NBER Reporter Winter 06.Qxd

NBER Reporter

NATIONAL BUREAU OF ECONOMIC RESEARCH

Reporter OnLine at: www.nber.org/reporter WINTER 2005/6

Program Report IN THIS ISSUE Public Economics Program Report: Public Economics 1 Research Summaries: Contribution of Science … to Production 8 James M. Poterba* Foreign Direct Investment Behavior of MNCs 11 Race and … American Economic History 15 Globalization and … Economic History 18 Researchers in the NBER Public Economics Program study many of the core issues that have been at the center of recent national policy NBER Profiles 22 Conferences 24 debates. While the Public Economics Program is broadly concerned with Bureau News 36 the economic role of government, an expansive definition that includes Bureau Books 57 some research in virtually every sub-field of economics, two of its most Current Working Papers 58 important research themes are the economics of taxation and the analysis of social insurance programs. Since the last Public Economics Program Report in 2001, the United States has undergone substantial tax reform in the form of the 2003 and 2004 tax bills. Because many of the tax reform provisions that were enacted in 2001 and 2003 are scheduled to expire later this decade, further tax reforms have already been enacted in a sense. Policy debate about the extension of these tax provisions, and about the structure of the tax system more generally, seems very likely to continue through the next few years. Tax reform has been widely discussed and there have been substantial NBER Service Brings changes in the last five years. In contrast, Social Security reform has also You New Data for Free been widely discussed, but there have been no significant changes in the program’s structure. Public programs for retirement income support have A new, free NBER email serv- been active topics of discussion in many industrialized nations. In the ice gives you daily email links to all United States, the earnest discussion of Social Security reform began when U.S. government data releases, a Presidential commission suggested several reform proposals in 2001. including unemployment, trade, Since then, various policy analysts and legislators have advanced a range of interest rates, GDP, etc. We keep track of your preferences and different proposals for reform. They differ in the role that they envision for email you the requested links the government in providing retirement income, and in their potential when they are released. To sign up effects on the long-run fiscal balance of the Social Security system. for any or all of the government Medicare, which portends to become an even more costly entitlement pro- releases, visit www.nber.org/releases and register your choices. IT’S FREE!! * Poterba directs the NBER’s Program on Public Economics and is a Professor of Economics at MIT. In this article, the numbers in parentheses refer to NBER Working Papers.

NBER Reporter Winter 2005/6 1 gram over the long-term future, has attracted less policy attention than Social Security. NBER Reporter The NBER Public Economics Program includes a very diverse group of researchers. NATIONAL BUREAU OF ECONOMIC RESEARCH Nearly 120 Faculty Research Fellows and Research Associates claim affiliation with the program, although only half of those The National Bureau of Economic Research is a private, nonprofit research organization founded in 1920 and devoted to objective quantitative analysis of researchers cite Public Economics as their pri- the American economy. Its officers and board of directors are: mary affiliation. Program affiliates have a long President and Chief Executive Officer — Martin Feldstein tradition of analyzing tax policies and of study- Vice President for Administration and Budget — Susan Colligan Controller — Kelly Horak ing social insurance programs such as Social Security. They also study a very wide range of BOARD OF DIRECTORS other topics, including environmental econom- Chairman — Elizabeth E. Bailey Vice Chairman — John S. Clarkeson ics, political economy, and health economics. Treasurer — Robert Mednick Program members meet twice each year at pro- DIRECTORS AT LARGE gram meetings, and again for a variety of work- shops during the NBER Summer Institute. In Peter Aldrich Jessica P. Einhorn John Lipsky Elizabeth E. Bailey Martin Feldstein Laurence H. Meyer the last four years, there have been eight pro- John Herron Biggs Jacob A. Frenkel Michael H. Moskow Andrew Brimmer Judith M. Gueron Alicia Munnell gram meetings and more than twenty Summer John S. Clarkeson Robert S. Hamada Rudolph A. Oswald Institute group meetings. Since late 2001, pro- Don R. Conlan George Hatsopoulos Robert T. Parry Kathleen B. Cooper Karen N. Horn Richard N. Rosett gram affiliates have disseminated 605 working George Eads Judy Lewent Marina v. N. Whitman Martin B. Zimmerman papers, or more than one sixth of all NBER papers, and published six books and a number DIRECTORS BY UNIVERSITY APPOINTMENT of special issues of academic journals. George Akerlof, California, Berkeley Joel Mokyr, Northwestern One recent innovation in the Public Jagdish W. Bhagwati, Columbia Andrew Postlewaite, Pennsylvania Michael J. Brennan, California, Los Angeles Craig Swan, Minnesota Economics group is the creation of several Glen G. Cain, Wisconsin Uwe Reinhardt, Princeton Ray C. Fair, Yale Nathan Rosenberg, Stanford working groups that tackle specific research Franklin Fisher, MIT David B. Yoffie, Harvard issues related to various topics in public policy. Saul H. Hymans, Michigan Arnold Zellner, Chicago Marjorie B. McElroy, Duke One such group, which Martin Feldstein and I DIRECTORS BY APPOINTMENT OF OTHER ORGANIZATIONS have co-directed, focuses on the Behavioral Responses to Taxation. Its members are drawn Richard B. Berner, National Association for Business Economics Gail Fosler, The Conference Board from the U.S. Treasury Department, the Dr. Arthur Kennickell, American Statistical Association Congressional Budget Office, and the Joint Richard C. Green, American Finance Association Thea Lee, American Federation of Labor and Congress of Committee on Taxation, as well as from Industrial Organizations Robert Mednick, American Institute of Certified Public Accountants NBER’s ranks. This group has met to discuss Angelo Melino, Canadian Economics Association completed research about, and the research Jeffrey M. Perloff, American Agricultural Economics Association John J. Siegfried, American Economic Association agenda for, the link between tax rates and vari- William W. Lewis, Committee for Economic Development Gavin Wright, Economic History Association ous dimensions of taxpayer behavior such as labor supply, capital gain realizations, and the The NBER depends on funding from individuals, corporations, and private foundations to maintain its independence and its flexibility in choosing its reporting of aggregate taxable income. The research activities. Inquiries concerning contributions may be addressed to Working Group has scheduled meetings just Martin Feldstein, President & CEO, NBER 1050 Massachusetts Avenue, Cambridge, MA 02138-5398. All contributions to the NBER are tax before or just after Program Meetings, or during deductible. the NBER’s Summer Institute, to maximize par- The Reporter is issued for informational purposes and has not been reviewed by ticipation by the NBER affiliates. A second such the Board of Directors of the NBER. It is not copyrighted and can be freely reproduced with appropriate attribution of source. Please provide the NBER’s group directed by NBER Research Associate Public Information Department with copies of anything reproduced. Douglas Shackelford of the University of North Requests for subscriptions, changes of address, and cancellations should be Carolina focuses on Financial Accounting and sent to Reporter, National Bureau of Economic Research, Inc., 1050 Massachusetts Avenue, Cambridge, MA 02138-5398. Please include the Taxation. It includes researchers in the fields of current mailing label. accounting, finance, and public finance. Its agenda includes issues at the intersection of public

2 NBER Reporter Winter 2005/6 ------finance and accounting, for example significantly reducing the relative tax bur- of a substantial reporting response to explaining the growing disparities den on dividends relative to corporate changes in tax rates (10273, 10044). between book and tax income for U.S. retained earnings that generate capital Taxable income is the sum of many corporations and evaluating the impact gains. The 2004 tax bill introduced a tran- components, each of which is may be of various tax reform proposals on sitory tax holiday for firms repatriating affected to different degrees by changes accounting earnings and corporate bal- earnings from foreign subsidiaries, and it in marginal tax rates and in the structure ance sheets. created a range of specialized provisions of the tax base. Not surprisingly, a sub- A brief report such as this cannot do to encourage specific business activities. stantial body of research has examined justice to the breadth of research that is Researchers in the NBER Public the components of taxable income and carried out by Public Economics Economics Program have analyzed the their sensitivity to the tax system. One Program members while also explaining economic effects of tax changes similar item that has attracted attention is the the substantive contributions of this to those embodied in each of these tax mortgage interest deduction. It has been research. I have therefore decided to bills. As data on taxpayer response to the actively discussed of late because the focus here on four broad areas: taxation, tax reforms has become available, they President’s Advisory Panel on Tax social insurance programs, political econ- have also provided a rapid evaluation of Reform suggested tightening current lim- omy, and the economics of the state and actual behavioral changes in response to its. Several recent studies have explored local government sector. Because taxa- tax rules. the economic effects of the current tion issues are studied exclusively by The tax changes of 2001, which deduction rules and the distribution of researchers in the Public Economics were temporary, phased-in gradually, and the resulting tax deductions across group, while Aging, Economic Fluctua- included an immediate tax rebate as a income groups and geographic locations tions and Growth, and Well-Being of means of stimulating economic activity, (10322, 9284). Because house prices dif- Children Program researchers study have generated several lines of research fer widely, there are large differences in some of the other issues (and their on tax policy and household behavior. A the average value of mortgage interest research is described in other Program number of studies have explored how deductions across states, with much high- Reports), I will devote more than equal consumers responded to the increase in er values on the East and West Coast time to the issues related to taxation. I aftertax income that resulted from the than in Mid-western states. Other will describe how Program members immediate tax rebate (10784, 9308). research has considered the effect of have approached a variety of topics, and These studies suggest that most house- changes in the home mortgage interest I will briefly summarize their key research holds spent between half and three quar- deduction in other nations, with particu- findings. This report unfortunately ters of their tax rebate within six months lar emphasis on the United Kingdom excludes far more research than it of receiving the rebate, and that the (11489, 9207). includes, and I apologize to the spending effects were greatest for house- Another active topic of research on researchers whose work is not mentioned holds with low levels of financial wealth. the individual income tax is the link in this summary. Other research has examined how the between current tax rules and incentives changes in investment incentives affected for entrepreneurial activity. Two studies The Economic Effects of Tax corporate investment activity (10415), have explored how the progressive struc- Reform and how the reduction in individual ture of the individual income tax, and the income tax liabilities raised the impor- interplay between the individual and the There have been three important tance of the Alternative Minimum Tax corporate income tax, affect entrepre- federal tax changes in the last five years. for many individual taxpayers (10072). neurial activity (9226, 9015). The impor- The 2001 Economic Growth and Tax One of the central issues in analyzing any tant role that start-up businesses play in Relief Reconciliation Act reduced mar- tax reform that changes marginal tax supporting research and development ginal tax rates under the federal income rates is how it will affect the amount of and encouraging job growth makes it tax, although budgetary pressures neces- taxable income reported on tax returns, important to understand how tax incen- sitated temporary rather than permanent since this determines the revenue effects tives affect the creation of new enterpris- tax reductions. The 2003 Job Growth of the tax reform. NBER researchers es. Other topics that have attracted atten- and Taxpayer Relief Reconciliation Act have played a central role in developing tion are the influence of tax incentives on reduced marginal tax burdens on divi- estimates of the elasticity of taxable purchases of health insurance and health dend income by as much as 20 percent- income with respect to marginal tax rates, care (10977, 9567, 8657, 9855), the link age points for some households, thereby and recent work supports earlier findings between taxation and labor supply

