameras to announce “a $2.1 billion homeownership investment” in Indianapolis by Fannie Mae. The mayor stood behind a Fannie Mae lectern; overhead was a huge ban- ner reading “HOUSE INDIANA—1-800 FANNIE.” As part of the ceremony Fannie announced that it would open an Indiana Partner- ship Office in Indianapolis within sixty days—“an important demonstration of the strength of Fannie Mae’s commitment to the city,” the mayor said. The Congressional Budget Office in its 1996 report on GSEs said the principal function of the partnership offices was to “enhance Fannie Mae’s political base.” CBO quoted one of Fannie’s executives saying that “for a relatively small investment, Fannie Mae will 118 INFLUENCING THE POLITICAL PROCESS be recognized as a force for good in each of the cities or states . . . by so doing Fannie will have 25 networks of support.” Fannie is busily expanding this network. The twenty-five offices reported by CBO in 1996 have increased to thirty-eight, with three in the planning stage.

Conclusion

Fannie Mae’s political muscle, its tough handling of its adversaries, and its aggressive public relations campaigns might be dismissed as just another example of excess by a private corporation. But neither Fannie Mae nor Freddie Mac is an ordinary corporation. The overwhelming political power of these GSEs raises serious questions about the willingness or the ability of the federal government to regulate them. The same is true of affordable housing goals, a critically important mission assigned the GSEs. Can these goals be set high enough and enforced with the vigor required to catch up with the needs of low- and moderate-income families for housing? Are the complaints from private companies valid that the GSEs are using their powers to reach beyond their charter to create unfair competition in the mortgage markets? Could Fannie and Freddie’s functions be conducted without the present level of subsidies? Could Fannie and Freddie be required to pass on more of the subsidy to homebuyers rather than diverting it for lucrative executive salaries, lobbying, and political activities? In 1997 James Johnson, then Fannie Mae’s chairman, received $5,441,232 in salary, bonuses, stock options, and other compensation. Lawrence Small, the president and CEO, received compensation of $2,948,751; Jamie Gorelick was the recipient of $1,850,993 in 1997 after joining Fannie as its vice-chairman in May 1997. Can Congress summon the courage to address these issues, to revisit the GSE charters, and to reform the present GSE system? Or have Fannie and Freddie built a political–public relations wall so strong that Congress will not even dare try to breach it? Some organizations opposing the expansion of Fannie and Freddie argue for tighter restrictions in the GSEs’ charters so that they cannot extend their activities beyond the residential mortgage market. Such tightening would enable the two to continue to perform their secondary market function but would prevent them from competing with banks and other companies that do not enjoy the same financial and other advantages. The implementation of such a policy would entail the drafting of some restrictive charter provisions and laws as well as the administration of those changes through regulations and administrative actions. As John Weicher points out, the history of these efforts is not encouraging. Unlike most organizations chartered to carry out a government mission, Fannie and Freddie’s independent obligations to their shareholders require them to maximize profits and to limit the effect of restrictions on their operations. This situation makes them extremely difficult to control through the usual mechanism of law and regulation. Weicher details governmental efforts to gain control over the policies pursued by Fannie and Freddie.

119 7 Setting GSE Policy through Charters, Laws, and Regulations

John C. Weicher

This chapter addresses the following question: Given the political will of Congress and the administration, can charters, laws, and regulations effectively control the scope of the GSEs’ activities? Assum- ing that policymakers do want to limit the GSEs, can they enforce their policy decisions? In practice the answer to this question may seem self-evident. If Congress and the administration have the political will, then charter violations or charter stretching can be addressed politically or administratively. If Congress lacks the political will, then even an independent regulator will be limited, at least in the long run and perhaps much sooner. Even assuming the political will, which I do assume for most of this essay, issues remain regarding whether and how that political will can be made effective. When issues arise, someone must act, even given the political will to limit GSE activities. Congress may not be in a position to act, purely because of timing concerns with its schedules and agendas within a two-year legislative cycle. But some entity must interpret the charter and the laws between congressional directives. Accord- ingly, Congress has delegated regulatory responsibility to executive

John C. Weicher was the director of urban policy studies at the Hudson Institute when he wrote this chapter. He is now the assistant secretary for housing and the commissioner of the Federal Housing Authority, Department of Housing and Urban Development. The views expressed in this chapter are not necessarily those of either the Hudson Institute or HUD. Editor’s note: Since this chapter was written in May 2000, the proposed capital standards discussed on pages 121–23 have been issued as a final rule, in July 2001. The rule requires 567 pages of the Federal Register.

120 JOHN C. WEICHER 121 branch agencies: to HUD and the Office of Federal Housing Enter- prise Oversight since 1992. Thus, in the first instance, effective policy concerns charters and laws but leads to questions of the efficacy of regulation. Although this chapter focuses on the program activities of the GSEs and the question of approval of new programs, relevant examples from other GSE policy issues shed some light on how—and how not—to set policy effectively. These are (1) capital standards and (2) housing goals—requirements to purchase mortgages on low- and moderate-income housing and other categories of special concern to Congress. After this discussion, the chapter considers new program approval and related issues.

The Interaction between Legislation and Regulation

The 1992 GSE regulatory legislation—the Federal Housing Enter- prises Financial Safety and Soundness Act of 1992, Title XIII of the Housing and Community Development Act of 1992— includes extensive language on both capital standards and housing goals. But Congress took rather different approaches— nearly opposite ones— to these two issues. It tried to write a precise capital standard into law, specifying the economic scenarios that the GSEs should be able to survive. Congress listed its categorical objectives for housing goals but with one exception left HUD as the regulator to establish both qualitative and quantitative standards for meeting those objectives. The regulatory processes put in motion have been quite different, with outcomes perhaps opposite to expectations. OFHEO took six years to formulate capital standards as proposed regulations. HUD promulgated final regulations on housing goals within three years after establishing interim goals within a year. Certainly other factors contribute to the different schedules be- sides the specificity of the 1992 legislation. OFHEO, as a new agency, was trying to set up shop in a new administration. And establishing capital standards is inherently harder than setting housing goals. But the 1992 legislation contributed significantly as well.

Capital Standards. The capital standard is a stress test. Each GSE is supposed to have enough capital to withstand specified adverse economic conditions. The legislation is quite explicit about those economic circumstances. It addresses both interest rate risk and credit risk. The specified interest rate change is a 600 basis point increase, or decrease, in the Treasury ten-year rate occurring within a year; 122 SETTING GSE POLICY the rate then remains unchanged for nine years. The specified scenario for credit risk is a rate of mortgage default loss and severity equal to the worst historical regional experience in the United States over a two-year period; the region must include at least 5 percent of the population. The changes are assumed to occur simultaneously. The legislation raises a variety of problems. A combination of interest rate changes and credit losses may not be at all consistent in terms of economic analysis. Some pattern of interest rate fluctuation accompanies any given credit risk situation. A more reasonable standard would include an actual historical experience and would require the GSEs to withstand the interest rate and credit conditions that occurred at the time, rather than choosing separate standards and combining them. In addition the scenario for credit risk had to be defined in terms of the GSEs’ experience; their relevant data extend only from 1978, and their worst regional experience occurred in the early 1980s, when interest rates rose sharply and put Fannie Mae in a negative net worth position for several years. Ten years later Congress was overly concerned with interest rate risk because it was still fighting “the last war.” The resulting capital standard proposed by OFHEO in 1998 seems to be a scenario with house price declines, mortgage default rates, and losses that are lower than those in the stress tests of the private rating agencies. The mandate to use regional experience has constrained OFHEO because the rating agencies based their stress tests on individual metropolitan areas. The 1992 legislation limited OFHEO’s discretion in other ways. It specified a 30 percent add-on to the stress-test capital requirement to cover management and operations risk. It also specified the mortgage product classifications to be used in determining credit risk. The legislation required OFHEO to propose a capital standard within eighteen months of its own establishment, which would have meant by December 1994. When 1997 came around and no capital standard had been proposed, Congress directed the General Account- ing Office to assess OFHEO’s capacity to fulfill its mission. GAO reported that one reason for delay was “the complexity of the development process as specified in the act” (GAO 1997, 7) and referred particularly to the stress test, both the individual credit and interest rate components and their combination, and the classification system of mortgage products. GAO did not limit the causes for the delay solely to the complexity of the legislated stress test, by any means; it noted the problems of starting a new agency. In some respects JOHN C. WEICHER 123

GAO suggested that Congress left too much discretion to OFHEO, for example, in identifying the region for the credit risk component of the stress test. (The benchmark region turned out to be Oklahoma, Louisiana, Arkansas, and Mississippi in 1983–1984; when the legislation was written, the general expectation had been Texas, probably in 1982–1983.) Republican members of the House Banking Committee objected to the stress test in the House version of the legislation on the grounds that it was too specific—to a degree “not desirable or workable in a law” (Wylie et al. 1991, 109). When passed, the legislation did not include the same capital standard, but the House version was, if anything, less prescriptive. This episode demonstrates that an econometric model cannot be written into legislation very well. Clearly Congress should have been thinking more broadly about the type of economic situation that it wanted the GSEs to withstand. Then Congress needed to give the researchers at the regulatory agency enough leeway to specify the details in a coherent way instead of hardwiring the stress test.

Housing Goals. The 1992 legislation established three types of housing goals: low- and moderate-income families; underserved areas, defined as central cities and rural areas; and a special affordable category for low-income housing in low-income areas and for very low-income families. At least 1 percent of mortgage purchases should meet the special affordable housing goal. The HUD secretary is required to establish specific goals in these categories. The legislation also established two-year transition goals of 30 percent for each of the first two categories, with dollar amounts of $2 billion and $1.5 billion for Fannie Mae and Freddie Mac, respectively (on the basis of 1990 or 1991 activity, this amounted to less than 0.5 percent for each GSE). Some of these terms have precise meanings in housing legislation. The Cranston-Gonzalez National Affordable Housing Act of 1990 defined low income as being below 80 percent of an area’s median income (either metropolitan area or nonmetropolitan county) and very low income as below 50 percent of an area’s median income. In a perhaps more familiar context, the poverty line is about 30 to 35 percent of the national median income: low income is about 250 percent of the poverty line and very low income is about 150 percent of it. The designations however are necessarily imprecise because poverty is determined nationally rather than locally. Moderate income 124 SETTING GSE POLICY does not have a precise meaning; prior to the 1990 act, the term low and moderate income was defined for FHA mortgage insurance in terms of the mortgage amount rather than family income. Two of these goals are not specific in stating what the GSEs must do. They do not specify exactly whom the GSEs should serve but establish income and geographic categories that are of special public policy concern. Neither do they specify a numerical goal. Within the legislative framework HUD has been able to develop more detailed goals by regulation. It has defined low and moderate income as income below the area’s median income. It has decided to define underserved areas within cities, rather than as cities as a whole, by census tract on the basis of income and minority population (so that, for example, Chevy Chase within the District of Columbia does not qualify). Underserved areas have census tract median income of no more than 90 percent of area median income or minority population of at least 30 percent and tract median income no more than 120 percent of area median income. A similar definition is used for rural areas (nonmetropolitan counties). The county median income must be no more than 95 percent of the state or national nonmetropolitan median, or minorities must comprise at least 30 percent of the population and the county median income be no more than 120 percent of the state nonmetropolitan median income. HUD has set numerical goals for these categories as a percentage of GSE mortgage purchases or securitizations based on analyses of their importance in the conventional conforming mortgage market. For 1995–1997 HUD calculated the low- and moderate-income share at 57 percent, the underserved area share at 33 percent, and the special affordable housing share at 29 percent. These numbers all exceed the transition goals established in the 1992 legislation. They also exceed HUD’s estimates in 1995, when it first established numerical goals after the expiration of the transition goals. HUD is proposing to raise the goals; its March 2000 proposed regulation for the years 2001–2003 sets them at 50 percent, 31 percent, and 20 percent, respectively (HUD 2000). These goals have not been established without controversy, including objections from the GSEs in their comments on the proposed rule in 1995. But the regulator had the authority to establish under- standable goals within the legislative framework and to modify them as market conditions changed. From these two experiences it seems reasonable to conclude that the GSEs can be constrained to serve public policy purposes, given JOHN C. WEICHER 125 the political will to do so. But the foregoing discussion concerns only the process of setting the goals, not whether the goals themselves are meaningful. That issue is beyond the scope of this chapter. Here it may be noted that despite HUD’s efforts the GSEs are not particularly active in serving the target populations. Their market share is about 20–25 percent of the mortgages in these various categories, and the loans that they buy are not particularly low–down-payment mortgages; this situation suggests that they are not particularly increasing homeownership in America because they are not particularly serving families on the margin of homeownership (Weicher 1999, 17–18). The gestation period of the housing goals is also important. The 1992 legislation and the 1995 HUD regulation were a long time coming. They are the outcome—at least the current outcome—of a policy debate that started in the late 1970s. In 1978 HUD Secretary Patricia Roberts Harris was concerned that Fannie Mae was not doing enough to promote national housing policy goals. HUD issued proposed regulations mandating Fannie Mae to direct 30 percent of its purchases to low- and moderate- income housing and 30 percent to properties in central cities. HUD cited section 309(h) of the FNMA charter act as authority for its actions: “The Secretary may require that a reasonable portion of the corporation’s mortgage purchases be related to the national goal of providing adequate housing for low and moderate income families, but with reasonable economic return to the corporation.”1 HUD also cited section 302(b), requiring the approval of the secretary to deal in conventional mortgages. In the ensuing political brouhaha (including 1,217 negative comments on the proposed regulations and 16 positive comments), Secretary Harris retreated; HUD established goals and insisted on its right to regulate, but the goals did not have any real force (HUD 1987, 162–66). The issue was essentially dormant until 1989. A 1985 GAO report concluded that HUD was not monitoring Fannie Mae’s low- and moderate-income housing performance (GAO 1985, ch. 6), but no immediate political action occurred. The next development came when the Financial Institutions Reform, Recovery, and Enforcement Act (FIRREA) changed the charter act to add a public purpose concerning low- and moderate-income housing. Before 1989 the charter act authorized Fannie Mae to “provide special assistance . . . for the financing of (1) selected types of home mortgages (pending the establishment of their marketability) originated under special hous- 126 SETTING GSE POLICY ing programs designed to provide housing of acceptable standards . . . for segments of the national population which are unable to obtain adequate housing under established home financing programs.” The language meant subsidized housing, produced under several programs between 1961 and 1974. FIRREA eliminated this purpose and substituted “provide ongoing assistance to the secondary market for home mortgages (including mortgages secur- ing housing for low- and moderate-income families involving a reasonable economic return).” FIRREA also changed the Federal Home Loan Mortgage Corporation Act so that the language applied to Freddie Mac as well. The issue attracted substantial attention in the policy discussions leading up to the 1992 legislation. Secretary Harris had established the low- and moderate-income goal in terms of the purchase price of the house, not the income of the buyer. Price and mortgage data were more readily available, and HUD assumed that price and income were fairly highly correlated, although data from the American Housing Survey could have been used for a serious analysis of the relationship even in 1978. As the legislative discussions focused on affordable housing goals, HUD’s analysis of AHS data showed that fewer than half the buyers of homes that met the 1978 regulatory price standard actually had low or moderate incomes (incomes below the area median). About half the homes in the conventional conforming market met the 1978 standard; only about 30 percent met the income criterion; only about 20 percent met both. This contributed to the legislative decision to set goals based on income and to leave their implementation to the regulator. The 1992 legislation also modified the language concerning public purpose. It changed home to residential and rewrote the par- enthetical language as “(including activities related to mortgages on housing for low- and moderate-income families involving a reasonable economic return that may be less than the return earned on other activities).” It also added a new purpose: to “promote access to mortgage credit throughout the Nation (including central cities, rural areas, and underserved areas).” These changes in the charter acts appear to be reinforcing the housing goals legislation by telling the GSEs to reach further down in the income distribution. The chronology covers about fifteen years. It suggests a further qualification to the conclusion that the GSEs can be constrained to serve public policy purposes, given the political will to do so. Again, the process is not easy—and it takes a long time. JOHN C. WEICHER 127