------NBER Reporter Winter 2005/6 3 (10316, 10935, 10139, 9429, 8774), the of participant contributions may play in out rates to study how firm characteris- sensitivity of capital gain realizations to raising participation rates (10419). tics, such as stock option holdings of top marginal income tax rates (10275, 9674, Another strand of research has consid- managers, affected the change in payout 8745), the economic effects of excise ered how assets held in tax-deferred in the aftermath of the tax change. taxes on goods such as cigarettes (8872, accounts should be valued from the per- Dividend increases were smaller at firms 8777) and alcohol (8562), and the impact spective of a household trying to com- where managers have substantial hold- of tax incentives on charitable giving pute a balance sheet that includes both ings of options that would decline in (10374). The Earned Income Tax Credit taxable and tax-deferred assets (10395). value if the firm paid out earnings as div- and its effect on the labor supply of low- Related work has studied the asset alloca- idends than at firms without such execu- income households has been widely stud- tion choices that households make when tive option holdings (11002). ied (11768, 11454, 11729). While many of they participate in tax-deferred accounts, While the tax treatment of dividends these studies emphasize specific issues of and it has contrasted these choices with has attracted particular attention in the income tax policy, NBER researchers are the predictions of simple models of tax- last two years, NBER researchers have still examining the broad issues raised by efficient asset location (9268). also studied many other aspects of cor- fundamental tax reform, such as the eco- The tax changes of 2003 focused on porate income taxation. The decline in nomic effects of shifting toward a con- the taxation of corporate capital income. corporate tax payments during a period sumption tax (9492, 9596), the funda- By reducing the maximum individual of high profitability, and the popular mental determinants of tax evasion income tax rate on dividend income to 15 claim that U.S. corporations were moving (8551), the theory of optimal taxation percent, instead of the top rate of more operations offshore to reduce their tax (10490, 10407, 10119, 10099, 9415, than 35 percent that prevailed in previous burden, have attracted an expanding set 9046), and the accurate measurement of years, the 2003 reform substantially of researchers to issues of corporate tax- the distribution of tax burdens across reduced the tax incentive for firms to ation. Two recent studies have explored households (8978, 8829). retain earnings or repurchase shares the source of the decline in corporate tax Discussions of consumption taxa- rather than to distribute cash dividends. revenues (9477, 9535), and a substantial tion and of fundamental tax reform NBER Public Economics researchers body of research has investigated the focus attention on the current tax rules have been studying the link between tax effect of international tax rules on the that affect saving, and in particular on rules and corporate financial policy since behavior of multinational firms (11717, opportunities for earning before-tax the program was created, as the 11196, 10806, 10936, 8854). These stud- returns in a variety of specialized “Business Taxation and Finance” group, ies generally find that large disparities in accounts such as Individual Retirement nearly thirty years ago. Not surprisingly, effective tax burdens across nations have Accounts and 401(k) plans. The analysis the dramatic change in dividend tax bur- the effect of shifting the geographical of retirement saving programs has been a dens stimulated many new research proj- pattern of reported income and of some very active area of research in Public ects. These include new studies of the corporate activities, although internation- Economics, and the ongoing research responsiveness of dividend payout with al tax considerations do not appear to has drawn insights from behavioral eco- respect to tax rates (10321, 10391, 10572, fully account for changes in corporate tax nomics (11518), financial economics, and 10841, 11449), and of the impact of the receipts over time. Other studies have many other sub-fields. In addition to 2003 tax reform on the market value of analyzed the determinants of corporate studying how the availability of these sav- firms with different payout policies tax avoidance and tax planning (11241, ing programs affects wealth accumula- (11452). The empirical findings suggest 11341, 10858, 10690, 10471, 11504) and tion (11680, 9096, 8610), a number of that in the months following the dividend the interplay between tax avoidance and studies have shown that participants in tax reduction, firms increased dividend financial fraud (10978). Researchers have employer-provided saving programs are payouts at a rate that had not been seen investigated the differences between tax- very sensitive to default options, peer for several decades. The tax change was a able income and book income, and the behavior, and other considerations that catalyst that reversed a decades-long potential consequences of moving are usually outside the neoclassical eco- decline in corporate dividend payout. toward a tax system that relied to a greater nomics analysis of saving choices (11554, Many firms that were paying dividends extent on book income for the computa- 11726, 9131, 8885, 8655). Research has increased their payouts, and many other tion of tax liability (11067, 8866), as well explored the potential use of plan default firms initiated cash dividends. Research as the role of new and sophisticated provisions to encourage saving (11074) on dividend policy has moved beyond financial products in affecting corporate as well as the role that employer matching the simple documentation of higher pay- income tax liabilities (9243).

4 NBER Reporter Winter 2005/6 ------Much of the recent research on cor- taxes and spending of unifying the budg- as the sensitivity of such calculations to porate income taxation has focused on ets of the Medicare and Social Security various assumptions (11060, 10969, emerging issues in the corporate sector, trust funds with the rest of the budget 10085, 9845). Other work has examined and a number of studies have linked this (10953), and the link among anti-deficit popular perceptions of future Social work back to long-standing concerns rules such as limits on government bor- Security benefits (9798); these percep- such as the economic incidence of the rowing and fiscal policy outcomes tions can have an important effect on corporate income tax (11686, 9916, (10788, 11065). More generally, recent current saving decisions. 9374). Other studies have also explored policy debates concerning federal deficits The projected shortfall of Social potential reforms of the current corpo- have led to renewed interest in the evolu- Security payroll taxes relative to benefit rate tax structure by analyzing the design tion of fiscal policy in the United States payments has stimulated numerous pro- of cash-flow corporate taxes in open and elsewhere (11630, 11600, 10788, posals for Social Security reform in the economies (10676, 9843) and the way 10023, 9012), and to new analyses of United States (8592, 11098). Some that the General Agreement on Tariffs how budget deficits and government researchers have explored the aggregate and Trade (GATT) might treat various debt levels affect interest rates (10681). efficiency effects of adopting a “private changes in the apportionment rules that Recent research on this issue has moved accounts” Social Security program are applied to the worldwide income of beyond earlier studies that considered (11622, 11101). Others have focused on multinational firms (9060). only the contemporaneous correlation specific design features of “private The individual and the corporate between asset markets and deficits, and accounts” programs, such as alternative income taxes are the focus of most begun to model expected future fiscal asset allocation restrictions and return Public Economics research on tax policy, policy and its impact on interest rates. guarantee programs for such accounts but there is always some research on (11300, 11084, 9195, 8906, 8732, other tax instruments. Several recent Social Security and Other 8731). The experience of Chile, a nation studies have examined the estate tax and Social Insurance Programs that adopted a privatized account system tried to summarize its incentive effects, in the early 1980s, has been carefully One of the reasons that long-term both in theoretical models (11408) and in chronicled (8924), and researchers have fiscal policy projections have attracted so practice (9456, 11025, 9661, 11767). One tried to predict the labor market effects much interest is the impending growth of explanation for the substantial flow of of a private accounts system (10305). Social Security and Medicare, programs wealth from one generation to the next is An important strand of research has that provide retirement income and that elderly households hold wealth to considered the labor market effects of health insurance to elderly households. prepare for the possibility of late-life existing Social Security programs either These programs represent the federal medical needs or other costs. This sug- by exploiting international differences in government’s largest long-term commit- gests that the impact of the tax code on Social Security programs to generate dif- ments. It is therefore no surprise that household behavior may be affected by ferences in retirement incentives (11290, these programs have been the focus of the structure of insurance markets. An 9407) or by examining the incentives an active research agenda by scholars emerging literature is beginning to created by the Social Security program affiliated with the Public Economics explore this interaction (11185). in the United States (10905, 9183, Program. Much of this research is also Most of the research described 10030). Social Security redistributes part of the NBER Programs on Aging above focuses on the detailed provisions resources within cohorts as well as and on Health Care, and is consequently of the tax rules affecting individuals or across generations, and recent research summarized in other Program Reports. firms. But the unusual nature of the 2001 has examined how the Social Security Because the issues in Social Security tax changes, in particular their temporary program has affected the economic sta- reform are particularly central to public character and the role of budgetary rules tus of the elderly and its distribution economics, I will describe several comin leading Congress to enact such tax (10466, 8911,8625). ponents of this research, and then dis- policies, has stimulated new research on While a substantial group of NBER cuss social insurance research more gen- the broad subject of budget rules and the researchers studies issues related to Social erally. link between such rules and policy out- Security, an even larger group investigates A number of studies have consid- comes such as the budget deficit. the wide array of other social insurance ered the long-term fiscal health of the Researchers have studied the effect of programs that currently operate in the Social Security program and computed sunset provisions on budget outcomes United States and other developed the present discounted value of prom- (10694), the potential impact on federal nations. This research touches on many ised payouts less projected taxes, as well

------NBER Reporter Winter 2005/6 5 different programs and topics. Some creation of government-provided social Public Economics Program, but many of work offers theoretical guidelines for the insurance programs, while other work them are united by a central focus on the design of social insurance programs recognizes that the provision of social determinants of electoral or legislative (11386, 11250, 10792). Two central issues insurance may alter the operation of pri- rules, and the consequences of different that arise in evaluating any social insur- vate insurance markets (11039, 10989, rules. Some work offers an explanation ance program are: the extent to which the 9714, 9031). The welfare effects of pub- for the political factors that underpin the program alleviates the problem that it is lic programs can be very sensitive to the choice of different electoral rules in dif- designed to address and the extent to private market response. Studies of char- ferent U.S. cities (11236). Another strand which it causes unintended distortions in itable work by religious organizations, of research examines the factors that the behavior of recipients (8730). Many and how such work is affected by the influence bargaining power within legisla- studies have examined one or both of provision of government transfer pro- tures (10530, 8973) and the link between these issues in the context of specific grams, represent a new direction for pub- such power and legislative outcomes social insurance programs, such as unem- lic economics researchers. Analysis of (10385, 9748). A third group of studies ployment insurance (11760, 10500, transfers during the New Deal suggests examine an even more general set of 10443, 10043), the Supplemental Security that as public spending on anti-poverty issues about the links between electoral Income program (11568), Medicaid efforts increased, private spending rules, the structure of political parties, (9058), Temporary Assistance for Needy through church-based relief efforts and the choice of economic policies Families (TANF) (8749), disability insur- declined. This suggests a novel channel (10176, 10040). Political institutions are ance (9155, 9148, 10219), housing assis- of crowd-out that has not been docu- increasingly recognized as affected by the tance programs (8709), and child care mented heretofore. Future research on underlying tastes of voters, the power of subsidy programs (9693). Other studies social insurance will undoubtedly contin- various interest groups, and the history of focus on various aspects of behavioral ue to explore both the way that potential political jurisdictions (9006). response that arise in a number of differ- beneficiaries respond to public programs, The recognition that political institu- ent social insurance programs. These and the effect of public programs on tions matter for policy outcomes raises include the effect of such programs on other components of the economic sup- the related question of whether one set household saving (10487), the decision of port network. of institutions may be more efficient in households with regard to benefit take- responding to some types of economic up (10488, 9818), the impact of social Political Economy, Legis- problems than another institutional insurance and transfer programs on labor lative Structure, and Policy structure. One specific context in which supply (9168), and the link between social Outcomes researchers have explored this question insurance programs and living arrange- concerns the choice between an appoint- ments (8774). As a result of empirical Much of public economics is con- ed regulator and an elected politician as studies such as these, policymakers have a cerned with the consequences of various the decisionmaker in particular settings much better description of the key inputs government policies. Research on the (10241). The findings suggest that politi- to social insurance program design. incidence of various taxes and on the cians will be more likely to outperform Researchers in public economics behavioral effects of various transfer regulators in settings that require com- have long recognized that it is important programs fits this description. An impor- pensating the losers from a policy action, to study government-provided social tant and growing strand of research, that do not involve specialized technical insurance programs in a broad context however, seeks to understand the link expertise, and that do not feature small that recognizes the many other ways in between political institutions and policy but powerful vested interests that benefit which resources may be transferred to outcomes. This work asks why certain from or lose from the policy choice. households that experience adverse eco- policies are enacted, not how such poli- Another broad issue of interest in nomic shocks of various kinds. Transfers cies would affect economic activity. This political economy concerns “election within families are one such alternative research on “political economy” crosses mechanics.” This area is concerned with mechanism. Recent research has empha- several NBER Programs, including the factors that affect voting, campaign sized two others: private insurance mar- Economic Fluctuations and Growth, spending, candidate selection, and elec- kets and transfers from religious organi- Industrial Organization, and Public toral outcomes. One example of such zations. Some research has considered Economics. research is the attempt to understand and how imperfections in private insurance There are many different elements of explain the apparent electoral advantage markets can provide a rationale for the political economy research within the of incumbent officeholders (10748).

6 NBER Reporter Winter 2005/6 ------Another is the analysis of voter participa- economics, as in its federal counterpart, is found, and, I expect, will continue to tion. Several studies have explored eco- the behavioral response of taxpayers find, an astonishing array of research nomic and other factors that influence confronting the tax system. An example topics to study. voter turnout (9896, 8720, 10797, 11794), of research related to this issue is a study as well as the likelihood that an election is of how property tax measures that Government Service close enough for an individual voter to grandfather a taxpayer’s taxable property In part because members of the rationally believe that she might have a value affect homeowner mobility Public Economics Program devote their significant chance of affecting the out- (11108). School finance also provides energies to studying government policies, come (8590). Other work explores the many opportunities for analyzing taxpay- they are frequently invited to effect of campaign finance rules on the er response. Recent work has analyzed serve in various governmental roles. A influence of interest groups and on poli- school finance reform programs in spe- substantial fraction of the Research cy outcomes (9601, 8693), and the broad cific states, such as Texas (10722), and Associates in the Program have devoted question of what determines the total looked at the effect of state-level aid pro- part of their careers to high-level policy amount of campaign contributions and grams on the tax and spending decisions advisory roles. This historical pattern has campaign spending (9409). of local school districts (10701). A fea- continued in recent years, as many pro- One intriguing line of research ture that distinguishes state tax analysis gram affiliates have taken a break from explores how the identity of elected offi- from its federal counterpart is the possi- their academic research and spent time in cials, which may be affected by electoral bility of taxpayer mobility across jurisdic- policymaking roles in Washington. rules, influences policy outcomes. This tions. One study (10645) explores the Research Associate R. Glenn Hubbard work evaluates an Indian electoral reform effect of such mobility in the context of served as the Chairman of the Council of that reserved a significant share of elect- state estate taxes, and finds some evi- Economic Advisers (CEA), while Harvey ed positions on local councils for women dence suggesting that older taxpayers Rosen, Mark McClellan, and Katherine candidates (8615). After this electoral with substantial estates migrate to states Baicker have served, or are serving, as reform, the set of policies chosen by the with low estate taxes. Another study members of the CEA. In addition to his local councils shifted toward support for examines the impact on local finances of role at CEA, Mark McClellan has also public programs that would be particular- winning a multi-jurisdiction battle for a served as Commissioner of the Food and ly beneficial for women rather than men. new plant (9844). A third study explores Drug Administration and as Director of These findings reinforce other studies how the ease of inter-jurisdictional the Centers for Medicare and Medicaid that suggest the important role played by mobility affects the relationship between Statistics. Jeffrey R. Brown has been electoral institutions that affect the char- tax rates and revenue collections (9686). nominated to the Social Security Advis- acteristics of winning candidates. Other Directions for Research ory Board. I served on the President’s State and Local Public Advisory Panel on Federal Tax Reform in The research summarized in the four 2005. Douglas Holtz-Eakin served as the Finance foregoing topic areas represents only a Chief Economist of the Council on While much of the tax policy debate fraction of the work carried out by Public Economic Advisors, and then as in recent years has centered on the feder- Economics researchers. Several other Director of the Congressional Budget al government, many important and studies, not mentioned above, illustrate Office. Mervyn A. King was appointed ongoing tax policy issues affect state and this range. Program affiliates have studied Governor of the Bank of England in local governments. Recent research by the design of terrorism insurance (10179, 2003. While some members of the Public Economics Program affiliates has 9271), the detection of teacher cheating Public Economics Program make life- examined a number of these issues. on behalf of students taking standard- time commitments to participate directly There are wide disparities in fiscal struc- ized tests (9413, 9414), and the effects of in the policy process, and they serve in a ture across states and localities. This pro- differential tax treatment of different variety of policy roles, many other mem- vides much wider variation in tax policies sized families on fertility behavior (8845). bers have taken only a single job in than at the federal level, and also results Because the public sector is involved in Washington and then returned to their in large differences across locations in fis- some way with virtually every aspect of academic careers. cal balance (11203). modern life in industrial democracies, A key issue in state and local public researchers interested in this field have

------NBER Reporter Winter 2005/6 7 Research Summaries

The Contribution of Science and Technology to Production

James Adams*

Economists have long recognized as industrial inventions or discoveries biased towards skilled labor, so that the that knowledge is a factor of produc- in basic science. From this root idea skill bias of firm R and D is localized tion, and even the most important fac- there flow a number of subsidiary in technology space.1 In addition, firm tor, given its role in labor quality and ideas. One is the reshaping of goods and industry R and D shift investment the design of capital goods. Still, it is production and the redirection of in plant capital towards equipment one thing to assert a general proposi- Research and Development (R and D) capital. This link should not be over- tion and quite another to provide con- that result from the accumulation of looked because equipment turns out to firmation of it in detail. My research is knowledge. A second is the distinction be skill-biased. Thus the skill bias of R part of a larger initiative at NBER that between knowledge that is internal to and D takes place through two distinct seeks to provide this information. In an organization, and outside knowl- channels, a direct one that operates essence, the work is a search for tangi- edge, or knowledge spillovers. A third through the small part of R and D that ble evidence of flows of knowledge, theme is the importance of limitations is targeted on the plant, and an indirect specifically scientific and technical on flows of outside knowledge or and potentially much larger one that knowledge, followed by an examina- knowledge spillovers that are imposed operates through the accumulation of tion of their effects on firms and other by absorptive capacity, human and equipment capital. institutions. Of course private incen- institutional constraints, and the intrin- The accumulation of outside tives, internal organization, public pol- sic relevance of the information. A knowledge, or knowledge spillovers, icy, and legal structure all affect the use fourth theme is the comparable impor- could alter the rate and direction of of science and technology by firms, tance of basic and often academic sci- industrial R and D. Using survey data universities, and federal laboratories. ence for production, besides that of from industrial R and D laboratories as Thus, broader aspects of modern industrial R and D. Finally, the research well as historical case studies, I find economies and of modern economics recognizes the role that contract that outside knowledge shifts R and D govern the role of knowledge in pro- design and public policy play in delib- effort towards learning about external duction. These provide many opportu- erate knowledge transfer between research and away from internal nities for research. firms and outside R and D performers. research.2 Similarly, in cross-equation The basic idea of the research is to These in turn influence the limits of tests I find that university R and D begin by specifying a vector of stocks the firm in R and D. In pursuing each increases learning expenditures target- of past knowledge flows in the pro- of these themes, the design, collection, ed on academia, and industrial R and duction function. The production and assembly of new and high quality D increases learning expenditures function may specify outputs of final economic data forms a critical part of devoted to industry, but not converse- or intermediate goods or it may speci- the work. ly. These results are observationally fy increments of new knowledge, such consistent with the view that outside Characterizing the Contri- opportunities alter the composition of bution of Knowledge industrial R and D, presumably in * Adams is a Research Associate in the more profitable directions, and are Using data on plants owned by NBER's Productivity Program and is a consistent with the historical case stud- chemical firms that span manufactur- Professor of Economics at Rensselaer ies. ing, I have found that firm R and D in Polytechnic Institute. His profile appears In all of this research, where the later in this volume. the same product area as the plant is