New Program Activities

The approval of new programs presents a situation that is the opposite of meeting housing goals. The policy question is typically whether the GSEs should be prevented from doing something rather than whether they should be required to do something. The competitive advantage that the GSEs derive from the privileges given to them in their charters motivates this area of policy. The issue also has a long history. When Fannie Mae was reconstituted in 1968 as a private corporation, it was limited to government- guaranteed mortgages, as it had been since its creation in 1938. Two years later, when the Emergency Housing Finance Act of 1970 created Freddie Mac with the authority to buy conventional mortgages, section 302 (b)(2) was added to the FNMA charter act and gave Fannie Mae authority to “purchase, service, sell, lend on the security of, or otherwise deal in” conventional mortgages—subject to the approval of the secretary of HUD. Parallel language covering home improvement loans and manufactured housing loans was added to the charter act in 1978 and 1980, respectively, as sections 302(b)(3) and 302(b)(4). The sections are in addition to the general regulatory authority of the HUD secretary contained in section 309(h). In 1984 the secretary was required to approve or disallow a new program request within forty-five days (sixty days if the secretary requested additional information from Fannie Mae). These provisions of the charter act were retained when FIRREA was enacted in 1989, and similar language was inserted in the corporation act. Freddie Mac’s previous programs to deal in conventional mortgages were grandfathered. The 1992 legislation modified the language in the acts. A new section 302(b)(6) replaced the approval requirements in the three sections of the charter act: Fannie Mae could not implement any new program before obtaining the approval of the HUD secretary. The legislation further specified that the HUD secretary should approve new programs proposed by either Fannie Mae or Freddie Mac, unless determining that the programs were not authorized by the charter or corporation acts or were not in the public interest.

Congress as New Program Regulator? Marvin Phaup of the CBO has argued that Congress is the effective regulator of program activity (Phaup 2000, 457–58). He points out that every proposed new program since 1992 has been approved by the HUD secretary. Con- 128 SETTING GSE POLICY gress, rather than HUD, effectively disallowed Fannie Mae’s Mort- gage Protection Plan. Fannie Mae would have provided cash-value life insurance and short-term disability and unemployment insurance to first-time homebuyers whose mortgages it purchased. Con- gress changed the law in 1997 to reduce the tax advantages of investing in cash-value life insurance, partly because of Fannie Mae’s plan. Fannie Mae then dropped the idea (Seiler 1999, 136–37). Phaup also points out that Congress forced Freddie Mac to abandon its investment in Philip Morris bonds. A more recent example also supports this hypothesis. Freddie Mac received statutory authority to go into the private mortgage insurance business late in 1998 as an amendment to the corporation act contained in legislation dealing with other issues; the amendment was not the subject of any congressional hearings. The authority was quickly repealed when Congress realized what it had done. This was purely a legislative issue; the HUD secretary as regulator had no role in it. The Clinton administration, however, did not favor the additional authority (McKinley 1999, 12). Although Phaup suggests that his generalization may apply only since 1992, earlier examples can be cited as well. One concerns second mortgages. In 1981 HUD Secretary Samuel Pierce allowed Fannie Mae to buy second mortgages on a temporary basis; with the juxta- position of a serious recession and high mortgage rates at that time, homebuyers were finding second mortgages necessary to complete purchases. When Fannie Mae requested a second extension of the authority in 1984, Secretary Pierce asked for public comments before making a decision. Congress overrode him by extending the authority for three years by statute and then made it permanent in 1988 (HUD 1987, 47–49, 171–73; Weicher 2000, 435–36). Real estate mortgage investment conduits provide another instance. The Tax Reform Act of 1986 gave the GSEs authority to issue REMICs subject to regulatory approval. After receiving approval from Treasury Secretary James Baker (who had the right to approve securities issuances but not new programs), Fannie Mae announced its intention to issue REMICs backed by conventional mortgages. Secretary Pierce claimed authority to determine whether Fannie Mae could issue such REMICs; Fannie Mae first rejected and then accepted his position. The HUD secretary gave conditional approval in 1987 and then gave permanent approval in 1988 under strong pressure from Congress, according to HUD staff (HUD 1987, 175–76). JOHN C. WEICHER 129

If Phaup is correct, then public policy can control the GSEs— not so much through their charters as through congressional acts. Further, what Congress appears to be saying is that the GSEs can extend their activities quite broadly in the mortgage market but perhaps not outside that market. In return “Congress wants assurances that it is not going to have to face the budgetary cost of a financial rescue of a GSE” (Phaup 2000, 458). Congress cannot monitor the budget situation on a continuing basis, and the legislature knows that, at least after the S&L debacle. Thus Congress leaves that determination to a regulator. But there are counterexamples. The clearest is HUD’s 1990 re- fusal to allow Fannie Mae to purchase debt obligations secured by conventional mortgages. HUD concluded that the requested authority was quite broad and would let Fannie Mae “engage in complex transactions significantly different from the examples given by F.N.M.A. in the documents submitted to HUD and from those engaged in previously” (HUD 1991a, 57). The authority would have allowed Fannie Mae to make advances against conventional mortgages, a type of activity limited to the Federal Home Loan Banks. After denying the request, HUD did not hear any particular objections from Congress. It might be argued that this case was special because of its timing—right after FIRREA, with the collapse of the S&Ls still front-page news and Congress unwilling to run appar- ent risks of further problems in the housing finance system. (HUD was also concerned that the program could significantly increase Fannie Mae’s financial risks.) Further, precedent existed from an earlier HUD ruling in 1985 that the charter act did not authorize Fannie Mae to buy mortgage-backed obligations issued by financing subsidiaries of financial institutions. For whatever reason, Con- gress did not override the program approval decision by the regulator (HUD 1991a, 56–57; Weicher 2000, 435–37). More recent evidence comes from an episode cited by Phaup: Freddie Mac’s purchase of Philip Morris corporate bonds. Congress did indeed object but acted in a way that suggested it saw a separate and important role for HUD. Representative James A. Leach, the chairman of the House Committee on Banking and Financial Services, asked GAO to investigate federal regulatory oversight of nonmortgage investments. GAO responded that HUD’s general regulatory authority allowed it to limit nonmortgage investments by the GSEs: “HUD has general regulatory authority over the housing enterprises. . . . HUD also has authority to approve new mort- 130 SETTING GSE POLICY gage programs that could contain nonmortgage investment components” (GAO 1998a, 7). Subsequently GAO responded to a further request from Representative Richard H. Baker, the chairman of the Subcommittee on Capital Markets, Securities, and Government-Spon- sored Enterprises, to review HUD’s implementation of its general regulatory and new program approval authorities (GAO 1998b, 7). Key congressional committee chairmen (both members of the Bank- ing Committee when the 1992 legislation was enacted) clearly did not think that Congress had de facto reserved new program approval authority to itself. Nor did GAO think so. (The substance of the issue was the ability of the GSEs to generate arbitrage profits by investing in corporate securities as a benefit from their agency status. There was particular concern over their investments in long- term securities—defined as longer than five years—that were dubi- ously related to their need for short-term liquidity. Making nonmortgage investments is indeed one way for the GSEs to continue their rapid growth.) Other episodes indicate the difficulty of congressional regulation. Activities can evade the congressional radar screen. Congress and the Clinton administration made clear in 1998–1999 that they did not want the GSEs in the business of private mortgage insurance. But while Freddie Mac’s authority was being repealed, Fannie Mae was putting in place something that looks quite a bit like mortgage insurance as it reduced the required insurance coverage if mortgages went through their automated underwriting system and if the homebuyer paid a fee. Freddie Mac put a similar program in place several months later (McKinley 1999, 13–14). Congress has not raised the same concerns over this activity.

Ancillary Services. The GSEs can work their way around the program limitations in the charter acts. In 1994 the GSEs started to lease their proprietary automated underwriting systems to lenders, although the charter acts prohibit them from originating mortgages. “The question of whether the provision of the charter limiting origination has been violated depends on what the definition of ‘originate’ is” (McKinley 1999, 11). Origination could be narrowly defined as simply the funding of a loan; it could be broadly defined as including the underwriting process. Clearly the GSEs have opted for the narrow limitation. Robin Seiler has pointed out that the GSEs are also supplying “credit reports, property valuations, title insurance and other origination-related services to lenders; providing JOHN C. WEICHER 131 funds to lenders immediately after loan papers are signed; and brokering the sale of the rights to service conforming mortgages” (Seiler 1999, 139). The activities make a good start on a comprehensive list of origination and settlement services. As McKinley observes, the GSEs “can draw much of the profitability” from the broadly defined origination process. How do these activities comport with the charter acts? Seiler (1999) says that “automated underwriting and other financial services provided to lenders are not new lines of business for Fannie Mae and Freddie Mac” as of 1998. The first stage in the process of moving into ancillary services was apparently Fannie Mae’s purchase of an underwriting software firm in 1985. Fannie Mae did not seek regulatory approval but justified its action under section 309(a) of the charter act, authorizing it to “do all things as are necessary or incidental to the proper management of its affairs and the proper conduct of its business.” Fannie Mae’s chairman, David Maxwell, also justified the purchase as “part of a larger role that Fannie Mae has always been expected to play in carrying out our Charter Act mandate” (HUD 1987, 46–47). The apparently long development stage before Fannie Mae could offer automated underwriting seems to have blunted criticism of its actions. But it is certainly doubtful that Congress intended—in 1985 or later—for the GSEs to move into the origination and settlement services markets.

Incremental Program Growth. The GSEs have often created new programs in a series of small steps. Several temporary or pilot programs turned into full-scale ones, such as the program for second mortgages. The special circumstances that led Secretary Pierce to approve a temporary program no longer applied in 1984, when he sought public comment on Fannie Mae’s request for a second extension, but the program was established enough for Congress to make it permanent anyway. In a similar situation in 1991, Secretary Kemp approved Fannie Mae’s pilot program to buy construction loans. Because FIRREA had limited the ability of the S&Ls to make these loans, builders were looking for alternative sources of funds. Secre- tary Kemp then allowed a broader program in 1992. The move into ancillary services is another instance, although in this case no external event created unusual, severe problems for participants in the housing or mortgage markets. The most striking and most far-reaching example is the Rom- ney letter. As mentioned, the 1970 legislation that allowed Fannie 132 SETTING GSE POLICY

Mae to buy conventional mortgages also made that authority subject to the approval of the HUD secretary. In 1971 Secretary George Romney gave Fannie Mae approval for a limited program of purchasing conventional first mortgages on a trial or demonstration basis. Fannie Mae has since argued that the letter permanently authorized it to purchase any kind of conventional first mortgage. The GSE has claimed that the Romney letter also applied to conventional mortgage instruments that were of minimal importance in the mortgage market in 1971, such as balloon mortgages, or even undreamed of at the time, regardless of the financial risks that the alternative instruments might pose. Indeed in support of its assertion that Sec- retary Patricia Roberts Harris could not establish housing goals, Fannie Mae claimed in 1978 that the Romney letter exhausted HUD’s regulatory authority over conventional mortgage programs (HUD 1987, 165). The Romney letter is a kind of special case of a GSE creating a new program in small steps. In that instance it took a small step and then argued some years afterward that the small step was really a giant step.