8 NBER Reporter Winter 2005/6 ------data allow a comparison I find statisti- citations divided by potential citations channel of collaboration in science.9 As cally significant effects of university sci- within cells that are classified by citing an alternative to citation, collaboration ence as well as industrial R and D on and cited fields and years. We estimate is undoubtedly more costly and more industrial R and D and industrial citation functions using the citation time-intensive but it offers the chance patents.3 Thus basic science as well as probability as the dependent variable, to acquire tacit knowledge that would applied research and development are where field and year effects are the not be available otherwise. important to industrial research. independent variables.7 Assuming that The paper describes trends and Another set of findings concerns citations represent scientific influence cross-sectional patterns in scientific limits on the influence of outside of papers cited, this probability is teams measured by authors per paper, knowledge on R and D performing equivalent to a utilization rate of cited and in institutional collaboration, meas- firms. In work with Adam Jaffe, I find literature by an average citing paper. ured by the location of team members that the effect of firm R and D on plant Thus, our finding that the citation prob- in separate institutions. The data are productivity is amortized by geographic ability is 10 to 100 times greater within steeply trended. Team size increases by and technological distance. We also find fields than between fields can be read to 50 percent over the sample period. that the number of plants in a firm and imply that field boundaries amount to However, counts of institutions per industry dilute the impact on productiv- technological barriers, in part because paper increase by 60 percent. Counts of ity of firm R and D and of industry R of decreased relevance. The fact that foreign institutions, while comparatively and D spillovers.4 These results suggest cross-field citation parameters resulting rare, increase by five-fold. We conclude restrictions that may apply to economy- from the estimation are statistically sig- that team workers in science are becom- wide returns from spillovers. In other nificant in less than one fourth of the ing more geographically and even inter- work, I find that knowledge spillovers possible cases only serves to reinforce nationally dispersed. This trend acceler- from universities are more localized this conclusion. In the same paper we ates around the start of the 1990s, sug- than spillovers from other firms.5 This find within fields that citation probabil- gesting a decline in costs of collabora- finding is curious because published ities are greater from lesser universities tion. Our hypothesis is that the deploy- findings should not be localized. The to top universities than conversely, and ment of NSFNET and its connection puzzle is explained by the industry-uni- we find that citations to peer institu- to networks in Europe and Asia in the versity cooperative movement, which tions increase with rank. These results late 1980s are responsible for this encourages firms to work with local suggest that scientific influence increas- change. The hypothesis is not unrea- universities. The universities are subject es with quality of university depart- sonable, given research and journal to incentives that allow firms to make ments, which levels the capabilities of publication lags. use of their capabilities and to gain diverse institutions, but that reinforcing In addition the paper explores rea- access to the wider world of scientific effects of quality among peer institu- sons for teams and institutional collabo- research. The same is not true of access tions may instead sustain differences in rations. We find that more highly ranked to proprietary knowledge in other the capabilities of institutions. departments, departments whose scien- firms. In assessing the significance of the tists have earned prestigious awards, citation channel it is important to con- departments with larger stocks of feder- Channels of Knowledge Flow sider alternative channels of knowledge al R and D, and departments in private flow. This is despite the fact that in the universities are more likely to form large In work with J. Roger Clemmons literature of industrial R and D, one key teams and to engage in institutional col- and Paula Stephan that uses data on sci- channel of knowledge flow is found to laboration. In the case of firms and for- entific publications, a counterpart to be the scientific literature. The citation eign institutions especially, we find that industrial patents, I also find that tech- channel can be thought of as disem- placement of graduate students signifi- nological distance and other factors bodied and informal, in that it does not cantly increases collaboration. Finally, limit knowledge flows among universi- require meetings or formal knowledge- the evidence suggests that scientific out- ties.6 In this case we explore a citation sharing agreements, but it is not all- put and influence increase with team channel of knowledge flow that is con- inclusive.8 In recent research with J. size and institutional collaboration. ditional on reading and afterwards con- Roger Clemmons, Grant Black, and Since these factors imply an increase in tributing to the science literature. The Paula Stephan, which uses the same the division of labor, the results suggest size of the channel is summed up by the data on publications as the citation that scientific productivity increases citation probability. This equals actual study, I have explored an alternative with the scientific division of labor.10

------NBER Reporter Winter 2005/6 9 Limits of the Firm in R and D D sourcing in industrial laboratories.14 R&D and the Factor Intensity of Production, In this paper we find that sourcing and Skill Bias,” The Review of Econ- Consistent with the literature of appears to be driven by sentiments omics and Statistics 81 (August 1999): Property Rights Economics, contractu- towards Research Joint Ventures pp. 499–510. 2 J. Adams, “Endogenous R&D Spillovers al design and public policy clearly influ- (RJVs), the option to purchase and and Industrial Research Productivity,” ence the extent to which firms turn to acquire, and research with federal labo- outside partners for complementary R NBER Working Paper No. 7484, January ratories. When we turn to the effects of 2000, revised and extended, and published as and D assets and the extent to which sourcing, the evidence suggests that the “Learning, Internal Research, and 11 they benefit from knowledge transfer. primary motive is that of cost-saving. Spillovers,” The Economics of In papers that use the data on R and D This contrasts with RJVs, which con- Innovation and New Technology 15 laboratories alluded to in earlier sec- tribute to new products, and with inter- (January 2006): pp. 5–36. 3 tions of this article, I have explored this nal research, which contributes to both J. Adams, “Learning, Internal Research, and Spillovers,” and J. Adams, “Comparative aspect of the practice of industrial R patents and new products. All of this and D. Localization of Academic and Industrial suggests that deliberately shared R and Spillovers,” NBER Working Paper No. 8292, In work with Eric Chiang and D comes in different varieties designed May 2001, and J. Adams, “Comparative Jeffrey Jensen, I find that Cooperative to meet different objectives of firms. Localization of Academic and Industrial Research and Development Agree- Spillovers,” The Journal of Economic ments (CRADAs) comprise the main Ongoing Research Geography 2 (July 2002): pp. 253–78, channel by which federal laboratories reprinted in Clusters, Networks, and increase patents as well as firm R and Along with coworkers, I continue Innovation, S. Breschi and F. Malerba, eds, D. 12 The CRADA effect survives con- to study the role of science and tech- forthcoming, Oxford University Press. 4 trols for simultaneous equation bias, it nology in production. At present we are J. Adams and A. Jaffe, “Bounding the Effects of R&D: An Investigation Using survives inclusion of alternative effects engaged in a study of the factors that determine the speed of diffusion of sci- Matched Firm-Establishment Data,” of federal laboratories on firms, and it NBER Working Paper No. 5544, April entific research across sectors and fields is consistent across patents and R and 1996, and J. Adams and A. Jaffe, “Bounding D expenditure in industrial laboratories. of science, including a comparison of the Effects of R&D: An Investigation Using While subject to justifiable skepticism the speed of diffusion of science with Matched Firm-Establishment Data,” about the usefulness of incentives in that of patented technology. We are RAND Journal of Economics 27 this setting, the results suggest that also engaged in studies of the determi- (Winter 1996): pp. 700–21. 5 CRADAs may be beneficial precisely nants of industrial scientific discovery, J. Adams, “Comparative Localization of of the relationships between firm Academic and Industrial Spillovers.” because of the mutual effort that they 6 patents and stock market value, and sci- J. Adams, J.R. Clemmons, and P. Stephan, require of firms and government labo- “Standing on Academic Shoulders: Measuring ratories. In another paper with Chiang entific research both inside and outside Scientific Influence in Universities,” NBER and Katara Starkey, I have found that the firm. I continue to pursue long- Working Paper No. 10875, November 2004, Industry-University Cooperative standing interests in research contribu- forthcoming in Les Annales d’Economie Research Centers (IUCRCs) also con- tions of the university system.15 This et de Statistique. 7 tribute to research productivity of system is not only a current hotbed of A. Jaffe and M. Trajtenberg, “International Knowledge Flows: Evidence from Patent industrial laboratories.13 Their effect ideas, but the health of the system Citations,” The Economics of Innovation entails the participation of university going forward may prove critical to the United States and other economies. In and New Technology 8 (1999): researchers in consulting, collaboration pp. 105–36. and placement with firms. Both conclusion, I am confident that the 8 W.M. Cohen, R. Nelson, and J. Walsh, CRADAs and IUCRCs are incentive- study of the contributions of science “Links and Impacts: The Influence of Public based policy initiatives put in place and technology to the economy will Research on Industrial R&D,” Manage- around 1980 whose aim was to promote provide grist for the economists’ mill ment Science 48 (January 2002): pp. 1–23. 9 knowledge transfer from the public sec- for years and even decades to come. J. Adams, G. Black, J. R. Clemmons, and P. Stephan, “Scientific Teams and tor to private industry. The evidence Institutional Collaborations: Evidence from contained in the two papers suggests 1 J. Adams, “The Structure of Firm R&D U.S. Universities, 1981–1999,” NBER that they may have had an effect. and the Factor Intensity of Production,” Working Paper No. 10640, July 2004, and J. Finally, in a third paper with Mircea NBER Working Paper No. 6099, July Adams, G. Black, J. R. Clemmons, and P. Marcu I explore the behavior of R and 1997, and J. Adams, “The Structure of Firm Stephan, “Scientific Teams and Institutional

10 NBER Reporter Winter 2005/6 ------Collaborations: Evidence from U.S. Chiang, and J. Jensen, “The Influence of Knowledge and Productivity Growth,” Universities, 1981–1999,” Research Federal Laboratory R&D on Industrial Journal of Political Economy 98 (August Policy 34 (April 2005): pp. 259–85. Research,” The Review of Economics 1990): pp. 673–702. See also J. Adams and 10 J. Adams, G. Black, J. R. Clemmons, and and Statistics 85 (November 2003): Z. Griliches, “Measuring Science: An P. Stephan, “Scientific Teams and pp.1003–20. Exploration,” NBER Working Paper No. Institutional Collaborations: Evidence from 13 J. Adams, E. Chiang, and K. Starkey, 5478 (Reissued July 1997), and J. Adams U.S. Universities, 1981–1999.” “Industry-University Cooperative Research and Z. Griliches, “Measuring Science: An 11 D. Mowery, “The Boundaries of the U.S. Centers,” NBER Working Paper No. 7843, Exploration,” Proceedings of the Firm in R&D,” in Coordination and August 2000, and J. Adams, E. Chiang, and National Academy of Sciences 93 Information: Historical Perspectives on K. Starkey, “Industry-University Cooperative (November 1996): pp.12664–70; and J. the Organization of Enterprise, N. Research Centers,” The Journal of Adams and Z. Griliches, “Research Lamoreaux and D. Raff, eds., Chicago: Technology Transfer 26 (January 2001): Productivity in a System of Universities,” University of Chicago Press for NBER, pp.73-86. NBER Working Paper No. 5833, 1995. 14 J. Adams and M. Marcu, “R&D November 1996, and J. Adams and Z. 12 J. Adams, E. Chiang, and J. Jensen, “The Sourcing, Joint Ventures, and Innovation: A Griliches, “Research Productivity in a Influence of Federal Laboratory R&D on Multiple Indicators Approach,” NBER System of Universities,” Les Annales Industrial Research,” NBER Working Paper Working Paper No. 10474, May 2004. d’Economie et de Statistique 49/50 No. 7612, March 2000, and J. Adams, E. 15 J. Adams, “Fundamental Stocks of (1998): pp.127–62.

Foreign Direct Investment Behavior of Multinational Corporations

Bruce A. Blonigen*

There is increasing recognition ly 50 percent of U.S. trade flows occur- hypotheses about, say, when a factor that understanding the forces of eco- ring between affiliates of the same should matter for FDI, or even just a nomic globalization requires looking MNC, or what is termed “intra-firm particular form of FDI, and then find- first at foreign direct investment (FDI) trade”.1 ing creative ways to test these hypothe- by multinational corporations (MNCs): Despite the obvious importance ses in the data. that is, when a firm based in one coun- of FDI and MNCs in the world econ- try locates or acquires production facil- omy, research on the factors that Exchange Rates and FDI ities in other countries. While real determine FDI patterns and the One good example of this is the world GDP grew at a 2.5 percent impact of MNCs on parent and host effect of exchange rate movements on annual rate and real world exports countries is in its early stages. The FDI. For years, the conventional theo- grew by 5.6 percent annually from most important general questions are: ry was to compare FDI to bonds, for 1986 through 1999, United Nations what factors determine where FDI which exchange rate movements do data show that real world FDI inflows occurs, and what impacts do those not affect the investment decision. A grew by 17.7 percent over this same MNC operations have on the parent depreciation of the currency in the period! Additionally, MNCs mediate and host economies? As I discuss in a host country reduces the amount of most world trade flows. For example, recent survey of the empirical literature foreign currency needed to purchase Bernard, Jensen, and Schott find that addressing the first question — the the asset, but it also reduces the nomi- 90 percent of U.S. exports and imports determinants of FDI decisions — the nal return one receives in the foreign flow through a U.S. MNC, with rough- answers are not straightforward.2 In currency. Thus, the rate of return for particular, the literature has shown that the foreign investor does not change. we cannot simply conclude that factors Empirical studies of FDI seemed to * Blonigen is a Research Associate in the such as exchange rates or tax policies NBER's Program on International Trade and confirm this, often finding insignifi- have an unambiguous general impact Investment and the Knight Professor of Social cant effects of exchange rates. In con- on FDI patterns. Instead, meaningful Science at the University of Oregon. His profile tradiction to this, the popular press appears later in this issue. insights come from developing

------NBER Reporter Winter 2005/6 11 often points to host-country exchange MNCs based in the other country. The Tomlin and Wilson finds that domestic rate depreciations as a contributing fac- Organisation for Economic Co-opera- firms experience a 3 percent increase in tor to inward foreign investment tion and Development (OECD) has expected discounted profitability from booms, and worries about the selling of been a big advocate of BITs as a way to antidumping duties unless the foreign key national technological assets. enhance FDI across member countries. firms subject to the duties decide to tar- I find a resolution to this puzzle by Others contend that BITs are mainly iff-jump, in which case the domestic considering FDI that involves firm-spe- intended to share tax information firms do not experience any increase.9,10 cific assets (such as patents or manage- across countries in order to deter tax rial skills) — the type of assets previous evasion and to reduce administrative Information and FDI literature established as crucial to for- costs and, thus, should have little, or An almost unexplored issue in the mation of MNCs and FDI.3 Such assets even negative, effects on FDI flows.5 literature has been the role of informa- are typically intangible and easily trans- Ron B. Davies and I examine whether tion on FDI decisions. FDI requires ferred across a firm’s operations. Thus, the empirical evidence suggests that substantial fixed costs of identifying an the purchase prices of such assets such treaties increase FDI flows across efficient location, acquiring knowledge through FDI are in the host-country’s nations, as the OECD and many econ- of the local regulatory environment, currency, but returns can be generated omists presume.6 In separate studies, we and coordination of suppliers. Thus, anywhere the firm operates and are not examine the evidence for the U.S. and access to better information about necessarily tied to the home country’s for OECD BITs, respectively, in panel some host countries may make FDI to currency. This means that host-country data that span a variety of bilateral that location more likely. Ellis, Fausten, currency depreciations theoretically can country pairs over time. Across these and I find an interesting avenue for lead to increased acquisition of FDI, various samples and numerous specifi- investigating this hypothesis using particularly of firms that have firm-spe- cations, we find little evidence that information on Japanese industrial cific assets. This hypothesis is strongly these BITs increase FDI activity, a sur- groups called keiretsu.11 Horizontal keiret- confirmed for a panel of acquisitions of prising result in light of OECD promo- su are groups of firms across a wide U.S. firms by Japanese and German tion of these treaties. range of industries, typically centered firms and provides evidence for the around a main bank that owns signifi- notion in the popular press that curren- Trade Protection and FDI cant shares in these firms. A number of cy depreciations ease foreign firms’ pur- The notion that trade protection studies have focused on the potentially chases of U.S. host-country technologi- encourages FDI is folk wisdom for favorable financing received by keiretsu cal assets. economists, so much so that it is rarely firms from their main bank as one Taxes and FDI examined empirically. But my research impetus for greater investment by these into this relationship has also yielded firms, including FDI — but the evi- Another factor that the literature surprises. In a study examining all U.S. dence is mixed on this. However, the finds does not affect FDI in a straight- antidumping trade protection actions major firms in a keiretsu also get togeth- forward manner is tax policy. MNCs are from 1980 through 1995, I find that er on a regular basis in what are termed potentially subject to taxation in both FDI responses to these trade actions Presidential Meetings and presumably the host and parent country. However, (tariff-jumping FDI) occur only for share information more than other most parent countries have policies to firms with previous experience as firms would. My work with Ellis and reduce or eliminate double taxation of MNCs.7 Most firms facing such trade Fausten examines whether this infor- their MNCs. James R. Hines, Jr. and co- policies (many from developing coun- mation affects FDI choices, by estimat- authors have shown that the way in tries) have no such experience and do ing how much prior-year FDI by mem- which parent countries reduce double not respond with FDI. Instead, these bers of a firm’s keiretsu in a particular taxation on their MNCs (for example, firms must face either significant host country increases the likelihood allowing credits or deductions) can have antidumping duties or go through the that the firm will also choose that coun- quite different implications for FDI costly process of raising U.S. prices and try for its FDI. We find that prior-year activity.4 requesting recalculations of the duties.8 investment by a firm in the same keiret- Many countries also have negotiat- For domestic firms, whether foreign su will raise a firm’s probability of located bilateral investment treaties (BITs) to firms tariff-jump the antidumping ing an investment in that same host mutually reduce withholding taxes on duties matters significantly. Work with country by about 20 percent.

12 NBER Reporter Winter 2005/6 ------A related paper with Wooster izontal model of FDI where an MNC across countries is still an open issue. examines whether U.S. firms increase replicates its process across multiple Until this is resolved, using these mod- overseas investments when a new CEO countries to avoid trade frictions, and els as control variables in studies of who is foreign-born takes over.12 Our Helpman’s vertical MNC model where how new factors of interest affect FDI examination of CEO turnover among firms locate their production process can be misleading. Fortune 500 firms in the 1990s does abroad to take advantage of lower fac- An additional concern is that MNC show evidence of significant increases tor costs.14 A recent important step by models typically use a two-country in FDI when a “foreign” CEO takes Carr, Markusen, and Maskus (CMM) framework and empirical FDI specifica- over. It is difficult to disentangle was estimation of empirical specifications use bilateral FDI data. This whether such an effect is attributable to tions of FDI based on general equilib- assumes that FDI decisions to different better information of foreign markets rium models of MNCs.15 Their work markets are independent. There are a by the foreign CEO or to different per- shows that other factors missing from number of reasons to think this may sonal preferences influenced by a less gravity-based FDI specifications, partic- not be true. For example, U.S. firms U.S.-centric perspective. Regardless, the ularly factor endowment differences, may prefer to locate FDI in one coun- results suggest that there are likely other are important for explaining FDI pat- try and then export to neighboring important factors behind FDI patterns terns. countries (export-platform FDI). In than the standard economic ones so In recent work with co-authors I this case, more FDI in a particular host often mentioned in the literature. have explored the central question of country would mean less in neighboring how well these specifications actually fit ones. Alternatively, U.S. firms may have Estimating Long-Run the real-world data we observe. The vertical production relationships General-Equilibrium empirical specification estimated by between affiliates such that more FDI Determinants of FDI CMM was a starting point in this in a country will naturally be associated research, since its inclusion of endow- with more in neighboring ones because Much of the literature described to ment differences clearly outperforms a of production externalities. Davies, this point motivates analysis with partial standard gravity equation of FDI. In Naughton, Waddell, and I explore this equilibrium models of individual firm- initial work with the model, Davies, by explicitly modeling spatial interde- level FDI decisions. But we also want to Head, and I found that the CMM model pendence in empirical estimation of have empirical specifications of FDI had a specification of endowment dif- U.S. FDI patterns.18 We find that spatial that are grounded in theory and that do ferences that was not consistent with interdependence shows up significantly a good job of explaining FDI patterns the theory. Once corrected, the model in the data, although the nature of these across the world. Researchers looking at no longer provides evidence that verti- spatial relationships is strongly affected world FDI patterns have generally used cal FDI motivations are very important by the particular geographic features of variations of a gravity framework to in overall FDI flows between coun- the sample of countries one chooses to model FDI, specifying parent- and tries.16 Work with Davies and Wang examine. However, our finding that the host-country GDPs along with distance shows that specification error goes coefficients on the standard control as core determinants of FDI. These beyond this with not only the CMM variables in FDI studies are hardly models seemingly do well to describe model, but also with the gravity specifi- affected by including these spatial con- FDI patterns statistically, but while cation.17 Data on FDI between coun- siderations is relatively good news for Anderson and van Wincoop have solid- tries are highly skewed, with very large previous work using these empirical ified an appropriate gravity specification activity between developed countries specifications. as theoretically valid for trade patterns, and small or even no activity for very it is not clear this is true for FDI pat- small countries. We show that even after Conclusion 13 terns. logging variables, adding country fixed- The study of FDI and MNCs is Of course, deriving a theoretically effects, and splitting samples into devel- both fascinating and important for based empirical specification of FDI is oped countries versus less-developed understanding economic globalization. a fairly complicated problem. General countries, one is still not guaranteed of There has been substantial progress in equilibrium theoretical models of having normally distributed error terms. the literature in the past couple of MNCs and their FDI activities only first In other words, finding an appropriate decades, but it is complicated enough began to be developed in the mid-1980s specification that effectively models the that, in many ways, we are still in the with Markusen’s development of a hor- substantial heterogeneity in FDI activity