Parsing the Charter and Corporation Acts. Originate is not the only word with varying interpretations in the context of public policy toward the GSEs. The GSEs have had terminological arguments with their regulators on various occasions. The most noteworthy have concerned the language about the public purposes of the GSEs: what they are allowed to do and what they are expected to do. From its reconstitution in 1968 until FIRREA, Fannie Mae had two statutory public purposes: to provide supplementary assistance to the secondary market for home mortgages and to provide special assistance for selected mortgages (meaning mortgages on privately owned, federally subsidized lower-income housing) or during a housing or national recession. FIRREA replaced these with three new purposes: to provide stability in the secondary market for home mortgages, to respond appropriately to the private capital market, and to provide ongoing assistance to the secondary market for home mortgages by increasing the liquidity of mortgage investments, including mortgages for housing for low- and moderate-income families. FIRREA also amended the corporation act to assign the same purposes to Freddie Mac. Controversy over the first of these purposes—both what it means and whether the GSEs have fulfilled it—has occurred over at least JOHN C. WEICHER 133 the past twenty years. Both the pre-FIRREA and post-FIRREA first- purpose statements have commonly been interpreted by housing market analysts and policymakers to refer to countercyclical mortgage market activity; indeed that activity was part of the reason for its creation, and the countercyclical purpose was restated in the 1954 charter act as well as later. The research evidence is mixed: Fannie Mae may have acted countercyclically to a modest extent through the late 1970s but not during the double recessions of 1980 and 1981– 1982, while Freddie Mac did not behave particularly countercyclically during its first two decades. Freddie Mac had no specific public purpose to do so in its corporation act until FIRREA, but the S&Ls (which owned it) certainly expected the GSEs to help them during recessions. During the 1990–1991 recession both GSEs were active mortgage buyers—although, since this period coincided with the post-FIRREA structural changes in the mortgage market, how much of the GSE growth was countercyclical and how much was secular is not clear. Their share of the market continued to grow during the recent long economic expansion. The occurrence of only two recessionary periods in the past quarter-century may indicate that the public purpose is becoming less important. But the purpose remains extremely relevant to the capital requirements of the GSEs. How much capital they need to withstand an economic downturn depends not only on the severity of the downturn but on how they react to it. The timing of their reaction is especially important. As part of its regulatory activities between 1989 and 1992, HUD performed stress tests to estimate the amount of capital needed by the GSEs to withstand an economic downturn. Those analyses became part of the policy debate over the appropriate capital standard in the 1992 legislation. HUD used a scenario from the Great Depres- sion, modeled on a scenario developed and used by Moody’s to analyze the capital adequacy of private mortgage insurance companies. The analysis concluded that both GSEs could survive ten years of the depression scenario if they stopped doing business immediately at the beginning of the downturn, but neither could survive as long as seven years if they continued to do business for as long as two years after the downturn began (HUD 1991b, ch. 2; 1992a, ch. 3; 1992b, ch. 3). The significance of those results is that ten years qualifies a company for an Aaa rating, and seven years for an Aa rating. The GSEs argued that because they would stop doing business as soon as a downturn started, they could thus survive ten years of the depres- 134 SETTING GSE POLICY sion scenario and therefore did not need additional capital. Even if they could somehow identify the onset of the depression and immediately stop buying mortgages, such behavior would be inconsis- tent with their charter acts. They would hardly be providing stability in the secondary market for residential mortgages. In other respects the language of the charter acts raises broad questions about GSE activities. The second public purpose is “to respond appropriately to the private capital market.” One might question whether an appropriate response is to dominate the mortgage market to the extent of absorbing 90 percent of the conventional conforming market by 2003 (Wallison and Ely 2000). In addition, one might question how expanding into ancillary services such as origination and settlement could serve any legislated public purposes or whether providing such services is necessary to the proper conduct of the GSEs’ business. The charter acts also state that within the conventional mortgage market the GSEs are expected to restrict themselves to “mortgages which are deemed by the corporation to be of such quality, type, and class as to meet, generally, the purchase standards improved by private institutional mortgage investors.” In the mid-1990s both GSEs began to buy and securitize subprime mortgages—loans to borrowers with impaired credit. The subprime market has grown dramatically since FIRREA, with most growth coming in refinancing or making new first mortgages for current homeowners. The loans have markedly higher delinquency and default rates than prime mortgages do, and perhaps one-third to two-thirds are for the purpose of debt consolidation because the rates for even the riski- est category of subprime mortgages are lower than the rates on credit card balances. By entering the subprime market, the GSEs take greater risks than previously and in effect become consumer lenders (Weicher 1997; Seiler 1999, 140–41; McKinley 1999, 8–10). The GSEs frequently describe their powers in terms that bear no close relationship to the charter acts. They tend to refer to their mission—a not particularly clearly defined concept—rather than their purposes. The statement of congressional findings in the 1992 legislation (section 1302) reflects the GSE position: they “have important public missions that are reflected in the statutes and charter Acts establishing the [Federal Home Loan] Banks and the enterprises.” But the statement does not specify what missions, and the relevant laws do not define the term elsewhere. A vague mission is much easier to use as justification for GSE activities than more specific purposes, JOHN C. WEICHER 135 but with sufficient ingenuity the interpretation of the purposes can be stretched if necessary.

Conclusion

Constraining the GSEs on paper—by charter, law, and regulation— is clearly difficult. Some examples demonstrate where regulators have limited them by following the procedures specified in law. By and large, however, enforcing public policy decisions is a story of frequent skirmishes and tortured legal and lexicographical arguments. (The peculiar bifurcated regulatory structure process does not help enforcement, but that is a different subject.) Making the case that Congress has chosen itself to regulate the GSEs’ program activities is possible. The corollary of that proposi- tion is that the charters and laws have been effective means of setting policy for the GSEs. But both arguments are overstated. First, Congress is quite likely to make the basic decisions about the role of GSEs, as it did about REMICs and second mortgages. But many issues regarding new program activity are on a much smaller scale, and many are fairly arcane, as the private mortgage insurance issue illustrates. Also, the mortgage market like many markets is changing under the impact of new information technology, and the legislative process is a slow and awkward way to regulate the GSEs as these changes occur. Second, Congress is always concerned about the availability of enough mortgage credit for housing. This concern kept it from ad- dressing the S&L problems from the mid-1970s on, until those problems became too obvious to overlook and demanded urgent attention. Congress may therefore be chary of limiting the GSEs’ ability to provide mortgage money. That said, Congress probably would not favor a monopoly or duopoly structure in the mortgage market or favor GSE dominance of mortgage originations if it were confronted with those issues in so many words. But the issues typically do not confront Congress in those terms. Both Fannie Mae and Freddie Mac “were privatized almost by hap- penstance” in response to immediate political exigencies that are now irrelevant and long forgotten (Weicher 1999, 18). The same can be said about some of their expansions into new program areas. Policymakers reacted to a situation, such as a recession or a spike in interest rates. The constant in all these situations has been the GSEs’ financial interest (both their managers’ and their stockholders’ in- 136 SETTING GSE POLICY terest) in extending their exploitation of their government guarantees. That abiding interest confronts any public policy approach that attempts to constrain the GSEs to a particular set of activities and limits its effectiveness.

Note

1. Quotations and citations from the Federal National Mortgage Associa- tion Charter Act and the Federal Home Loan Mortgage Corporation Act are taken from the 1984, 1994, and 1998 compendiums of Basic Laws on Housing and Community Development prepared by the Office of Legislative Counsel of the U.S. House of Representatives.

References

McKinley, Vern. 1999. “’Expanding the Empires: What Are the Housing GSEs Doing Today, and How Do These Activities Serve Their Public Purposes?” Paper presented at conference, “Fannie Mae and Freddie Mac: Public Purposes and Private Interests,” American Enterprise Institute, Washing- ton, D.C. Phaup, Marvin. 2000. “Comment on ‘Regulating Government-Sponsored Enterprises.’” In Restructuring Regulation and Financial Institutions, edited by James R. Barth, R. Dan Brumbaugh, Jr., and Glenn Yago, pp. 455–58. Santa Monica: Milken Institute Press. Seiler, Robert S., Jr. 1999. “Fannie Mae and Freddie Mac as Investor-Owned Public Utilities.” Journal of Public Budgeting, Accounting, and Financial Management 11(1) (spring): 117–54. U.S. Department of Housing and Urban Development. 1987. 1986 Report to Congress on the Federal National Mortgage Association. Washington, D.C.: Government Printing Office. ———. 1991a. 1990 Report to Congress on the Federal National Mortgage Asso- ciation. Washington, D.C.: Government Printing Office. ———. 1991b. Capitalization Study of the Federal National Mortgage Associa- tion and the Federal Home Loan Mortgage Corporation. Washington, D.C.: Government Printing Office. ———. 1992a. 1991 Report to Congress on the Federal Home Loan Mortgage Cor- poration. Washington, D.C.: Government Printing Office. ———. 1992b. 1991 Report to Congress on the Federal National Mortgage Asso- ciation. Washington, D.C.: Government Printing Office. ———. 2000. “HUD’s Regulation of the Federal National Mortgage Asso- ciation (Fannie Mae) and the Federal Home Loan Mortgage Corporation (Freddie Mac).” 24 CFR Part 81. Federal Register, March 9: 12631–816. JOHN C. WEICHER 137

U.S. General Accounting Office. 1985. The Federal National Mortgage Associa- tion in a Changing Environment. Report GAO/RCED -85-102.Washington, D.C.: Government Printing Office. ———. 1997. Federal Housing Enterprises: OFHEO Faces Challenges in Imple- menting a Comprehensive Oversight Program. Report GAO/GGD-98-6. Wash- ington, D.C.: Government Printing Office. ———. 1998a. Government-Sponsored Enterprises: Federal Oversight Needed for Nonmortgage Investments. Report GAO/GGD-98-48. Washington, D.C.: Government Printing Office. ———. 1998b. Federal Housing Enterprises: HUD’s Mission Oversight Needs to Be Strengthened. Report GAO/GGD-98-173. Washington, D.C.: Govern- ment Printing Office. U.S. House of Representatives, Committee on Banking, Finance and Urban Affairs, Subcommittee on Housing and Community Development. 1984, 1994, and 1998. Basic Laws on Housing and Committee Development. Wash- ington, D.C.: Government Printing Office. Wallison, Peter J., and Bert Ely. 2000. Nationalizing Mortgage Risk: The Growth of Fannie Mae and Freddie Mac. Washington, D.C.: AEI Press. Weicher, John C. 1997. The Home Equity Lending Industry: Refinancing Mort- gages for Borrowers with Impaired Credit. Indianapolis: Hudson Institute. ———. 1999. “The Development of the Housing GSEs.” Paper presented at conference, “Fannie Mae and Freddie Mac: Public Purposes and Private Interests,” American Enterprise Institute, Washington, D.C. ———. 2000. “Regulating Government-Sponsored Enterprises.” In Restruc- turing Regulation and Financial Institutions, edited by James R. Barth, R. Dan Brumbaugh, Jr., and Glenn Yago, pp. 413–52. Santa Monica: Milken Institute Press. Wylie, Chalmers P., Marge Roukema, Doug Bereuter, Tom Ridge, Cliff Sterns, Paul Gillmor, Bill Paxon, John J. Duncan Jr., Mel Hancock, Frank Riggs, Dick Armey, Craig Thomas, and Sam Johnson. 1991. “Additional View on H.R. 2900.” In Government-Sponsored Housing Enterprises Financial Safety and Soundness Act of 1991: Report. Report 102-206, U.S. House of Repre- sentatives, pp. 109–10. Washington, D.C.: Government Printing Office. Among the many proposals for dealing with Fannie and Freddie over the long term, several would not change their charters but would attempt to reduce the risks that they pose to the taxpayers by gaining some control over their expansion, creating incentives for them to reduce their risk taking, or compensating the government for the risks assumed in continuing its support. Ron Feldman explores several of these ideas. His analysis suggests that some proposals—requiring the GSEs to get regular credit ratings, putting their debt “on budget,” and charging them a guaranty fee—would be difficult to implement. The fundamental problem is the unique nature of the GSE structure itself; the necessary information to implement these proposals is difficult to obtain or to compare with similar information from other financial institutions that do not enjoy government support. Accord- ingly, while these proposals have some value, they are distinctly subopti- mal. Further, the Feldman analysis suggests that those who wish to gain greater control over the risks that the GSEs create for the federal government might do better to look to more fundamental solutions, such as full privatization.

138 8 Improving Control over Fannie Mae and Freddie Mac

Ron Feldman

Analysts have presented several reasons why taxpayers and policymakers should worry about Fannie Mae and Freddie Mac, the two largest government-sponsored enterprises. The concerns include (1) excessive risk taking by Fannie and Freddie with the specter of future taxpayer bailouts, (2) excessive growth of Fannie and Freddie in their current markets and potential entry into new ones where they will have an unfair advantage, and (3) an excessive and ineffec- tive subsidy for the two firms. In this brief essay I argue that these concerns arise because Fannie and Freddie are out of control. By out of control I mean that Fannie and Freddie have only weak constraints on their actions. Policymakers designed the two to avoid the response systems that most firms and government agencies face and which produce effective feedback leading to a desirable level of activity. When considering policy toward Fannie and Freddie, focusing on the absence of effective constraints offers at least three important benefits. First, this perspective highlights the fact that Fannie and Freddie, rather than policymakers, exercise significant control over their public subsidy. For the most part the two entities determine the subsidy’s size and the timing of its use. Policymakers cannot act as effective stewards over public resources under these conditions. Second, the perspective clarifies the persistent nature of the problems with Fannie and Freddie. How Fannie and Freddie manifest their out-of-control behavior changes over time. Critics have given more attention recently to subsidized expansion than to solvency

The views expressed here are those of the author and not necessarily those of the Federal Reserve Bank of Minneapolis and the Federal Reserve System.

139 140 IMPROVING CONTROL risk. However, policy concerns in some form will exist as long as the two are GSEs. Third, reforms that address the disease rather than the symptoms present more effective and more attractive options to policymakers. Certainly each of the three concerns mentioned has a strong intellectual foundation and has drawn attention to Fannie and Freddie. But the three may not present as compelling a case for reform to policymakers as the underlying, long-term problem identified here. Policymakers should have an intuitive grasp of how a system encouraging excessive activity produces undesired results. All these implications support enactment of GSE reforms that address the underlying and persistent lack of constraints on Fannie and Freddie. This chapter reviews several of the most frequently suggested reforms for Fannie and Freddie to determine if they would constrain Fannie and Freddie’s actions. All reforms would prove quite difficult to implement effectively even though they have attractive attributes in theory. Nevertheless, reforms would raise the cost of GSE status and would thus encourage Fannie and Freddie to relinquish their charters.

Flaws in the Fannie Mae and Freddie Mac System

Commercial firms clearly operate in complex systems where they face competition that influences and constrains their actions. A firm taking on additional risk, for example, must pay higher costs for funds as the chance of default rises. The higher costs will constrain the firm’s growth. At some point the risks incurred by the firm may prevent additional funding. In another familiar story a firm’s entry into new markets will attract competitors whose supply of similar products will drive down profits and will limit expansion. Reliance on such mechanisms to allocate resources among competitive uses reflects confidence that such systems (or invisible hands) tend to maximize society’s welfare. Society has granted special power for the allocation of resources to government agencies. While the market system does not effectively constrain federal agencies, policymakers have created other limiting systems. Most important, the government budget process forces policymakers to trade off the growth of one program against the alternative use of those resources. Policymakers curtail the growth of a program when its expansion denies resources to others that they support. As Herb Stein (1997) once observed about rising government spending, “If something cannot go on forever, it will stop.” RON FELDMAN 141

Fannie and Freddie have attributes of both private firms and government agencies and should face constraints from both markets and government systems. But in fact neither controls them. More specifically, debt holders, competitors, and policymakers do not effectively control Fannie and Freddie. Government support relieves Fannie and Freddie of the market discipline of fixed-income creditors that have an implied protection from loss. The absence of market discipline manifests itself in both price and quantity. In terms of price, creditors demand a yield on Fannie and Freddie debt that approximates the risk-free rate rather than the higher rate that they would charge a firm with a risk of default similar to Fannie and Freddie’s absent implied federal support. In terms of quantity, the credit support provided to Fannie and Freddie and the Treasury-like characteristics of their borrowing instruments give the two the ability to raise funds in amounts and at times when non-GSEs with similar risks cannot (the liquidity of Fannie and Freddie debt in turn reduces the required yield). When Fannie Mae appeared insolvent on a mark-to-market basis, it still borrowed large amounts of funds at rates only 100 basis points above other GSEs (CBO 1991, 10). Creditor discipline against Fannie and Freddie will arise if creditors perceive gaps in the implied support that the government offers the GSEs. Moreover, Fannie and Freddie do face the disciplining force of their equity holders. The ownership of stock in the firms by senior Fannie and Freddie management creates additional incentives to limit risk taking. The restraining forces play a particularly important role in periods such as the current one, when a GSE charter generates significant profits. But other incentives can play an important role, especially when charter value falls. In particular a well-diversified equity holder might benefit more from additional risk that a firm assumed in comparison with fixed-income creditors that received only their promised principal and interest. Thus equity-holder discipline is not a perfect substitute for fixed-income holder discipline. The incentive of equity holders, including managers, may be to take on excessive risk when the value of equity falls. Only with additional risk taking can they hope to recover the value of their shares. Further, Fannie and Freddie do not appear to face effective competition from non-GSE firms. Though limited in scope, the available empirical evidence suggests that Fannie and Freddie exercise duopoly power (Hermalin and Jaffee 1996). Market power, according to these analyses, comes from their financial advantage based 142 IMPROVING CONTROL on GSE status. Some market observers share this view. One equity analyst recently argued that Fannie and Freddie would be “giving up the source of value” of the companies if they voluntarily shed GSE status (Barta 2000, C1). Indeed the market value of the firms often falls when reforms to curtail their GSE status gain support. The GSEs can use their subsidized status to expand their core business, make investments in new products, and gain market share beyond the potential in an unsubsidized environment. Because their subsidy grows when they issue more debt and because a lack of competition allows the two to retain some of that subsidy, Fannie and Freddie have an unusual incentive to raise funds and expand. And to the degree to which such expansion and subsidy exploitation increase equity value, the alignment of the interest of management with shareholders could encourage, rather than constrain, troubling behavior. In addition Fannie and Freddie do not face the potential tradeoffs and resulting financial constraints from inclusion in the federal budget process (Feldman and Koppell 1999). More generally, the annual budget process provides a forum for congressmen to question the representatives of federal agencies and to gather detailed information such as previous performance, allocation of public resources, current activities, and future objectives. The process also provides an opportunity to clarify—in both a formal and informal manner—congressional preferences regarding the activities of the agency under review. Moreover, those involved with the budget process can conduct additional monitoring or review with minimal additional cost after this initial investment. Finally, the opaque subsidy transmission embedded in the GSE structure and the complex activities of the GSEs complicate an outside evaluation of Fannie and Freddie’s actions. Taxpayers, for example, cannot easily determine how GSEs use their public credit support and thus have limited opportunity to advocate restrictions on Fannie and Freddie’s activities. The opacity of the firms also ham- pers regulation and supervision necessary to constrain their behavior effectively.

Implications of the Flawed System for Policy

Several closely related implications follow from the observation that Fannie and Freddie do not face the same level of constraint as do the vast majority of government agencies and private firms. In short, RON FELDMAN 143 policymakers have a strong rationale for enacting reforms to address the underlying and persistent absence of constraints that permits Fannie and Freddie to control public resources.

The Disease and Its Symptoms. The absence of a robust feedback control system for GSEs manifests itself in three ways that are normally identified as separate and distinct concerns. First, the lack of effective constraints from creditors creates the problem of moral hazard and promotes excessive risk taking. Second, the absence of effective constraints from policymakers contributes to a poorly targeted subsidy. Third, the absence of constraints in competitive markets because of Fannie and Freddie’s subsidized status leads to concerns about the inferior allocation of resources and unfair expansion. The nonexistence of competition allows Fannie and Freddie to retain more subsidy than they might otherwise. The GSEs’ retention of their subsidy has led to criticism of the effectiveness of the transmission method for this subsidy. Being outside the federal budget process invites setting subsidies at excessive levels. Focusing on the underlying concern suggests that policymakers look to the fundamental, weak systems of feedback and constraint when enacting legislation.

Control of the Subsidy. The absence of effective constraints almost allows the GSEs to determine the size of their own subsidy. The subsidy arises through an implied guarantee. Policymakers have for- feited direct control over the subsidy. The methods of estimating the subsidy used by the Congressional Budget Office and the Treasury confirm that Fannie and Freddie control the size of the subsidy by altering the level of risk that they assume as well as the volume of the debt that they issue subject to the implied guarantee. Specifically CBO and the Treasury calculate the size of the subsidy by estimating the yield that Fannie and Freddie would have to pay on their debt and mortgage-backed securities (MBSs) if the two did not have GSE status. To help determine the non-GSE yield, analysts look to debt and MBSs issued by non-GSEs similar in structure and in other respects to Fannie and Freddie. To generate an estimate of GSE’s subsidy, the difference between the yield paid by Fannie and Freddie on MBSs and debt and that paid by non-GSE issuers is multiplied by the amount of outstanding MBSs and debt. A critical part of the comparison involves assessing the credit risk that Fannie and Freddie debt and MBSs would have absent GSE status and then determining the firms’ credit rating without federal support. The 144 IMPROVING CONTROL lower the credit rating and the greater the amount of debt and MBSs issued, the greater the subsidy is. Fannie and Freddie essentially choose the amount of risk that they assume and the volume of their debt issuance and thus control the size of the subsidy. That said, the GSEs’ charters provide at least some restrictions on their activities (because of legal restrictions and the incentive to Fannie and Freddie to maintain the value of the charters). However, concerns about the GSEs’ recent expansive activities imply that the charters—or at least the enforcement of the charters—may not se- verely limit the range of their activities. A second restriction comes from government regulation and supervision. Regulation is a cum- bersome tool to manage the risk taking of firms as complex and dy- namic as Fannie and Freddie. In contrast the capital rule developed by Fannie and Freddie’s financial regulator, the Office of Federal Housing Enterprise Over- sight, offers a more sophisticated method for setting Fannie and Freddie’s capital and in theory managing its risk taking. However, it is not clear that the capital rule transfers control of the subsidy back to policymakers. Consider a period when the GSEs have a subsidy of approximately several billion dollars. In this example OFHEO applies its capital rule and finds that one GSE does not have to hold any additional capital while the other need only engage in additional de minimis hedging to meet the capital rule. In other words the entities can both meet the capital rule and receive a subsidy of an amount roughly equal to what they received before the capital rule existed. One would be hard pressed to argue that such a capital rule has given additional control to policymakers; this hypothetical example approximates events over the past several years. The capital rule cannot perform tasks for which it was not designed. The current capital rule was designed to ensure that Fannie and Freddie could pass a specific type of financial test and was not created to control the subsidy. Unsurprisingly the capital rule does not require much change from the current regime, given observations about its formulation: “The stress test was not performed before the bill was passed. Instead, it was negotiated between the Bush Administration and the GSEs and written into law by Congress without analysis of how much capital will be required to meet it” (Weicher 1994, 76, 57). Fannie and Freddie’s control over their subsidy should provide additional support for reforms to constrain their behavior. Policymakers are trustees of the public’s resources: some of the re- RON FELDMAN 145 sources go to Fannie and Freddie. How can the public expect policymakers to carry out their stewardship when the entity receiving the subsidy has greater control over its size than the policymaker?

Persistence of the Problem. Congress shaped the GSEs as a combination of a public agency and a private firm. The combination inherently reduces the constraints that either system normally produces. As a result, problems with Fannie and Freddie cannot be dismissed as one-time events but instead must be treated with reforms that take into account this endemic feature.

How Might Reforms Address the Root Problem?

Identifying the root problem provides a rationale for reforming current GSE policy. It also provides a common and logical standard against which to evaluate several of the most commonly suggested GSE reforms. How well do each of the reforms effectively constrain GSE behavior? I first briefly describe a reform and its potential for constraining Fannie and Freddie’s activities. The reforms include (1) requiring Fannie and Freddie to obtain a credit rating, (2) charging the GSEs a cost of capital or guarantee fee, (3) placing the GSEs in the federal budget process, and (4) restricting the GSEs’ traditional activities. The next section reviews the significant problems in implementing reform.

Required Credit Ratings. Under most versions of this reform, the federal government would contract with well-known firms to issue a credit rating for Fannie and Freddie. Both the Treasury and OFHEO contracted with Standard & Poor’s to provide such a credit rating, most recently in 1997. S&P considers their output a “risk to the government.” The firm supposedly evaluates the “risk that FN or FM will become financially troubled and require federal financial assistance” (S&P 1997). S&P assigns the rating from its conventional issuer rating system (for example, AAA, AA). Recent legislation sponsored by Congressman Richard Baker, the chairman of the House banking committee (now the Committee on Financial Services), would require an annual assessment by two credit rating agencies; the Treasury has supported this requirement.1 Congress could implement the reform in two ways. The credit rating could serve as an information-only item. Alternatively Con- gress could link the rating to restrictions on activities (for example, 146 IMPROVING CONTROL no new, net debt issuance without maintenance of an AA or higher rating) or regulatory action. Credit ratings linked to some restriction offer a superficially attractive method for constraining the risk taking and unfair expansion of Fannie and Freddie. The method is a straightforward way to limit the amount of risk that the GSEs take and in part the amount of their subsidy. Estimates of the subsidy depended on the probable credit rating of the two sans GSE status. The reform would have credibility with policymakers because credit ratings are ubiquitous in financial markets and have become increasingly accepted in bank regulatory settings (for example, capital rules for securitization and financial modernization).

Cost of Capital–Guarantee Fee. Analysts have suggested levying a fee against Fannie and Freddie to compensate the government for their use of public credit (see, for example, Miles 1995). In theory the reform could effectively restrict the risk taking of Fannie and Freddie and their subsidized expansion by pricing GSE risk taking on the margin. That is, Fannie and Freddie would face a series of prices that would increase as the amount of their risk taking increased or as the volume of their debt issuance (and subsidy) increased. The higher prices would effectively act as a form of feedback control. By reducing volume and risk taking, the fee could also reduce the GSEs’ subsidy and thus address concerns about subsidized expansion. In reality, however, proposed methods of levying the fee are much less sensitive. Rather than reflecting a price that would shift with GSE risk taking, the fee would normally be a fixed amount over time and would be levied on the basis of the amount of debt outstanding, for example. Although such a fee would have a limited ability to effect risk taking on an ex ante basis, it could still reduce the GSEs’ subsidy. The fee system could move that amount closer to the optimal level and could alter expansion efforts.

Inclusion in the Federal Budget. Inclusion in the federal budget would require an estimate of the GSEs’ subsidy or the expected loss that they would impose on the government. As with the credit-rating system, Congress could include the estimate in the budget as an information-only item. Alternatively the estimate could become part of the budget process. Mimicking the system established by the Credit Reform Act of 1990, Congress could require an appropriation equal to the estimate (Kane 1996). RON FELDMAN 147

Either as information or as a part of the budget process, the reform would differ from the current treatment. The federal budget should analyze the GSEs’ financial condition and the exposure to risk that they create for taxpayers. However, the quality of such analysis has varied a great deal over the 1990s. Earlier budgets, such as those prepared for fiscal year 1992 or 1993, had extensive reviews. The fiscal year 2000 budget does not provide an empirical estimate of potential costs, does not comment in detail on Fannie and Freddie’s risks, and relies on unedited financial data and self-description from the GSEs. Including the estimate in the budget process would directly address the issue of missing constraints on Fannie and Freddie. Assuming that inclusion in the budget process would require Con- gress to balance support of GSEs with other priorities, the reform does not establish any direct limits on Fannie and Freddie. Such scrutiny of a GSE subsidy should have implications for the amount of their subsidy and thus for their risk taking and subsidized expansion.

Limits on Traditional Activities. A wide range of reforms would limit the traditional activities of Fannie and Freddie in the mortgage markets. Examples include limiting the amount of mortgages that the GSEs can finance (for example, by lowering the conforming limit) and restricting the options for financing their activities (for example, by limiting their ability to finance themselves with debt). The reforms could directly affect Fannie and Freddie’s actions, but the actual outcome is unclear. How would these reforms actually limit risk taking or the generation of subsidies? The inability to determine how reforms would actually constrain Fannie and Freddie’s activities foreshadows a more general concern.

Roadblocks to Implementation

All reforms discussed above have some potential to constrain the activities of Fannie and Freddie; in some cases the reforms might substitute—at least in theory—for the usual systems of feedback. That said, policymakers would face steep challenges in implementing the reforms. Moreover, the rationale for implementing these reforms as opposed to taking more direct steps toward privatization, for example, is not clear. 148 IMPROVING CONTROL

The difficulty in implementing the reforms arises from the same factor that reduced the effective constraints on Fannie and Freddie in the first place, the GSE structure. All reforms require information on the amount of risk that Fannie and Freddie assume and the resulting subsidy. Yet estimating the subsidy presents significant practical problems. Consider the comparison method. Because the GSEs can issue debt in amounts and structure (for example, callable or without a senior; subordinated structure) that other firms cannot, analysts cannot find debt or MBSs issued by non-GSE firms readily comparable with GSE debt or MBSs. The difficulty does not imply that CBO and the Treasury used the wrong methodology or that their estimates were incorrect. Rather the GSE structure means that such estimates always come with an inherently high degree of uncertainty. More generally the GSE activities are not transparent to outsid- ers. The GSEs’ amount of risk and the subsidy depend in part on their ability to hedge their increasingly large balance-sheet mortgage portfolio. Such an assessment requires a significant amount of private information. Even when Fannie and Freddie provide analysts with sufficient data to begin an estimate of risk taking or the subsidy, the output from the analysis comes on an ex post basis. Some reforms rely on fairly exact and timely measures of Fannie and Freddie’s risk taking, such as the guarantee fee, or on good point estimates of the subsidy, such as the budget option, to constrain GSE action effectively. Clearly the limitations outlined above indicate that these reforms would be difficult to implement in an effective manner. The practical limitations also reduce the potential benefits of requiring credit ratings. As a general rule, credit ratings are known to lag behind the financial condition of the rated firm and the market’s assessment of risk. The problem is especially acute for an opaque financial firm (CBO 1991, 50–55; Morgan 1999). Such uncertainties diminish the ability of policymakers to support these reforms. Policymakers should support reforms that more effectively constrain Fannie and Freddie’s activities on the grounds that the reforms improve the stewardship of public resources. But policymakers would naturally want to understand how these reforms would affect the purchase of a home. After all, Congress views Fannie and Freddie as a tool for directing subsidies to those borrowing funds for homeownership. Legislators may not want to support reforms with uncertain methods to affect the total subsidy and the amount passed through Fannie and Freddie to taxpayers. RON FELDMAN 149

A second challenge of implementation concerns unintended consequences. Many reforms seek to constrain Fannie and Freddie’s behavior by reducing their subsidy. Some analysts believe that much of the GSEs’ value derives from their special status and subsidy. Subsidy reduction diminishes the value of the GSEs and thereby gives management a strong reason to increase risk to save their net worth. Finally, support for constraining the GSEs’ actions and exerting greater policy control over their subsidy may suggest a course of action different from these reforms. Congress designed GSEs to receive difficult-to-control subsidies and to escape from the usual market and government constraints. Support for the reforms above would indicate a strong underlying concern with the structure of the GSE itself and reservations about GSE status. Congressional testimony by Thomas Woodward expressed these sentiments effectively:

The whole idea of granting privileges to the enterprise is to confer a benefit that helps the enterprise be successful. Con- sequently, requiring the firm to compensate the government for the privileges simply undoes the benefit. As a general rule, therefore, if a fee is justified, then it is difficult if not impossible to justify the special status and privileges in the first place. . . . Thus, the notion of paying for the privileges is inseparable from the issue of continuing GSE status . . . a “user fee” for GSE privileges is simply a substitute mechanism for ending GSE status. (Woodward 1997)

Privatization by a Thousand Cuts

The concerns expressed about Fannie and Freddie take different forms but ultimately arise from the lack of constraints that these two face because of their GSE status. This chapter reviewed several reforms for their ability to impose otherwise absent constraints. Al- though they could improve on the current condition, all reforms would face significant hurdles in implementation, again because of GSE status. The removal of GSE status represents the most logically coherent option for those concerned about the various manifesta- tions of Fannie and Freddie’s out-of-control behavior. Enactment of the more limited reforms discussed in this essay offers a transitional 150 IMPROVING CONTROL tool whose primary contribution may be to raise the costs of retaining GSE status. Through the proverbial thousand cuts, the reforms could cause Fannie and Freddie to abandon their GSE status.