------NBER Reporter Winter 2005/6 13 process of uncovering what we don’t Practices of U.S. Antidumping Activity,” “Gravity with Gravitas: A Solution to the know. I am excited to work on filling NBER Working Paper No. 9625, April Border Puzzle,” NBER Working Paper No. more gaps in our understanding in my 2003, and, forthcoming, Canadian Journal 8079, January 2001, and American future research efforts. of Economics, documents the rapidly rising Economic Review, 93(1), March 2003, trend in U.S. antidumping duties and the pp. 170–92. sources of this trend. B.A. Blonigen and S.E. 14 J.R. Markusen, “Multinationals, Multi- 1 A.B. Bernard, J.B. Jensen, and P.K. Schott, Haynes, “Antidumping Investigations and the Plant Economies, and the Gains from Trade,” “Importers, Exporters and Multinationals: A Pass-Through of Exchange Rates and Journal of International Economics, Portrait of the Firms in the U.S. that Trade Antidumping Duties,” NBER Working 16(3–4): pp. 205–26, and E. Helpman, A Goods,” NBER Working Paper No. 11404, Paper No. 7873, October 1999, and Simple Theory of International Trade with June 2005. American Economic Review, 92(4), Multinational Corporations,” Journal of 2 B.A. Blonigen, “A Review of the Empirical September 2002, pp. 1044–61, and B.A. Political Economy, 92(3), pp. 451–71. Literature on FDI Determinants,” NBER Blonigen and J.-H. Park, “ Dynamic Pricing 15 D.L. Carr, J.R. Markusen, and K.E. Working Paper No. 11299, May 2005, and in the Presence of Antidumping Policy: Maskus, “Estimating the Knowledge-Capital forthcoming, Atlantic Economic Journal. Theory and Evidence,” NBER Working Model of the Multinational Enterprise,” 3 B.A. Blonigen, “Firm-Specific Assets and Paper No. 8477, September 2001, and NBER Working Paper No. 6773, October the Link Between Exchange Rates and American Economic Review, 94(1), 1998, and American Economic Review, Foreign Direct Investment,” American March 2004, pp. 134–54, address the eco- 91(3), June 2001, pp. 693–708, and J.R. Economic Review, 87(3), June 1997, pp. nomics of firms’ strategic pricing decisions in Markusen, and K.E. Maskus, “Discrimi- 447–65. the face of antidumping duties. nating Among Alternative Theories of the 4 For example, see J.R. Hines, “Altered States: 9 B.A. Blonigen, K. Tomlin, and W.W. Multinational Enterprise,” NBER Working Taxes and the Location of Foreign Direct Wilson, “Tariff-jumping FDI and Domestic Paper No. 7164, June 1999, and Review of Investment in America,” NBER Working Firms’ Profits,” NBER Working Paper No. International Economics, 10(4), Paper No. 4397, May 1997, and American 9027, June 2002, and Canadian Journal of November 2002, pp. 694–707. Economic Review, 86(5), December 1996, Economics, 37(3), August 2004, pp. 16 B.A. Blonigen, R.B. Davies, and K. Head, pp. 1076–94, and M.A. Desai, C.F. Foley, 656–77. “Estimating the Knowledge-Capital Model of and J.R. Hines, “Foreign Direct Investment in 10 A related issue is how FDI may affect trade the Multinational Enterprise: Comment,” a World of Multiple Taxes,” NBER protection policies (that is, reverse causality), NBER Working Paper No. 6773, October Working Paper No. 8440, August 2001, and which I address with co-authors in B.A. 1998, and American Economic Review, Journal of Public Economics, 88(12), Blonigen and R.C. Feenstra, “Protectionist 93(3), June 2003, pp. 980–94. December 2004, pp. 2727–44. Threats and Foreign Direct Investment,” 17 B.A. Blonigen and R.B. Davies, “The 5 For example, see T. Dagan, “The Tax NBER Working Paper No. 5475, March Effects of Bilateral Tax Treaties on U.S. Treaties Myth,” New York University 1996, and in Effects of U.S. Trade FDI Activity,” International Tax and Journal of International Law and Protection and Promotion Policies, R.C. Public Finance, 11(5), September 2004, Politics, Summer 2000, pp.939–96. Feenstra, ed.,Chicago: University of Chicago pp. 601-22, and B.A. Blonigen and M. 6 B.A. Blonigen and R.B. Davies, “The Press, 1997, pp. 55–80, and B.A. Blonigen Wang, “Inappropriate Pooling of Wealthy Effects of Bilateral Tax Treaties on U.S. and D.N. Figlio, “Voting for Protection: Does and Poor Countries in Empirical FDI FDI Activity,” International Tax and Direct Foreign Investment Influence Legislator Studies,” NBER Working Paper No. Public Finance, 11(5), September 2004, Behavior?” American Economic Review, 10378, March 2004, and Does Foreign pp. 601–22, and B.A. Blonigen and R.B. 88(4), September 1998, pp. 1002–14. Direct Investment Promote Develop- Davies, “Do Bilateral Tax Treaties Promote 11 B.A. Blonigen, C.J. Ellis, and D. Fausten, ment? T.H. Moran, E.M. Graham, and M. Foreign Direct Investment?” NBER “Industrial Groupings and Foreign Direct Blomstrom, eds., Institute for International Working Paper No. 8834, March 2002, and Investment,” Journal of International Economics, April 2005, pp. 221–44. Handbook of International Trade, Economics, Vol. 65(1), January 2005, pp. 18 B.A. Blonigen, R.B. Davies, G.R. Volume II: Economic and Legal 75–91. (An earlier version was circulated as Waddell, and H.T. Naughton. “FDI in Analysis of Laws and Institutions, J. “Industrial Groupings and Strategic FDI: Space: Spatial Autoregressive Relationships in Hartigan, ed., Blackwell Publishers, 2005. Theory and Evidence” NBER Working Foreign Direct Investment,” NBER Working 7 B.A. Blonigen, “Tariff-jumping Antidump- Paper No. 8046, December 2000). Paper No. 10939, December 2004, and ing Duties,” NBER Working Paper No. 12 B.A. Blonigen and R.B. Wooster, “CEO “Spacey Parents: Spatial Autoregressive 7776, July 2000, and Journal of Turnover and Foreign Market Participation,” Patterns in Inbound FDI,” NBER Working International Economics, 57(1), June NBER Working Paper No. 9527, March Paper No. 11466, July 2005. 2002, pp. 31–50. 2003. 8 B.A. Blonigen, “Evolving Discretionary 13 J.E. Anderson and E. van Wincoop,

14 NBER Reporter Winter 2005/6 ------Race and Twentieth-Century American Economic History

William J. Collins*

Two recent anniversaries have put a ic work on the evolution of racial dispar- quality of white schools at the expense spotlight on the economic history of ities in educational outcomes, the politi- of black ones, and as the high school race-related public policy in the United cal economy and impact of changes in movement took off in the North (where States — the fiftieth anniversary of the race-specific employment and housing few blacks lived).2 In fact, there appears Supreme Court’s Brown v. Board of policy, and racial disparities in housing to have been some racial divergence in Education decision and the fortieth market outcomes, including assessments years of schooling for some birth anniversary of the Watts Riot in Los of how severe riots affected the cities in cohorts in this period. But the incentives Angeles. The Brown decision was a land- which they occurred. for investing in children’s schooling were mark in the mid-century reorientation of strong (despite labor market discrimina- race-related policy, as the machinery of Race and Schooling tion), and the overall post-1870 story is government slowly responded to the dominated by a theme of black-white In 1860, approximately 90 percent imperatives of the Civil Rights convergence. The literacy rate gap of African-Americans were slaves, and Movement. The Watts Riot, in contrast, among those born from 1870 to 1909 few slaves (perhaps 10 percent) learned marked the onset of a wave of civil dis- was about 20 percentage points, and for to read and write at even a minimal level turbances that broke out in predomi- the 1910-14 birth cohort to the 1950-4 of competence. Throughout the South, nantly black neighborhoods across the birth cohort, the racial difference in aver- it was illegal to teach slaves to read and country. Although they were fundamen- age grades of schooling fell from three write; consequently, African-Americans tally different manifestations of African- to less than one. The work highlights the entered into the post-Emancipation peri- American discontent, the Brown case and importance of intergenerational factors od with very little exposure to formal the 1960s riots have two things in comin the transmission of human capital, schooling. The literacy rate gap between mon: first, they often serve as points of and it provides an important backdrop blacks and whites born between 1800 departure in discussions of race and to current debates about racial differ- and 1860 (and still alive in 1870) was labor market, housing market, and edu- ences in test scores and educational approximately 70 percentage points. cational disparities; second, whatever attainment.3 Robert A. Margo and I use micro-level their political or symbolic significance, A subsequent paper, co-authored census data, a simple model of parents’ scholars have yet to reckon fully their with Orley Ashenfelter and Albert Yoon, incentives to invest in their children’s economic significance. attempts to set the Supreme Court’s schooling, and historical sources to Much of my work has attempted to Brown v. Board decision in the context of describe the long-run process of racial measure the effects of such events and the history described above.4 Resources convergence in schooling attainment to describe the underlying political econ- for black and white schools in the South from 1870 onward.1 Despite imperfec- omy and historical forces that con- began to equalize about two decades tions and limitations of the data, it is tributed to their occurrence. This is before the 1954 Brown decision, but it clear that the key mechanism driving the undertaken with the overarching goal of was not until the late 1960s that southern convergence was “cohort replacement” building a more comprehensive and schools truly desegregated. The paper — new generations of African- quantitative story of the economics of addresses two main questions. First, after Americans entered the labor force with race in twentieth century America. In estimating the labor market returns to more and better schooling (relative to this research summary, I describe specif- school quality for southern-born black whites) than older generations that exit- men, we ask how much more would they ed the labor force. There was nothing have earned (in 1970) if they had attend- automatic about this process, especially * Collins is a Research Associate of the ed schools with the same measurable NBER's Development of the American in the 1890-1930 period when the disen- characteristics as white schools in their Economy Program and an Associate Professor franchisement of southern blacks birth state? For the 1920s birth cohort, of Economics at Vanderbilt University. His enabled administrators to ratchet up the profile appears later in this issue. we estimate a 6 to 9 percent earnings