Note

1. The two rating requirement assumedly reduces the potential for grade inflation although it is not clear that more ratings improve their quality. See Cantor and Packer 1996.

References

Barta, Patrick. 2000. “Freddie Mac, Battling a Chorus of Critics, Spurns a Proposal to Split Up Operations.” Wall Street Journal, June 30, C1. Cantor, Richard, and Frank Packer. 1996. “Multiple Ratings and Credit Stan- dards: Differences of Opinion in the Credit Rating Industry.” Federal Reserve Bank of New York Staff Report 12. Congressional Budget Office. 1991. Controlling the Risks of Government- Sponsored Enterprises. Washington, D.C.: Government Printing Office. Feldman, Ron, and Jonathan Koppell. 1999. “Congressional Oversight of Fannie Mae and Freddie Mac: Does Their GSE Structure Matter?” Paper presented at American Enterprise Institute conference, “Fannie Mae and Freddie Mac: Public Purposes and Private Interests,” Washington, D.C. Hermalin, Benjamin E., and Dwight M. Jaffee. 1996. “The Privatization of Fannie Mae and Freddie Mac: Implications for Mortgage Industry Struc- ture.” In Studies on Privatizing Fannie Mae and Freddie Mac, pp. 225–302. Washington, D.C.: Department of Housing and Urban Development. Kane, Edward J. 1996. “Statement of Shadow Financial Regulatory Commit- tee on Extending the Credit Reform Act to GSEs.” Statement 131. Febru- ary 12. Washington, D.C. Miles, Barbara L. 1995. “Proposed Offset Fees for Fannie Mae and Freddie Mac.” Congressional Research Service Memorandum to House Banking and Financial Services Committee, November 1. Morgan, Don. 1999. “Whether and Why Banks Are Opaque.” In Proceedings of the Federal Reserve Bank of Chicago’s Thirty-fifth Annual Conference on Bank Structure and Competition. Chicago: Federal Reserve Bank of Chicago. Standard & Poor’s. 1997. Final Report of Standard and Poor’s to the Office of Federal Housing Enterprise Oversight. Contract HEO9602C. RON FELDMAN 151

Stein, Herb. 1997. http://Slate.msn.com/ItSeemsToMe/97-05-15/ ItSeemsToMe.asp#Bio. Weicher, John C. 1994. “The New Structure of the Housing Finance System.” Federal Reserve Bank of St. Louis Review 76 (July–August): 58. Woodward, Thomas. 1997. Statement before the Subcommittee on Capital Markets, Securities, and Government-Sponsored Enterprises, Committee on Banking and Financial Services, and the Subcommittee on Govern- ment Management, Information, and Technology, Committee on Gov- ernment Reform and Oversight, U.S. House of Representatives, July 16. Many commentators have remarked on the fact that Fannie and Freddie— despite their government support and special advantages—contribute less toward the provision of low-income or affordable housing than ordinary banks do. Congress and the Department of Housing and Urban Develop- ment have endeavored to get the GSEs to pull their weight in this process—without much success. But the result stems naturally and inevitably from the GSE structure, which couples a government mission with an obligation to pursue maximum profitability. Unlike banks, which receive an explicit subsidy and are heavily regulated and supervised, Fannie and Freddie have been able to avoid strict regulation while exploiting an implicit subsidy. In the process they have been able to fulfill their obligation as publicly held companies to achieve high levels of profitability while pro- claiming that they have made major commitments to support affordable and low-income housing. The following essay by Jonathan Brown shows that Fannie and Freddie’s supposed commitments to affordable and low-income housing are almost entirely pretense. In case after case Brown details that the GSEs do not provide financing support for mortgages in low-income areas at anything like their rate of support for middle-income housing. And this is true even when the low-income mortgages qualify as conventional and conforming loans. The lack of such support significantly demonstrates the difficulty in forcing the GSEs to divert a portion of their subsidy for any purpose other than increasing shareholder values and rewarding management. That ori- entation is also a significant political fact because a substantial amount of the congressional support for Fannie and Freddie comes from representatives of minority and low-income communities, who either are not aware of the fact that their communities are not receiving housing support from Fannie and Freddie or do not choose to treat the fact as important.

152 9 Reform of GSE Housing Goals

Jonathan Brown

HUD’s proposed rules for housing goals for Fannie Mae and Freddie Mac and the many public announcements of Fannie Mae and Freddie Mac on their housing commitments have fostered the general impression that the goals have been a success, although some adjustments are needed to strengthen the goals. A perspective of local community reinvestment and civil rights, however, reveals major problems in the design and implementation of the GSE housing goals. The problems stem in large part from serious flaws in the structure of the goals. An extensive review of GSE performance in metropolitan statistical areas (MSAs) across the nation indicates that in many MSAs in 1998 the GSE market share of one- to four-family mortgage loans in low- and moderate income and minority neighborhoods was roughly half the GSE market share of such loans in upscale, nonminority neighborhoods. My analysis measured GSE market share as the ratio between Fannie Mae and Freddie Mac’s combined purchases of one- to four- family mortgages and the total number of conventional one- to four-family mortgage loan originations reported under the Home Mortgage Disclosure Act (HMDA) with loan amounts less than $227,000, the 1998 GSE ceiling for mortgages on one-unit homes. GSE purchases are based on the Census Tract File in the GSE Public Use Database. Because the file does not distinguish between home purchase loans and refinancing loans, I consolidated these two types of mortgage loans in computing GSE market shares. Implicit in the GSE concept of housing goals is the idea that the GSEs should at minimum support the mortgage market in low- and moderate-income and minority neighborhoods to the same extent that they support the mortgage market in affluent, nonminority neighborhoods. Clearly the minimum standard of support is not

153 154 REFORM OF HOUSING GOALS being met where the GSE market share of mortgage loans in low- and moderate-income or minority neighborhoods is much less than the GSE market share in affluent or nonminority neighborhoods. Similarly the GSEs fall short of the minimum standard when the GSE market share of mortgage loans to low- or moderate-income or minority borrowers is well below the GSE market share of loans to more affluent or nonminority borrowers. Figures 9–1 through 9–4 provide stark examples of how the GSE market share tends to decline sharply in inner-city neighborhoods. (The 1998 Chicago, Cleveland, and Akron maps were submitted to the Department of Housing and Urban Development according to the pending rulemaking on GSE housing goals by Illi- nois Association of Community Organizations for Reform Now and the Metropolitan Strategy Group of Ohio. The 1996 Chicago map was presented by Essential Information at a 1998 conference on the role of the GSEs.) The map’s diagonal overlay identifies minority census tracts. The Chicago map overlay identifies census tracts in which nonwhite and non-Asian persons are 70 percent or more of the census tract population. The Cleveland and Akron diagonal over- lays indicate census tracts in which minority persons are 40 percent or more of the census tract population. The maps strongly correlate the racial composition of census tracts and the GSE market share. Also comparison of the 1996 and 1998 Chicago maps shows that the overall pattern of GSE market share disparity did not change much during this two-year period. To provide further insight into the nature of GSE housing goal performance in the Chicago MSA , I prepared a set of statistical tables on 1998 GSE market shares by census tract race and census tract income categories and by borrower race and income categories. I based the tables on GSE loan purchase data and HMDA loan origination data for the nine-county Chicago primary metropolitan statistical area (PMSA). Table 9–1 cites GSE market share data by census tract racial category. The table shows a decline in GSE market share from 75.10 percent in predominantly nonminority census tracts to only 33.28 percent in census tracts in which minorities constitute 75 percent or more of the population. In the Chicago PMSA these two census tract categories account for almost 60 percent of the 1,776 census tracts. Table 9–2 reports GSE market share data by census tract income category and shows a decline in GSE market share from 77.54 percent in upper-income census tracts to only 39.01 percent in low-income JONATHAN BROWN 155

FIGURE 9–1 Fannie Mae and Freddie Mac Support for Conventional Home Mortgage Loan Market, Chicago Metropolitan Area, 1996

Fannie Mae and Freddie Mac market share of Diagonal overlay for minority neighborhoods conventional loans 1996, by census tract Minority census tract: African-American, Municipal boundaries: Chicago, Maywood, Riverdale, Harvey Hispanic, Native-American, and Other-Race Minority census tract persons (excluding White and Asian persons) No conventional loan originations comprise 75% or more of census tract population. 0–15% 15–30% Fannie Mae and Freddie Mac market share: Combined number of 30–45% home mortgage loans purchased by Fannie Mae and Freddie Mac 45–60% in 1996 as a percentage of total conventional home mortgage loan 60% or more originations in 1996.

Home mortgage loan: Home purchase loan or refinancing loan secured by a 1-4 unit dwelling (excluding loans above $227,000— the Fannie Mae and Freddie Mac loan-size ceiling in 1996).

Sources: HUD, 1996 GSE Single-Family Public Use Data; FFIEC, 1998 HMDA data; and 1990 U.S. Census data. 156 REFORM OF HOUSING GOALS

FIGURE 9–2 Fannie Mae and Freddie Mac Support for Conventional Home Mortgage Loan Market, Chicago Metropolitan Area, 1998

Fannie Mae and Freddie Mac market share of Diagonal overlay for minority neighborhoods conventional loans 1998, by census tract Minority census tract: African-American, Municipal boundaries: Chicago, Maywood, Riverdale, Harvey Hispanic, Native-American, and Other-Race Minority census tract persons (excluding White and Asian persons) No conventional loan originations comprise 75% or more of census tract population. 0–15% 15–30% Fannie Mae and Freddie Mac market share: Combined number of 30–45% home mortgage loans purchased by Fannie Mae and Freddie Mac 45–60% in 1998 as a percentage of total conventional home mortgage loan 60% or more originations in 1998.

Home mortgage loan: Home purchase loan or refinancing loan secured by a 1-4 unit dwelling (excluding loans above $227,000— the Fannie Mae and Freddie Mac loan-size ceiling in 1998).

Sources: HUD, 1996 GSE Single-Family Public Use Data; FFIEC, 1998 HMDA data; and 1990 U.S. Census data. JONATHAN BROWN 157

FIGURE 9–3 Fannie Mae and Freddie Mac Support for Conventional Loan Market, Cleveland Metropolitan Area, 1998

Percentage of conventional home mortgage loan originations in 1998 purchased by Fannie Mae and Freddie Mac, by census tract.

Minority population > 40%

0–20% 20.1–40% 40.1–60% 60.1–90% 80% or more Home mortgage loan: Home purchase loan or refinancing loan secured by a 1-4 family dwelling (excluding loans above $227,000—the Fannie Mae and Freddie Mac loan-size ceiling in 1998).

Sources: 1998 GSE Single-Family Public Use Data; 1998 HMDA data; and U.S. Census Bureau 1990. Prepared by Metropolitan Strategy Group.

census tracts. The census tract income-related decline in GSE market share in table 9–2 is similar to the census tract race-related decline in table 9–1 except that not as many census tracts have the lowest market share interval in the former. Generally one can obtain a sharper focus on GSE market share disparities by looking at census tract race rather than census tract income. Table 9–3 compares the GSE market share in minority census tracts and nonminority census tracts for separate census tract income intervals. While the GSE market share in nonminority census tracts ranges from 61.27 percent to 69.62 percent for a census tract within various moderate- and middle-income intervals, it never rises 158 REFORM OF HOUSING GOALS

FIGURE 9–4 Fannie Mae and Freddie Mac Support for Conventional Loan Market, Akron, Ohio, 1998

Percentage of conventional home mortgage loan originations in Minority population > 40% 1998 purchased by Fannie Mae and Freddie Mac, by census tract.

0–25% Home mortgage loan: Home purchase loan or refinancing loan secured by a 25.1–40% 1-4 family dwelling (excluding loans above $227,000—the Fannie Mae and Freddie Mac loan-size ceiling in 1998). 40.1–75% 75% or more

Sources: 1998 GSE Single-Family Public Use Data; 1998 HMDA data; and U.S. Census Bureau, 1990. Prepared by Metropolitan Strategy Group.

above 42.61 percent for minority census tracts in comparable census tract income intervals. Although such tract-based analysis does not by itself establish disparate treatment or disparate impact in violation of the federal discrimination laws, it does indicate major disparities between neighborhoods along racial lines in accessing the GSE-operated secondary mortgage market. The disparity needs to be addressed directly in implementing the GSE housing goals. JONATHAN BROWN 159

TABLE 9–1 GSE Market Share, by Census Tract Racial Category, Chicago PMSA, 1998 GSE Census Tract Minority Census Loan GSE Market Population (%) Tracts Originations Purchases Share 75 or more 472 23,616 7,860 33.28 50–75 141 12,546 6,846 54.57 30–50 182 20,194 12,825 63.51 20–30 146 24,887 16,971 68.19 10–20 267 66,815 47,842 71.60 Less than 10 550 167,043 125,446 75.10 Tracts with no population data 18 Total tracts 1,776

Table 9–4 provides GSE market share data by borrower income category. The table indicates that borrower income-related disparities in GSE market shares are not as great as neighborhood-related disparities. Further, the GSE market share disparity is substantial only for low-income borrowers but is not much of a factor for borrowers with incomes exceeding 80 percent of MSA median family income. Table 9–5 provides GSE market share data by borrower racial category and shows that the GSE market share is dramatically lower for African-American borrowers and moderately lower for Hispanic borrowers. The failure of the GSE housing goals to provide residents of many minority neighborhoods and low- and moderate-income neighborhoods and, more generally, minority borrowers and low- and moderate-income borrowers with the same access to the secondary mortgage market enjoyed by most Americans is rooted in several major structural flaws in 1992 GSE housing goal legislation. First, the legislation used unduly broad definitions to determine the focus of the GSE housing goals, with the result that they are not properly targeted on underserved sectors of the mortgage market. Much of this inadequate targeting or dilution is associated with misapplication of terms commonly used in the Community Rede- velopment Agency (CRA) and community development activities. 160 REFORM OF HOUSING GOALS