------NBER Reporter Winter 2005/6 15 loss. Second, given that school desegre- level. The unevenness of the diffusion of Although the state fair housing laws gation in the South occurred fairly sud- anti-discrimination policy provides were usually somewhat stronger in cov- denly, is there evidence that pre and opportunities to study both the political erage and enforcement provisions than post-desegregation cohorts of southern- economy and the effects of such policies the federal Fair Housing Act of 1968, born black students fared differently in before federal coverage applied a compar- they were still considered by many con- labor markets or in terms of educational atively uniform standard to all places at temporary observers to be too weak and attainment (in 1990)? With numerous the same time. too blunt to make a big difference. The caveats attached to the interpretation, To explore the political economy results are consistent with that sugges- the answer is “yes”. Relative to same- that facilitated (or hindered) the spread tion. aged non-southern-born blacks, the of fair employment and fair housing post-desegregation southern-born co- laws, I combined historical sources and Home Ownership, Housing horts earned more than the pre-desegre- hazard models.6 In short, I learn that Values, and Riots gation southern-born cohorts, by about throughout the period under study, A series of co-authored papers 10 percent. African-Americans were a relatively explores the economic history of race, small and poor segment of the non- residential segregation, home ownership, The Political Economy and southern population. Nonetheless, the and housing values.9 The racial gap in the Effects of Early Anti- legislation gradually moved forward home ownership rate (by household because the efforts of black political Discrimination Laws heads) was nearly the same in 2000 as it groups (such as the NAACP) were was in 1900, approximately 25 percent- World War II catalyzed the Civil strongly reinforced by labor unions (par- age points. Around mid-century, the gap Rights Movement. Because of wartime ticularly the CIO) and Jewish groups. widened as whites rapidly increased their production exigencies, African-American The econometric estimates and historical rate of home ownership and as blacks leaders had more political leverage than accounts of state-level legislative cam- moved to central cities (where ownership ever before. As labor markets tightened, paigns complement one another in this rates were low), but between 1960 and African-American workers were in rela- interpretation. The hazard model coeffi- 1980 the gap narrowed. Even so, the tively high demand; still, they were initial- cients, which can be used to project the ownership gap remains large, and in ly excluded from high-paying defense- likelihood of the adoption of anti-dis- 2000 approximately half of the gap industry jobs. A. Philip Randolph, a crimination laws, also confirm the could not be accounted for by racial dif- prominent black labor leader, demanded notion that federal intervention was crit- ferences in income, education, location, (by threatening a march on Washington) ical in the South. or household composition. This is that President Roosevelt issue an But did the sub-federal anti-discrim- approximately the same size as the executive order to outlaw discrimination ination policies make any real difference “unexplained” portion of the gap in in defense work. The President’s for black workers and households? The 1940. In a separate paper, Margo and I Fair Employment Practice Committee results from detailed analyses of individ- find that there was considerable black- attempted to enforce the first widely ual-level census data are mixed. I find white convergence in the ratio of mean applicable anti-discrimination policy, that black workers, especially women, values of owner-occupied housing thereby opening some new employment residing in states that adopted fair between 1940 and 1970 (from about 0.36 opportunities for black workers and pro- employment laws in the 1940s had larger to 0.60) but, again, there has been little viding a model regulatory agency for improvements in their labor market out- 5 change since. It is notable that the vast future anti-discrimination initiatives. comes during the 1940s than black work- majority of black-white convergence in The policy and the Committee expired ers in similar states that did not adopt ownership and housing values occurred when the war ended, but the drive to leg- fair employment laws. But I find no such before the federal Fair Housing Act and islate similar policies and committees at evidence for black workers in states that related anti-discrimination policies and the state and federal level continued. adopted fair employment laws in the before large numbers of black families At the federal level, the Civil Rights 1950s.7 In a separate paper on fair hous- moved to the suburbs. Act of 1964, the Voting Rights Act of ing laws and housing markets in the The unprecedented wave of riots 1965, and the Fair Housing Act of 1968 1960s, I find little evidence of a signifi- that rolled through black neighborhoods were the culmination of the political cant positive effect on the quality of in the mid- to late-1960s looms large in struggle to advance such legislation. But housing enjoyed by black households or the literature on race, housing, and cities, long before these milestones, race-spe- on the level of residential segregation.8 cific policy changed rapidly at the state but few studies attempt to measure the

16 NBER Reporter Winter 2005/6 ------riots’ economic impact. In two papers, the policy effects thus far. Second, in a Paper No. 128, June 2000, and Margo and I set out to measure the co-authored paper with Martha Bailey I Explorations in Economic History 40 effect of the riots on the labor market demonstrated the importance of the (2003): pp. 24–51; and W.J. Collins, “The outcomes and owner-occupied property rapid decline in household service Political Economy of State Fair-Housing Laws Prior to 1968,” NBER Working values of African-Americans.10 Most of employment, especially for black women Paper No. 10610, July 2004, forthcoming in our analysis focuses on cross-city regres- and especially after 1940, a decline that Social Science History. sions of changes in labor market out- coincided with a dramatic reorganization 7 W.J. Collins, “The Labor Market Impact of comes or property values on measures of of intra-household production and a rise State-Level Anti-Discrimination Laws, riot severity, pre-riot trends, and several in married women’s labor force partici- 1940-1960,” NBER Working Paper 8310, city-level characteristics. Instrumental pation.11 Currently, Bailey and I are col- May 2001, and Industrial and Labor variable estimates that exploit exogenous lecting data on electrification, household Relations Review 56 (2003): pp. 244–72. 8 variation in the weather around the time appliances, domestic servants, and W.J. Collins, “The Housing Market Impact of State-Level Anti-Discrimination Laws, of Martin Luther King’s murder and in women’s fertility and labor market out- 1960-1970,” NBER Working Paper No. city government structure provide an comes to shed light on the early-to-mid 9562, March 2003, and Journal of Urban alterative perspective on the riots’ effects. twentieth century connections between Economics 55 (2004): pp.534–64. When possible, we also examine patterns women’s work in the home and work in 9 W.J. Collins and R.A. Margo, “Race and of change at the census tract level and the market. Home Ownership: A Century-Long View,” using individual-level data. Nearly all of NBER Working Paper No. 7277, August the evidence suggests that the riots had 1999, and Explorations in Economic 1 W.J. Collins and R.A. Margo, “Historical negative and long-lasting effects (until at History 38 (2001): pp. 68–92; “Residential Perspectives on Racial Differences in Schooling Segregation and Socioeconomic Outcomes: least 1980) on the median value of black- in the United States,” NBER Working Paper When Did Ghettos Go Bad?” Economics owned residential property and smaller, No. 9770, June 2003, forthcoming in the Letters 69 (2000): pp. 239–43; and “Race but nontrivial, effects on the median Handbook of the Economics of and the Value of Owner-Occupied Housing, value of all residential property. For the Education, E. Hanushek and F. Welch eds. 1940–1990,” NBER Working Paper No. 1960s, the base results suggest approxi- New York: North-Holland. 7749, June 2000, and Regional Science 2 mately a 15 percent decline in the value C. Goldin, “America’s Graduation from and Urban Economics 33 (2003): pp. High School,” Journal of Economic of black-owned property in cities that 255–86. History 58 (1998): pp. 345–74. 10 W.J. Collins and R.A. Margo, “The Labor had severe riots compared to those that 3 See, for example, D. Neal, “Why Has Market Effects of the 1960s Riots,” NBER did not. Our estimates of the effects on Black-White Skill Convergence Stopped?” Working Paper No. 10243, January 2004, labor market outcomes are more mixed, NBER Working Paper No. 11090, January and in Brookings-Wharton Papers on but on the whole they suggest a signifi- 2005. Urban Affairs 2004, W. Gale and J. Pack cant negative riot effect on black income 4 O. Ashenfelter, W.J. Collins, and A. Yoon, eds. Washington, DC: Brookings Institution, and employment. For example, the base “Evaluating the Role of Brown v. Board of 2004, pp. 1–34; and “The Economic results suggest approximately a 10 per- Education in School Equalization, Aftermath of the 1960s Riots in American Desegregation, and the Income of African cent decline in median black family Cities: Evidence from Property Values,” Americans,” NBER Working Paper No. NBER Working Paper No. 10493, May income in cities that had severe riots 11394, June 2005, forthcoming in American 2004. compared to others. Law and Economics Review. 11 M.J. Bailey and W.J. Collins, “The 5 W.J. Collins, “Race, Roosevelt, and Wartime Wage Gains of African-American Women in New Work Production: Fair Employment in World War the 1940s,” NBER Working Paper No. II Labor Markets,” American Economic 10621, July 2004. For related work, see Two new projects will follow close Review 91 (2001): pp. 272–86; “African- J. Greenwood, A. Seshadri, and G. on the heels of those described above, American Economic Mobility in the 1940s: Vandenbroucke, “The Baby Boom and Baby though with less focus on race-specific A Portrait from the Palmer Survey,” Journal Bust,” American Economic Review (95) issues. First, I hope to study the long-run of Economic History 60 (2000): 1, March 2005, pp. 183–207; W.J. Collins economic impact of early urban renewal pp.756–81; and M.J. Bailey and W.J. and R.A. Margo, “Historical Perspectives on and slum clearance projects (particularly Collins, “The Wage Gains of African- Racial Differences in Schooling in the United American Women in the 1940s,” NBER in the 1950s and 1960s), which is cur- States,” NBER Working Paper No. 9770, Working Paper No. 10621, July 2004. June 2003, forthcoming in the Handbook of rently unknown. Like much of the work 6 W.J. Collins, “The Political Economy of the Economics of Education, E. described above, anecdotal impressions State-Level Fair-Employment Laws, Hanushek and F. Welch eds. New York: have outstripped systematic analyses of 1940–1964,” NBER Historical Working North-Holland.

------NBER Reporter Winter 2005/6 17 Globalization and New Comparative Economic History

Alan M. Taylor*

Globalization is probably one of the long run in the economic environment today. most overused words in economics, as it (institutions, regimes, policies, and so on) Much of my own research, including is in many other realms of academic and and economic outcomes (growth, infla- a large project in collaboration with my public debate. Nonetheless, it cannot be tion, trade, capital movements, and so fellow NBER Research Associate avoided, if only because an understand- on). Maurice Obstfeld, has been concerned ing of the modern world requires us to In that vein, I have been working to with this question of measuring market confront it. Economically, its potential address several important questions that integration over time, with a focus on benefits seem all too apparent: for exam- help us understand economic globaliza- global capital markets.1 There is no ple, the fast growing industrializing tion over the last 100–150 years, allowing agreed upon method for evaluating mar- economies of Asia are well connected to us to understand the economic outcomes ket integration, although we have made global markets for goods and capital. of today with a deeper perspective. In some progress recently using nonlinear Conversely, no economically isolated this research summary I highlight two theoretical and empirical models to better country has prospered. As UN Secretary strands of this work: the evolution of estimate transactions costs in markets General Kofi Annan has pointed out: global capital markets and the evolution using high-frequency price data. For “The main losers in today’s very unequal of world trade. These topics address such most applications both price and quanti- world are not those that are too exposed issues as: how can we measure the extent ty criteria remain relevant. Each have to globalization, but those who have been of globalization? What explains the rise their weaknesses — quantities may flow, left out.” and fall of globalization in different eras and prices may converge, between loca- My recent research has focused on and in different countries? What are the tions despite large obstacles — and aux- the causes and consequences of global- costs and benefits of globalization? iliary assumptions and information must ization, and is based on an interdiscipli- be carefully considered using either crite- nary approach that straddles internation- The Ebb and Flow of Global rion. Yet what we find, broadly, is that al economics, economic growth, and eco- Capital global capital markets were just as nomic history. Methodologically, an his- impressive in their degree of integration The forces of economic globaliza- torical approach has appeal because the a century ago as they are today. Some tion appear particularly strong at present, global “economic laboratory” provides very simple quantity criteria can sum up but economic historians have been at data not only across space (for cross- the story. pains to point out that we are now living country comparisons) but also across For example, we can look simply at in the second era of globalization, not time (from previous centuries to the the ratio of the stock of foreign invest- the first. The first stretched from rough- present era). Historical data contain more ment in the world to global GDP. Plotted ly 1870 until the start of World War I in variation than contemporary data alone, over time, this series has a distinctive 1914 and saw unprecedented integration providing a wealth of information to be shape. It rose dramatically from 1870 to in international market for goods, capital, exploited. An emerging sub-field of New 1914, from 7 percent to 18 percent. From and labor. Since a key issue for the intel- Comparative Economic History is devot- 1914 to 1950 it fell precipitously to just 5 lectual enterprise of New Comparative ed to exploring relationships in the very percent. It rose slowly but stayed fairly Economic History is whether the past low through the 1980s, and it then surged can provide useful lessons for the pres- quickly in the last two decades of the * Taylor is a Research Associate in the ent, we have first to answer the question twentieth century from 25 percent to 92 NBER's Programs on the Development of the of whether this past era in any way percent. The data suggest that we have American Economy, International Trade and resembles the present. A first challenge is Investment, and International Finance and indeed lived through two eras of global- to assess quantitatively when and where Macroeconomics, and a professor of economics ization, and using this yardstick, the inter- the extent of market integration in the at the University of California, Davis. His national movement of capital one hun- past bore any resemblance to that seen profile appears later in this issue. dred years ago was no less impressive