TABLE 9–2 GSE Market Share, by Census Tract Income Category, Chicago PMSA, 1998 GSE Census Tract MFI as a Census Loan GSE Market Percentage of MSA MFI Tracts Originations Purchases Share Less than 50 280 8,631 3,367 39.01 50–80 374 29,541 14,717 49.82 80–90 168 22,722 12,919 56.86 90–100 209 36,401 23,260 63.90 100–110 167 39,927 27,440 68.73 110–120 141 43,729 31,931 73.02 More than 120 409 133,944 103,858 77.54 Tracts with no income data 28 Total tracts 1,776

For purposes of the low- and moderate-income housing goal, low- and moderate-income borrowers are defined to include not only borrowers with incomes less than 80 percent of MSA median family income (as they are for CRA purposes) but also borrowers with incomes between 80 percent and 100 percent of MSA median family income. Yet table 9–4 shows that in the Chicago PMSA borrowers with incomes in the 80–100 percent range of MSA MFI (lower-middle- income borrowers) have a GSE market share (69.03 percent) almost the same as the average GSE market share for the Chicago PMSA as a whole (68.92 percent). Why are GSE housing goals needed to encourage GSE purchases from this middle subsector of the mortgage market? Moreover, if the GSE housing goals are successful in encouraging the GSEs to change underwriting and other policies to stimulate more GSE purchases of loans to low- and moderate-income borrowers, wouldn’t borrowers in the 80–100 percent range of MSA MFI also benefit from such changes? Similarly, for purposes of the special affordable housing goal, the GSE legislation defines low-income borrowers to mean borrowers with incomes less than 80 percent of MSA median family income rather than the 50 percent of MSA median family income standard used for CRA purposes. Yet, as table 9–4 shows, borrowers with incomes less than 50 percent of MSA mean family income have the JONATHAN BROWN 161

TABLE 9–3 GSE Market Share, Minority and Nonminority Census Tracts, Chicago PMSA, 1998 GSE Loan GSE Market Originations Purchases Share 280 tracts with MFI less than 50% of MSA MFI: 272 minority tracts 374 tracts with MFI of 50–80% of MSA MFI 253 minority tracts 18,484 7,845 42.44 121 nonminority tracts 11,057 6,872 62.15 168 tracts with MFI of 80–90% of MSA MFI 32 minority tracts 3,367 1,061 31.51 136 nonminority tracts 19,355 11,858 61.27 209 tracts with MFI of 90–100% of MSA MFI 33 minority tracts 3,422 1,458 42.61 176 nonminority tracts 32,979 21,802 66.11 167 tracts with MFI of 100–110% of MSA MFI 8 minority tracts 1,137 435 38.26 159 nonminority tracts 38,790 27,005 69.62

NOTE: Minority tract: minority population percentage of 50% or more. Nonminority tract: minority population percentage less than 50%.

most difficulty in accessing the GSE secondary market and clearly the most need for a “special affordable housing goal.” In addition the GSE legislation established a geographic GSE housing goal that included all central cities, rural areas, and other underserved areas as defined by HUD. Clearly this was an excessively broad definition of geographic areas with restricted access to the GSE-operated secondary market, although HUD has to some extent narrowed its scope by means of its regulatory authority to define underserved geographies. Second, the GSE legislation prohibits HUD from establishing enforceable subgoals for the low- and moderate-income housing goal and the geographic area goal. Subgoals are a logical tool to ensure 162 REFORM OF HOUSING GOALS

TABLE 9–4 GSE Market Share, by Borrower Income Category, Chicago PMSA, 1998 Loan GSE GSE Market Borrower Income Category Originations Purchases Share Low 20,851 10,631 50.99 Moderate 64,537 40,351 62.52 Lower-middle 51,595 35,614 69.03 Upper-middle 46,936 34,438 73.37 Upper 119,538 92,206 77.14 No borrower income 12,568 4,552 36.22 MSA total 316,025 217,792 68.92

NOTE: Low income: borrower income less than 50% of MSA MFI. Moderate income: borrower income greater than 50% but less than 80% of MSA MFI. Lower-middle income: borrower income greater than 80% but less than 100% of MSA MFI. Upper- middle income: borrower income greater than 100% but less than 120% of MSA MFI. Upper income: borrower income greater than 120% of MSA MFI.

that the GSEs adequately consider the most underserved sectors of the mortgage market. For example, census tracts with minority population percentages exceeding 50 percent, or even 75 percent, would be good candidates for an enforceable subgoal. Third, and perhaps most important of all, the GSE legislation failed to establish separate GSE housing goals (or sets of housing goals) for one- to four-family mortgage loans and multifamily mortgage loans. As a result, in implementing all three statutory housing goals, HUD has devised an accounting system that measures the number of housing units financed by GSE loan purchases and combines one- to four-family units and multifamily units. This consolidation of one- to four-family and multifamily units has significantly diluted the impact of the special affordable housing goal, the one statutory GSE housing goal that made some effort to achieve proper targeting. Because a high percentage of multifamily housing loans qualify for the special affordable housing goal, the GSEs can significantly reduce their obligation to purchase one- to four-family housing loans under the special affordable housing goal by increasing their pur- JONATHAN BROWN 163

TABLE 9–5 GSE Market Share, by Borrower Racial Category, Chicago PMSA, 1998 Loan GSE GSE Market Borrower Racial Category Originations Purchases Share White 227,212 165,569 72.87 African-American 24,521 8,556 34.89 Hispanic 21,312 13,016 61.07 Asian 12,448 9,562 76.82 Targeted minority 49,809 23,696 47.57 No race 26,556 18,965 71.42 MSA total 316,025 217,792 68.92

NOTE: Targeted minority: excludes white and Asian borrowers.

chases of multifamily mortgage loans. According to HUD roughly 50 percent of the multifamily mortgage loan purchases of the GSEs qualify for the special affordable housing goal. In 1997, multifamily loans accounted for 44 percent of Fannie Mae’s special affordable housing goal units and 31 percent of Freddie Mac’s. In short the commingling of one- to four-family housing loans and multifamily housing loans within the same special basket enables the GSEs to trade off performance on the one- to four-family side against performance on the multifamily side. Reinforcing these statutory weaknesses, HUD set the special affordable housing goal too low in the 1995 GSE rulemaking. Al- though HUD estimated that such units were 20–23 percent of the market, it set the goal for 1997–1998 at only 14 percent of GSE loan purchases measured on a housing unit basis. Improving the working of the GSE housing goals requires several basic reforms. First, in the one- to four-family side of the mortgage market, GSE market shares should define underserved sectors. Such market share analysis should be efficiently employed to identify underserved census tract categories, borrower income categories, and borrower racial categories. Such GSE market share data were not available to Congress in 1992 when it enacted the GSE housing goal legislation. Now that the data are available, Congress should 164 REFORM OF HOUSING GOALS examine the information closely and restructure the statutory GSE housing goals to provide for better targeting. In implementing the GSE housing goal legislation, HUD has generally relied on loan origination and loan denial rates derived from HMDA data to identify underserved segments of the mortgage market. While the approach has some value, GSE market share data provide a much sharper lens for identifying market segments that have restricted access to the GSE-sponsored secondary mortgage market. HUD should use this approach in its current rulemaking to bring as much targeting to the GSE housing goal rules as possible within the statutory framework. Second, the legislation on GSE housing goals needs to be re- structured to provide separate housing goals or sets of housing goals for one- to four-family housing and multifamily housing. The reform is needed to prevent the GSEs from trading off improved housing goal performance on the multifamily side for reduced housing goal obligation on the one- to four-family side. Most of this trade-off takes place within the category of the special affordable housing goal. Multifamily housing loans and one- to four-family housing loans targeted toward underserved sectors are the most important elements of any strategy to address affordable housing needs. GSE housing goals should be structured to encourage the GSEs to take affirmative measures on both fronts. Third, better targeted GSE housing goals for one- to four-family housing loans are needed to encourage the GSEs to purchase more CRA-type mortgage loans originated by banks and savings associations. One of the main reasons why the GSE market share is so low in most minority and low- and moderate-income census tracts is that many mortgage loans originated in such tracts are CRA-type loans that are often not sold to the GSEs. In some cases those loans have below–market rate features, and the GSEs will not purchase such loans without discounting them to market rates. Given the large indirect subsidies received by the GSEs and their substantial resources including strong earnings streams, the GSE should share more equally with banks the costs of effective CRA implementation. For example, with CRA-type loans with below-market rates, the GSEs should absorb some cost of marking these loans to market. Such action would make banks more willing to sell CRA-type loans to the GSEs and would in turn significantly expand the capacity of banks to originate CRA-type loans. JONATHAN BROWN 165

Fourth, HUD should establish a process for administrative review of GSE performance at the local level when GSE market share data indicate that the GSEs have significantly underserved minority census tracts or low- and moderate-income census tracts within a particular MSA or rural area. Local government officials, community groups, and lenders should be encouraged to participate in this administrative review. The administrative process should encourage the GSEs to make commitments to improve their housing goal performance in the local community. Review of GSE performance within individual MSAs or rural areas is important because the GSEs can substantially underserve individual MSAs and still meet national targets if they are evalu- ated only in terms of national targets. Setting uniform targeting standards appropriate for all MSAs or rural areas in the nation is difficult. In particular the relationship between the minority percentage of census tract population and the GSE market share varies dramatically in different parts of the nation. In Brooklyn, New York, neighborhood access to the GSE-operated secondary market may not be greatly restricted until the minority percentage rises well above 50 percent. By contrast in Iowa such restricted access may be seen in census tracts with a minority percentage of only 15 percent. Fifth, separate and enforceable subgoals for one- to four-family mortgage loans should be established for home purchase loans and refinancing loans. The GSE housing goals could be implemented much more effectively if the home purchase loan market were sepa- rated from the volatile and increasingly controversial refinancing loan market for evaluation purposes. Finally, should the GSEs be given housing goal credit for the purchase of subprime loans? In many minority neighborhoods of the nation’s larger cities, mortgage lenders classified by HUD as subprime lenders originated more than 40 percent or even 50 percent of the refinancing loans in 1998. GSE purchase of a substantial share of these loans, with housing goal credit awarded for such purchases, would significantly increase GSE housing goal performance in these neighborhoods. However, the primary mission of the GSEs should be to bring as many borrowers as possible into the main- stream of prime mortgage credit financing, or at least to provide a strong secondary market for legitimate CRA-type lending. At this point it would be counterproductive to the underlying GSE mission to encourage the GSEs to purchase subprime loans by giving them housing goal credit for such purchases.

PART THREE Ideas for Change

Opponents of Fannie Mae and Freddie Mac have frequently advanced the concept of full privatization, that is, cutting all ties to the federal government, as the best way to resolve the role of government-sponsored entities in the economy. However, considerable debate concerns exactly how this should or could be done. In the mid-1990s, another GSE, the Student Loan Marketing Association—otherwise known as Sallie Mae—voluntarily sought privatization. The structure that it used—a corporate holding company with a temporary GSE subsidiary—could serve as a template for similar privatization of Fannie and Freddie. Regardless of the final vehicle, the privatization of Sallie Mae demonstrates that such a step is feasible and does not raise insurmountable legal problems. In this chapter Mark Overend outlines the reasons that Sallie Mae sought privatization. Significantly, once Sallie Mae’s management recognized that the benefits of remaining a GSE had been substantially reduced, they concluded that privatization—and the freedom to innovate and extend activities that it offered—was preferable to continuing its activities under its restrictive government charter. Similarly, many analysts believe that if the benefits of Fannie and Freddie’s government ties were substantially reduced, the two would make the same choice. Chapter 10 also discusses the structure that Sallie Mae ultimately used to meet several concerns about the process and outcome of privatization.

169 10 The Privatization of Sallie Mae

Mark Overend

Through much of Sallie Mae’s existence, many analysts viewed its GSE status favorably. As a GSE, we had access to the agency markets that provided us with attractive funding costs. We were also exempt from registration and filing requirements with the Securi- ties and Exchange Commission—although for all intents and purposes we behaved as if we were not. And we were exempt from most state and local taxes. The 1990s ushered in dramatic changes. In 1993, with a stroke of the pen, Congress effectively eliminated our funding advantage. Equally important, Sallie Mae’s primary role of making a secondary market for student loans, its purpose for being, faded in the face of increased competition and the evolution of the capital markets. Clearly this GSE’s life cycle had run full course. Our attention, as well as Congress’s, shifted to privatization as the logical end. The success of the privatization effort demanded a solution that would accomplish several goals: • guard the interests of Sallie Mae’s shareholders and bondhold- ers, the providers of capital that allow a GSE to fulfill its mandate • protect the public policy concern of providing adequate liquidity in the higher education marketplace • recognize the significant role of the corporation in the overall competitive balance of the industry As the nation’s leading provider of origination and servicing for student loans, Sallie Mae was in a unique position for privatizing. The particular nature of the student loan business distinctly influenced some factors in the process. Other factors were more personal and related specifically to the future direction of the corporation. As the discussion about privatization came to the forefront, the student loan marketplace was experiencing significant changes that

170 MARK OVEREND 171 fueled both public and private interest in moving ahead. Sallie Mae’s traditional customer base had been banks and other lenders acting as loan originators. Those customers had increasingly come to rely on Sallie Mae not simply as a provider of capital but as the devel- oper of new products and service offerings—a shift that altered the relationships. Whereas lenders originally determined secondary- market transactions at the time of the sale, the market evolved into longer-term marketing and servicing alliances with predefined secondary-market transactions. Sallie Mae’s ability to provide expertise in marketing and product development across lender relationships had also begun to expand our focus solely from lenders to include schools and students, the originally intended beneficiaries of the student loan programs. Sallie Mae was determined to provide a national brand of products to meet the increasingly demanding expectations of schools for standard- ized loan delivery and top-flight service levels. To gain and maintain campus loyalty required a broader definition of student loan services: the market share of student loans often depended on providing tan- gential or related services for other administrative needs. Because limitations of the corporate charter prevented Sallie Mae from being a direct originator or pursuing some new business lines, the major investments and technological innovations brought by the corporation to the program did not necessarily lead to better economics or improved business opportunities. As a privatized entity, Sallie Mae could also better explore a wider variety of campus-based services and open the door to heightened brand awareness and new sources of revenue. Beginning in the early 1990s, the student loan industry faced perhaps its stiffest competition from the very agency tasked with regulation and oversight of the Federal Family Education Loan Pro- gram (FFELP). The Department of Education became a lender and service provider in direct competition with Sallie Mae and its loan origination partners with the introduction of the Federal Direct Stu- dent Loan Program. Government entry into our competitive marketplace demanded nimble and targeted responses from all other players in the industry. Sallie Mae was concerned that the restrictions of our charter would prevent an independent and cost-effective response to the new challenge. Government and political entities almost solely determine the infrastructure of the guaranteed student loan program from loan terms to insurance requirements to expectations of the basic level of servicing. Most notable, the law mandates interest rates instead of 172 PRIVATIZATION OF SALLIE MAE allowing the rates to be subject to traditional market forces. An active Congress interested in protecting the financial concerns of students and a non–market-based determination of interest rates could—and did—effectively eliminate the funding benefits of a GSE. Because Sallie Mae is a GSE with a unique role in the student loan program, the charter became a tool for legislators to make statements and shape policy. In the years leading up to the privatization effort, some legislative actions adversely affected the corporation’s perceived benefit as a GSE: • the imposition of an annual offset fee of 30 basis points that equated to a special form of taxation and effectively eliminated our agency’s funding advantage • the mandated role as a lender of last resort within the program for students who could not otherwise get guaranteed loans because of their institutions’ high default rates • the specific statutory minimum capital requirements • the restrictions on avenues of business allowed by the charter The deliberation over privatization did not occur solely in the halls of Congress, the floor of the exchanges, or the quads of our nations’ colleges and universities. An internal debate raged over the future of the company; the impact of a potential privatization effort had become a strategic point. No consensus developed on the exact for- mula for future success, but nearly all parties involved had concluded that our corporate viability depended on the freedom to respond in a timely manner to a changing marketplace without concern over charter restrictions or political action. The 1996 privatization legislation created a corporate entity con- sisting of a state-chartered holding company, which owned the federally chartered GSE as well as other non-GSE subsidiaries. Dissolution of the GSE was designed to take place over ten years. The entity remains open for its traditional secondary-market business and can continue to issue debt obligations as long as they ma- ture prior to the dissolution date. The structure of a holding company allowed Sallie Mae to continue to use the GSE for activities involved in student loan purchasing while the corporation began to explore and pursue new business opportunities. The privatized Sallie Mae subsidiaries encountered no restraints regarding new business lines. However, the publicly traded holding company and its non-GSE subsidiaries could not engage in purchasing student loans as long as the GSE continued in MARK OVEREND 173 that field.The enacted privatization legislation contained a key feature: the element of choice for Sallie Mae shareholders. Congress recognized that allowing shareholders a choice was prudent at the end of the public-private partnership—the shareholders of Sallie Mae after all had brought their capital to the table for almost thirty years. Shareholder rejection of the proposed privatization plan would have prompted an alternative wind-down plan for the GSE operations. The economics of the privatization allowed Sallie Mae to trade GSE benefits, which had been decaying under increased political pressure and public policy concerns, for an opportunity to pursue expanded business opportunities in a marketplace that had strate- gically changed. The legislation as approved provided a workable framework for the transition that minimized the impact to Sallie Mae’s core business while adding a significant opportunity for growth . In some respects it is still too early to evaluate fully the success and impact of Sallie Mae’s privatization. Most notable, the GSE wind- down continues. However, some meaningful signs indicate that the arguments put forth by the company and its supporters hold true. Sallie Mae is becoming an originator on many of our nation’s campuses. Our investments in new technologies are helping to revolu- tionize the entire student loan industry. The FFELP program, once faced with a serious threat from direct lending programs, has made strong inroads in its efforts to reclaim its market share. And we continue to report strong quarterly earnings. Not every line of business that interested us has borne fruit. Several such fledglings have already come and gone from the organization, while other directions never fully materialized. Some analysts see the situation as evidence that privatization led to solid business practice. The collective expertise of the corporation, instead of the strength of public sentiment and political winds, determined our decisions to enter new markets or provide new services. Sallie Mae will continue to identify, evaluate, and implement new ideas on the basis of the dictates of a strategic business plan. Privatization may or may not be the natural conclusion of the public-private partnership embodied in all GSEs. But in the case of student loans and Sallie Mae, all parties—consumers, campuses, American taxpayers, shareholders, and employees alike—have apparently benefited from this course of action. Assuming that full privatization does not or cannot occur, a number of proposals offer plans to draw off the GSEs’ subsidy. Some of these would impose a guaranty fee to recapture the subsidy for the government. Others would require Fannie and Freddie to increase their support of the low-income and affordable housing market substantially to divert the subsidy toward more worthy purposes. In the following essay Thomas H. Stanton presents a proposal to create a number of competitive GSEs—with powers and advantages identical to those of Fannie and Freddie—to force the subsidy to be distributed to the housing markets through competition. In Stanton’s plan Congress would authorize the chartering of an unlimited number of GSEs, all able to compete with Fannie and Freddie on identical terms. The new GSEs and Fannie and Freddie would be subject to a new and more pervasive regulatory regime to address to some degree the moral hazard problem currently associated with Fannie and Freddie. The most significant aspect of the proposal would presumably be its political force. The GSEs would have difficulty arguing against the establishment of competitive enterprises, and those who voted for it could claim that they would bring down mortgage rates through additional competition and thus avoid the GSEs’ charge that full privatization would raise mortgage rates. That point, however, serves only to emphasize that the creation of another set of housing GSEs is a second-best solution, which would find its greatest support if full privatization cannot be achieved.

174 11 An Alternative Approach to GSE Reform through National Mortgage Associations

Thomas H. Stanton

The public costs of the largest GSEs—Fannie Mae, Freddie Mac, and the Federal Home Loan Bank System—are increasing while their public benefits are decreasing. The public costs include a concentration of financial risk in three of the largest financial institutions in the United States and, in the case of Fannie Mae and Freddie Mac, market power that is reshaping the American system of mortgage finance on the basis of government subsidies rather than open competition. Past public benefits of the GSEs included many contributions to overcoming imperfections in the mortgage market, for example, through standardizing mortgage forms. However, most imperfections no longer exist in today’s efficient financial markets.1 From the perspective of the housing sector, perhaps the major function of the GSEs today is to convey a multibillion-dollar federal subsidy to the housing market outside the federal budget process. Because the GSEs serve this function, they have been able to claim that any effort to reduce potential public costs—say, by increasing GSE capital requirements—or to increase public benefits—say, by imposing truly significant affordable housing requirements— amounts to a so-called tax on housing. Thanks to their market power, the GSEs can protect their shareholders from the effects of policy changes merely by increasing the amount of the federal subsidy that they keep for themselves compared with the amount that flows to the housing sector. In one approach to the problem of a changing balance between public benefits and public costs, the government could design and implement an exit strategy so that the three housing GSEs could

175 176 NATIONAL MORTGAGE ASSOCIATIONS relinquish their government sponsorship and become completely private companies without special tax or regulatory benefits and without the ability to borrow in the federal agency debt market.2 The solution has not been feasible because of objections from the housing sector, reinforced by the immense political power of the GSEs, and also because of fears among their competitors that Fannie Mae and Freddie Mac have acquired such a market presence that they would continue to be dominant in the mortgage market even without government sponsorship (MSDW 2000, 18). This chapter presents an alternative type of GSE reform to re- dress the balance between public costs and public benefits and addresses concerns about the effects on the housing sector. Under this alternative approach, government would permit a federal regulator to charter numerous competing GSEs with legal powers identical to those of Fannie Mae and Freddie Mac. The regulator would have powers comparable to those of the Comptroller of the Currency, the federal agency that charters and supervises national banks. By fostering competition, this alternative approach would (1) reduce the concentration of financial risk present in the housing GSEs, (2) reduce the adverse consequences for innovation and competition that Fannie Mae and Freddie Mac cause through their market power, and (3) ensure that a greater proportion of the federal subsidy flows through the GSEs and into the housing market.3 Fannie Mae and Freddie Mac have enjoyed returns on equity of 20–34 percent in the 1990s (HUD 2000), far above the profitability of other financial institutions. In contrast to Fannie Mae and Freddie Mac today, the new GSEs would obtain more competitive and less generous rates of return. They also would exert competitive pressure that would reduce the duopoly returns enjoyed by the two largest GSEs. The resulting surplus could be used to lower mortgage rates and to improve the balance between public benefits and public costs of the housing GSEs. The first section of this chapter begins with an overview. Next the proposal is presented in institutional terms, including a discussion of the attributes of the regulator and the GSEs that would be chartered under the proposal. The new GSEs might be called Na- tional Mortgage Associations in deference to a similar statutory system that the Roosevelt administration helped to enact in title III of the National Housing Act of 1934 (ch. 847, 48 Stat. 1246, enacted June 27, 1934). Multiple “Nancy Maes” might thus join Fannie and Freddie in the housing finance playground. A consideration of some THOMAS H. STANTON 177 of the major policy issues includes the question whether the creation of national mortgage associations could be a useful step in ending agency status for GSEs.

Institutional Mechanics of Creating and Supervising New GSEs

Fannie Mae and Freddie Mac benefit from unique charters that Con- gress has enacted in legislation. No other incorporators, regardless of how well qualified, may obtain a similar GSE charter except through another special statute. Such a legal framework differs markedly from the chartering of national banks, federal savings and loan associations, and most other types of financial institution. A federal regulator, the Office of the Comptroller of the Currency (OCC) of the Treasury Department, charters national banks, for example. The Office of Thrift Supervision (OTS), also in Treasury, charters federal savings and loan associations. Congress should create a new federal regulator with authority comparable to that of the OCC or OTS. That regulator would have the authority to charter and supervise new national nortgage associations. The new federal regulator might be called the Office of National Mortgage Association Supervision (ONMAS), just as the OTS is named after the institutions that it supervises.

The Regulator. The regulator would have the authority to charter many national mortgage associations. The new Office of National Mortgage Association Supervision would charter national mortgage associations that would compete against one another and against other types of financial institutions, much as many national banks compete in the financial markets today. The regulator would re- semble the OCC in its authority to charter institutions, supervise them for safety and soundness, oversee the scope of their permitted business activities, apply requirements to serve public purposes that may not be as lucrative as other activities (for national banks, these are requirements of the Community Reinvestment Act; for the housing GSEs, these might include contributions to an affordable housing fund), and appoint conservators or receivers in cases of serious financial difficulty. To ensure the primacy of safety and soundness considerations, the new regulator would be an autonomous bureau within the Treasury Department. 178 NATIONAL MORTGAGE ASSOCIATIONS

The New GSEs. National mortgage associations would have the same statutory powers as Fannie Mae and Freddie Mac. For example, they would be authorized to “purchase, service, sell, lend on the security of, and otherwise deal in” conforming mortgages,4 and engage in all other activities that Fannie Mae or Freddie Mac may carry out under their charters. The law would link the statutory authority of national mortgage associations to that of Fannie Mae and Freddie Mac: an expansion of statutory authority for either Fannie Mae or Freddie Mac would automatically extend to the national mortgage associations chartered by the regulator. Conversely any increase in statutory burdens on national mortgage associations, especially capital requirements, would automatically extend to Fannie Mae and Freddie Mac.5 So long as the national mortgage association charter retains the basic benefits available to Fannie Mae and Freddie Mac, various types of companies, possibly including Federal Home Loan Banks as well as completely private mortgage lenders, would likely find it financially attractive to establish and operate a mortgage business through the new charter. If the current GSEs drop their fees to try to drive out the new competition, the mortgage market will benefit from lower mortgage rates. Some large institutions may find it useful to create a national mortgage association as a way to put pressure on the GSEs to lower their fees; others may be able to develop scale fast enough to mount actual competition, much as the Federal Home Loan Bank of Chicago has used GSE status to create a growing market for itself.

Open Issues. Some issues require special attention. Several relate primarily to the differences between GSEs and other financial institutions with federal backing: (1) supervisory powers of the new regulator, (2) capital requirements, and (3) affordable housing requirements for the new GSEs. A fourth issue relates to the role of the Federal Home Loan Banks in the new system of housing finance institutions. Consider each of these in turn. Supervisory powers. Ideally all GSEs would be subject to supervision by a regulator that possesses the statutory powers of a federal bank regulator with respect to safety and soundness (see Stanton 1991a). By contrast to the Office of Federal Housing Enterprise Over- sight, the federal bank regulators have a full tool kit of enforcement powers without much of the crippling language of the OFHEO stat- THOMAS H. STANTON 179 ute, have the discretion to set and change capital requirements on a case-by-case basis, and also have the authority to close a failed institution. OFHEO’s powers are almost so weak as to be a parody of the powers of the federal bank regulators (see, for example, Stanton 1999, 2001, and Carnell 1999). The GSEs have rejected the approach of providing comparable powers to their regulator (see, for example, Bacon 1992). However, under this new proposal to create multiple national mortgage associations, the case for giving the regulator the supervisory powers of the federal bank regulators would become even more compelling.6 Capital. The matter of capital is more difficult for several reasons. First, the issue of risk-based capital for commercial banks itself is somewhat in disarray. Second, the GSEs have shown the ability to block any effort at increasing their capital requirements to levels comparable to those that apply to commercial banks. Yet, allowing any institutions, especially potentially numerous national mortgage associations, to operate on the basis of the thin capital requirements that apply today to Fannie Mae and Freddie Mac runs a risk. Compromise might help. For the first several years after enactment of legislation, national mortgage associations would benefit from whatever low capital standards applied to Fannie Mae and Freddie Mac. The regulator of national mortgage associations would have statutory authority, comparable to that available to the federal bank regulators, to increase capital requirements on any GSEs that pose special problems of safety and soundness (see 12 U.S.C. secs. 1831o(g)(1) and 3907). Four or five years after enactment, as the national mortgage associations begin to increase the scale of their activities, the government could phase in higher capital standards for all of the GSEs to a level consistent with the evolving capital standards for commercial banks. One stock research report recently pointed out that Fannie Mae and Freddie Mac could raise “significant amounts of preferred capital . . . with little impact on EPS [earnings-per-share] growth” (Bernstein 2000). Affordable housing requirements. Despite the best efforts of the Department of Housing and Urban Development, Fannie Mae and especially Freddie Mac continue to lag the primary mortgage market in serving minorities and others with special affordable housing needs (HUD 2000; see also Apgar 2000). It may be time to concede that the present system of affordable housing goals has not worked 180 NATIONAL MORTGAGE ASSOCIATIONS well. Instead the new legislation might require all national mortgage associations, plus Fannie Mae and Freddie Mac, to contribute to an affordable housing fund on the same basis that the Federal Home Loan Banks currently contribute to such a fund (12 U.S.C. sec 1430(j)). The new regulator could administer that fund and ensure that the money is properly allocated to affordable housing needs. While the fund is not an essential part of the proposal, all GSEs—Fannie Mae, Freddie Mac, and the new national mortgage associations— can and should be required to make some genuine contribution to affordable housing needs as partial compensation for the benefits that they receive from their federal charters. Federal Home Loan Banks. Given the substantial advantages of a GSE charter vis-à-vis banks and thrift institutions, several banks and thrifts will likely seek to convert their mortgage operations to the new institutional form of a national mortgage association. If so, the role of the Federal Home Loan Banks in providing agency-status advances will be even more diminished in policy significance. (Agency status is a term that applies to the ability of GSEs to borrow in the federal agency credit market on the basis of the perception of an implied federal guarantee. The implied federal backing allows GSEs, despite their actual financial condition, to fund themselves at preferential rates close to those of federal Treasury borrowing.)7 Moreover, the competition from national mortgage associations could reduce conforming mortgage rates by some perceptible amount (perhaps 10–20 basis points?). That reduction would disad- vantage the Mortgage Partnership Finance Program vis-à-vis Fannie Mae and Freddie Mac. One solution would be to permit the Federal Home Loan Banks to become national mortgage associations by converting their charters. 8 The new charter must be flexible enough in its legislative design to permit cooperative ownership as well as investor ownership. This would permit the financial institutions that own Federal Home Loan Banks either to retain their cooperative control over a national mortgage association or, as they choose, to become investors in an investor-owned national mortgage association. The legal mechanics of such conversion will require careful attention. In particular, the converted institutions will need to possess sufficient perpetual stock to meet the statutory capital requirements applicable to national mortgage associations. THOMAS H. STANTON 181

Greater Policy Issues

At least three greater policy issues remain for consideration: (1) the pervasive political influence of the GSEs and their potential to block legislative provisions that would strengthen the supervisory powers of the new regulator, (2) the problem of GSEs that may be considered too big to fail, and (3) the use of national mortgage associations as a transition mechanism looking toward removal of agency status from the obligations and mortgage-backed securities of the housing GSEs.

Strengthening the Supervisory Authority of the New Regulator. The GSEs undisputedly exert considerable dominance over the political process (Stanton 2001). Political influence has been the hallmark of government-sponsored institutions since the days of the two Banks of the United States (Stanton 1994). A direct legislative effort to increase capital requirements for the housing GSEs, for example, would incite allegations that this would cause higher mortgage rates. The proposal to create national mortgage associations must be crafted with that issue in mind. The advent of increased competition in the conforming mortgage market could help to lower mortgage rates. The savings to consumers would then offset the costs of any slight increase in burden on the GSEs resulting from increased bank-type supervisory powers for the new regulator. As discussed, the statute might retain the existing standards regarding inadequate capital at least for the short run but should provide that the regulator have the discretion to require increased capital for any institution posing special concerns relating to safety and soundness. The ability of the regulator to increase capital on such a selective basis would not increase the cost of funding mortgages through the other national mortgage associations. Ultimately the government needs to try to avoid the specter of Stanton’s law, that “risk will migrate to the place where government is least equipped to deal with it.” Because of the capability of the financial markets to arbitrage across differential regulatory requirements, the government must synchronize the capital standards and supervisory requirements applicable to other financial institutions with those that apply to national mortgage associations.9 Firms will shift their mortgage portfolios, for example, so that the institutions with the lowest capital standards hold them, other things being equal. 182 NATIONAL MORTGAGE ASSOCIATIONS

GSEs That Are Too Big to Fail. The housing GSEs are immense institutions that create an exceptional and growing concentration of risk for the financial system. At year-end 1999 Fannie Mae had total assets of $575 billion plus net MBS outstanding of $679 billion for a total size of $1.25 trillion. Freddie Mac had total assets of $387 billion plus total MBSs outstanding of $538 billion for a total size of $925 billion. Because of their size alone, these GSEs are arguably too big to fail. As Freddie Mac has stated: In light of long-standing public policy of preventing a crisis in the financial markets, it is not likely that Congress would fail to intervene to address a future crisis in the mortgage markets. . . . Consequently, repealing Freddie Mac’s and Fannie Mae’s charters could result in eliminating the benefits of the implied guarantee without eliminating the risk to the taxpayer. (Federal Home Loan Mortgage Corporation 1996) The proposal to create national mortgage associations helps to address the problem by helping to disperse the funding of conforming mortgages among many different financial institutions. In addition the legislation might include a provision that limits the total holdings of conforming mortgages by any single institution to a stated fraction, say, 5 or 10 percent, of the entire conforming mortgage market for that year.10 During a mandatory transition Fannie Mae and Freddie Mac could reduce their mortgage portfolios, either by selling to other institutions or by funding the mortgages through securitization. Finally, it would be useful to adopt effective constraints on bail- ing out GSEs for fear that they were too big to fail. The problem here, as Richard Carnell points out in chapter 4, is that the government finds it difficult to manage the implicit government guarantee for fear of making the guarantee even more concrete than it is. Just as imaginative legal craftsmanship allowed the government to create the implicit guarantee, so too a well-drafted legal provision is needed to protect taxpayers from having their money used to bail out a failed GSE except under conditions similar to those that apply to a bailout of a failed commercial bank.

Creation of National Mortgage Associations as a Step to Removal of Agency Status. The creation of competing national mortgage as- THOMAS H. STANTON 183 sociations and the limitations on concentration and the strengthening of supervisory powers, as described, together can help to reduce the financial risk exposure of taxpayers to the potential costs of failure of any single GSE.11 In addition the creation of national mortgage associations could provide an elegant transition to permit the government to remove agency status from the obligations and mortgage-backed securities of the housing GSEs. In other words the government might take the following legislative actions: (1) permit the creation of national mortgage associations, (2) place a sunset provision to take effect, say, five years after the date of enactment in the legal provisions that together create the perception of agency status of GSE obligations and mortgage-backed securities, and (3) perhaps create a sunset provision for the Federal Home Loan Bank Act in its entirety, again five years after date of enactment. In addition some of the more questionable provisions of the GSE charter acts—such as the exemptions from federal securities laws and from state and local income taxes—need to be updated. Even without agency status, firms are likely to find the national mortgage association charter to be valuable. In their proposals to displace private firms that offer credit life insurance and mortgage insurance, Fannie Mae and Freddie Mac have contended that their charters are broad enough to permit a combination of loan products and insurance products, for example. Also the provision in the federal charters for a more uniform legal framework across the fifty states benefits the mortgage market. The creation of national mortgage associations could establish competitive conditions that ease the transition to a mortgage market without today’s government-sponsored enterprises. Competing firms would have the opportunity to build some scale in their mortgage operations before agency status ended. Similarly, if the Federal Home Loan Banks were included in the legislation, they would have the chance to operate under an improved charter for some years before making the transition to non–agency-status institutions. The United States took a similar approach to deregulation of geographic limits on commercial banks. By permitting regional banks to grow during transition, legislation allowed strong institutions to emerge before the regulatory barriers to interstate banking disap- peared completely. The proposal for national mortgage associations similarly could allow institutions to gain financial strength before they lost the protection of federal laws that impede more efficient competition. 184 NATIONAL MORTGAGE ASSOCIATIONS

Finally, any proposal for a more efficient mortgage market must not lose sight of the issue of affordable housing. As discussed, the creation of national mortgage associations would help to support affordable housing through the establishment of an affordable housing fund that would receive a percentage of the annual net income of each housing GSE. The question becomes, What happens to the fund if the government removes agency status from the housing GSEs? Several options for supporting such a fund, either from public or private sources, exist; the particular legislative package that policymakers decide to propose must fit in the contours of that support.

An Invitation for Comments

The essence of statecraft combines good policies with sound politics. The basic idea of national mortgage associations is to recognize the overwhelming political salience of the GSEs’ battle cry that any change that they oppose, whether higher capital requirements or increased affordable housing or a more capable federal regulator, amounts to a tax on housing. The idea of national mortgage associations creates the possibility for lower mortgage rates, thanks to increased competition, along with a reduction in the concentration of financial risk in the three housing GSEs and other improvements in the balance between their public benefits and public costs. Needed now is healthy debate, both on the general idea and on the particular elements that might be part of the package. The position of Fannie Mae and Freddie Mac on this issue can be anticipated. Other constituencies and analysts should make their views known as well.

Notes

1. One market imperfection that still exists in significant form relates to housing discrimination. Yet here the GSEs lag rather than lead the market. In 2000 HUD Assistant Secretary William Apgar testified that Fannie Mae and Freddie Mac’s “share of the affordable housing market is substantially smaller than their share of the total conventional conforming market. Lower income families, certain minorities, central-city residents, and immigrant populations continue to be underserved by Fannie Mae and Freddie Mac.” (Apgar 2000). THOMAS H. STANTON 185

2. See, for example, HUD 1996—prepared for the General Accounting Office, Department of Housing and Urban Development, Treasury, and Con- gressional Budget Office—and Dean, Moskowitz, and Cipriani 1999. 3. On the last point, see Seiler 1997. 4. Conforming mortgages are those that Fannie Mae and Freddie Mac may purchase under their charter authority, in 2001 up to a size limit of $275,000 for single-family mortgages. See 12 U.S.C. secs. 1717(b)(2) and 1454(a)(2). 5. The last point relates to historical precedent: the original statutory concept for national mortgage associations became flawed when competitors, then the savings and loan industry, managed to burden the statute with such requirements and limitations that it was not profitable for firms to establish national mortgage associations. See Semer et al. 1976. 6. Economists might argue that the shared monopoly of Fannie Mae and Freddie Mac results in the GSEs’ expecting to earn supernormal profits. The expectation of duopoly rents would be diminished as national mortgage associations began to compete effectively. The reduction in franchise value could increase the incentive of some competing companies to increase returns by taking risks, compared with a more monopolistic market structure. The solution is to improve oversight of safety and soundness so that it re- sembles federal supervision of commercial banks, and then today’s federally sponsored duopoly could be eliminated. 7. For a discussion of the so-called agency attributes of GSEs and an example of the limited disclosures that GSEs make, see Stanton 1991b, 40–44. 8. See also Stanton 1991b, 196–97. In comments on an earlier draft of this chapter, Alex Pollock indicated that a number of Federal Home Loan Banks could find it attractive to become national mortgage associations. 9. Commercial banks may be diversified in their types of lending as well as geographically. By contrast, while the GSEs may be diversified geographically, their lending is largely concentrated in a single type of loan, that is, the conforming mortgage loan. Other things being equal, such differences among institutions argue for the application of higher capital requirements to the financially less diversified institutions. 10. The 10 percent figure has precedent in provisions of the Riegle-Neal Interstate Banking and Branching Efficiency Act of 1994, which limit interstate mergers of depository institutions under that law to concentration limits of 10 percent of outstanding insured deposits in the United States. See 12 U.S.C. secs. 1831u(b)(2) and 1842(d)(2). 11. The recent proposal of the Treasury Department to reduce concentrated holdings of GSE obligations by banks and thrift institutions would be a useful precursor to such legislation. 186 NATIONAL MORTGAGE ASSOCIATIONS

References

Apgar, William. 2000. Testimony before the U.S. House Subcommittee on Capital Markets, Securities, and Government-Sponsored Enterprises, March 22. Bacon, Kenneth H. 1992. “Privileged Position: Fannie Mae Expected to Es- cape an Attempt at Tighter Regulation.” Wall Street Journal, June 19. Bernstein Research. 2000. The GSEs: Hegemony in the Mortgage Markets, p. 122. Carnell, Richard Scott. 1999. “Federal Deposit Insurance versus Federal Spon- sorship of Fannie Mae and Freddie Mac.” Paper presented at conference, “Fannie Mae and Freddie Mac: Public Purpose, Private Interests,” AEI, Washington, D.C. Dean, John E., Saul L. Moskowitz, and Karen L. Cipriani. 1999. “Implica- tions of Privatization of Sallie Mae.” Public Budgeting, Accounting and Fi- nancial Management 11 (1) (spring): 56–80. Federal Home Loan Mortgage Corporation. 1996. “Everything You Wanted to Know about Privatization but Had No Time to Ask.” May 13, p. 20. Morgan Stanley Dean Witter. 2000. Internet & Financial Services: Catching the Third Wave, p. 18. April 14. Seiler, Robert S. Jr. 1997. “Fannie Mae and Freddie Mac as Investment-Owned Utilities.” Journal of Public Budgeting, Accounting & Financial Management 11 (1) (spring): 134–44. Semer, Milton P., Julian H. Zimmerman, Ashley Foard, and John M. Frantz. 1976. “Evolution of Federal Legislative Policy in Housing Credits.” In Housing in the Seventies Working Papers, vol. 1, pp. 28–29. Washington, D.C.: Department of Housing and Urban Development. Stanton, Thomas H. 1991a. “Federal Supervision of Safety and Soundness of Government-Sponsored Enterprises.” Administrative Law Journal 5 (2) (summer): 395–484. ———. 1991b. A State of Risk. New York: HarperCollins. ———. 1994. “Nonquantifiable Risks and Financial Institutions: The Mer- cantilist Legal Framework of Banks, Thrifts and Government-Sponsored Enterprises.” In Global Risk-Based Capital Regulations, edited by Charles Stone and Anne Zissu, pp. 57–97. Irwin Professional. ———. 1999. “Devising an Effective Legal Framework for Supervising the Public Benefits and Public Costs of Government-Sponsored Enterprises.” Paper presented at conference, “Fannie Mae and Freddie Mac: Public Purposes, Private Interests,” AEI, Washington, D.C. ———. 2001. “Government Sponsored Entities (GSEs): Why Is Effective Government Supervision Hard to Achieve?” Paper presented at Thirty- seventh Annual Conference on Bank Structure and Competition, Federal Reserve Bank of Chicago, May 10. THOMAS H. STANTON 187

U.S. Department of Housing and Urban Development. 1996. Studies on Priva- tizing Fannie Mae and Freddie Mac. Washington, D.C.: Government Print- ing Office. ———. Proposed rule. 65 Federal Register 12631, p. 12678, March 9. ———, Office of Federal Housing Enterprise Oversight. 2000. 2000 Report to Congress, tables 3 (Fannie Mae earnings) and 13 (Freddie Mac earnings), pp. 102, 112. Washington, D.C.: Government Printing Office.

AEI STUDIES ON FINANCIAL MARKET DEREGULATION Charles W. Calomiris and Peter J. Wallison, series editors

High Loan-to-Value Mortgage Lending: Problem or Cure? Charles W. Calomiris and Joseph R. Mason

Is the Bank Merger Wave of the 1990s Efficient? Lessons from Nine Case Studies Charles W. Calomiris and Jason Karceski

Nationalizing Mortgage Risk: The Growth of Fannie Mae and Freddie Mac Peter J. Wallison and Bert Ely

Optional Federal Chartering and Regulation of Insurance Companies Edited by Peter J. Wallison

The Postmodern Bank Safety Net: Lessons from Developed and Developing Countries Charles W. Calomiris

Reforming Bank Capital Regulation: A Proposal by the U.S. Shadow Financial Regulatory Committee Shadow Financial Regulatory Committee

Regulating Financial Markets: A Critique and Some Proposals George J. Benston

Serving Two Masters, Yet Out of Control: Fannie Mae and Freddie Mac Edited by Peter J. Wallison