18 NBER Reporter Winter 2005/6 ------than that witnessed today. The data also period and the Bretton Woods era (the policy was subordinated to the goals of reveal two major reversals in the twenti- latter being the period studied by capital mobility and a fixed exchange rate. eth century: a steep decline in capital Feldstein and Horioka). The familiar In the interwar period, perhaps because movement in the interwar years and a story of two globalizations — and the of increasing democratic pressure, gov- steep rise in the 1980s, creating a distinc- same “U-shape” — emerges again. ernments felt the need to use tive “U-shape” pattern when these data The “U-shape” pattern, which recurs autonomous monetary policy; what gave are plotted. in many other tests based on a variety of was the peg (the collapse of the gold More formal tests are possible. For data and empirical methodologies, also standard) or, in some cases, capital mobil- example, turning from stock data to flow conforms to the broad contours of the ity. But the economic chaos and instabili- data, we can look at the correlation of history of macroeconomic policymaking ty of the interwar period was intolerable saving and investment rates across coun- in the world’s major economies that are to those planning the contours of the tries and across time. As the seminal the main focus of the study (quite differ- postwar global economy at Bretton work of Martin S. Feldstein and Charles ent patterns apply to developing coun- Woods, and fixed exchange rates were Y. Horioka points out, a small open tries). In the two eras of globalization, still viewed as a sine qua non for a stable economy need not see any correlation of capital controls were notably absent and world economy. The new arrangements domestic investment and saving in the were typically frowned upon; in between, would sacrifice capital mobility to keep short run, so any correlation between the at the bottom of the “U,” capital controls currencies on “adjustable” pegs to the two may be considered prima facie evi- became prevalent and came to be judged dollar and yet preserve monetary policy dence of capital market frictions.2 We can as the norm. How and why the history of autonomy. Still, this system could not gain some preliminary insight if we apply policymaking followed these twists and endure: capital movements (often dis- this methodology across time and space, turns then becomes an important ques- guised) grew in the 1960s, the adjustabili- using annual data from 1870 to the pres- tion. ty of pegs invited speculative attacks, and ent. We suppose that investment is driven A central concept in international importing rising inflation from the U.S. by where the best profit opportunities macroeconomics presents itself as a can- anchor currency imposed costs on the are, at home or abroad; saving is driven didate explanation: the “trilemma.” The other players. From 1971 onwards, the by consumption choices, which the trilemma posits that economic policy major economies have floated, adapting household can in principle de-link from cannot simultaneously achieve three to (even encouraging) capital mobility, firm investment choices; and the differ- goals — a fixed (or even managed) and resolving the trilemma in the only ence between saving and investment is exchange rate, which may be desired for other way that preserves policy autonomy. the current account. We need only to add stability purposes; international capital The trilemma sounds like a nice the caveat that, in the long run, the two mobility, which may be desired for access story, but what is its explanatory power must be correlated: the long-run budget to foreign capital; and autonomous mon- and historical relevance? This hitherto constraint of the economy dictates that etary policy, which may be desired for unexplored question can be addressed be “on average” the current account be in managing the business cycle or providing examining the degree of correlation balance, allowing for initial wealth. a lender of last resort. The logic is that between “local” and “world” interest Econometric results show that the under a fixed exchange rate and capital rates, controlling for the type of correlation between saving and invest- mobility, simple interest parity means that exchange rate regime and capital control ment almost never goes to zero — the local interest rate must equal the regime in operation. Tested in this way, indeed, the long-run budget constraint “world” interest rate, and monetary poli- the trilemma finds strong support in all tends to keep the measure between a cy is rendered ineffective (or impossible). historical eras from the Gold Standard to minimum of 0.5 and a maximum of 1 for Something has to give if monetary policy the present and under a wide variety of reasonable model simulations under a is to be enabled: either the exchange rate macroeconomic regimes. These findings wide range of parameters. Yet this range must float or capital controls must be provide an evidentiary base for our still provides a useful yardstick. Sure applied to suspend parity and admit inter- accounts of global macroeconomic his- enough, we find correlations in the data est differentials. tory; they also give much-needed empiri- close to the low of 0.5, implying low fric- Prevailing narratives that tell the his- cal weight to the idea of the trilemma, tions, in both eras of globalization: a cen- tory of the world in four parts (that is, the one of the most fundamental constraints tury ago and today. We find high correla- macroeconomic history since 1870) build that economic policymakers have all too tions close to the high of 1, implying high on the trilemma’s logic.3 In the classical often ignored, to their peril. friction, in between: during the interwar gold standard (1870–1914) monetary

------NBER Reporter Winter 2005/6 19 The Rise and Fall of World ural resources — arguably the factors that rose (relative to wholesale prices) helping Trade were uppermost in Heckscher and to turn the boom into a bust. Interwar tar- Ohlin’s minds.6 iff policy, especially in the 1930s, was also The historical patterns of globaliza- Turning to the gravity model, howev- an important culprit in the collapse of tion seen in capital markets also carry over er, results in something more like an world trade. to goods markets and the history of inter- empirical success with historical data, as I Finally, war matters. In very recent national trade. Circa 1870, the ratio of have found in research with various col- work, we have found that wars have a world trade to GDP stood at 10 percent, laborators. Again, this matches the empir- profound — and very persistent —effect rising to 21 percent by 1914, falling to 9 ical success of the gravity model using on trade between countries.7 In wartime percent by 1938, and then rising to 27 postwar data. Yet if our goal is to under- perhaps 90 percent or more of trade percent by 1992: a first phase of global- stand why trade rose and fell so markedly between countries simply disappears; but ization followed by that twentieth-century over time, an unadorned gravity model is even after the war ends, we find that it “U-shape” again. What can explain this not much help, since relative country sizes takes about ten years for trade to return to rise and fall (and rise) of world trade? and distances change little over time. normal “peacetime” levels. This also This has been another major goal in my Instead we need to include other meas- helps to explain the precipitous drop in research.4 ures of policies, institutions, and the interwar trade and the slow post-1945 To understand trade patterns in the changing economic environment, and recovery: globally about 10 to 20 percent long run requires that we adopt a theoret- some obvious candidates stand out here of world trade was probably destroyed by ical model and estimate its parameters for the late nineteenth and early-to-mid the “war effect” alone. We also find large using long-run data. Two models stand twentieth century: the rise and fall of the “negative externalities” from war, in the out as leading contenders for this job. gold standard, a monetary arrangement sense that even neutral countries suffer a First, the Heckscher-Ohlin model, in which was believed to be a stimulus of drop in trade when their trading partners which trade happens as a result of differ- world trade; the transportation revolution, enter a conflict. A speculative and rough ences in countries’ factor endowments; which dramatically lowered long-distance estimate of the costs of such “lost trade” second, the so-called gravity model, in shipping costs before 1914 through tech- finds that they might be significant in wel- which countries export differentiated nological change in shipping and the con- fare terms, of the same order of magni- products in proportion to their own struction of major canals; changes in tar- tude as the costs of lost human capital country size and subject to distance-relat- iff policy, particularly after 1914 when (measured by lost wages attributable to ed transport costs.5 trade policy activism became common; deaths or injuries). We have therefore Getting the Heckscher-Ohlin model and the impact of wars, particularly the been able to document a quantitatively to match real world data has generated two World Wars which affected a large important cost of war that is subject to endless problems with postwar data: rela- fraction of the world economy. large spillovers, and that has been little tive abundance (or scarcity) of a particu- The results of these studies give little understood until now. lar factor is a poor predictor of whether hope that a monocausal explanation will said factor will tend to be exported suffice to explain the history of world Future Research (respectively, imported) by any given trade. We find that the gold standard Historical research on the past evolu- country. Instead, the so-called factor con- made a difference, and when two coun- tion of the global economy sheds new tent of trade goes the “wrong” way. And, tries both go onto the gold standard their light on the causes and consequences of even more perplexing, the volume of bilateral trade rises by 42 percent, which economic integration and the problems trade is far too small to be consistent with helps to account for much of the rise in and challenges it may cause for people, the model — the so-called paradox of trade before 1914, and much of its disap- firms, and policymakers today and tomor- “missing trade,” which can only be solved pearance by 1939. There are direct paral- row. In recent years we have arrived at (or assumed away) theoretically with lels here, of course, with the contempo- new insights using the systematic, quanti- strong anti-trade axioms of home bias. It rary debate over the impact of common tative, cross-country and cross-time thus might be expected that with data currencies on trade, especially the long- approach of the New Comparative from distant history, from the period run impacts of the euro. We also find that Economic History, but there remain 1870-1939, we might also encounter the decline of transport costs likewise many unanswered questions. problems with the theory. This is broadly made a big difference in the 1870–1914 Understanding the frictions in the true, although we can find some weak period, explaining a large fraction of the global economy is a central task for support for the model as it applies to nat- trade boom; but after 1914, trade costs students of international trade and

20 NBER Reporter Winter 2005/6 ------finance in the past and present.8 pp.75–108; M. Obstfeld and A.M. Taylor, Feenstra, Advanced International Trade: Methodologically, we shall continue to “Sovereign Risk, Credibility, and the Gold Theory and Evidence (Princeton, N.J.: develop better techniques to assess how Standard: 1870–1913 versus 1925–31,” Princeton University Press, 2004). 6 globalization has evolved, and how well NBER Working Paper No. 9345, November The application of the Heckscher-Ohlin model 2002, and Economic Journal 113 (April to the late nineteenth century works rather better integrated markets are at any given time.9 2003), pp.1–35; “Globalization and Capital when its price predictions, rather than quantity We can then better understand how close Markets,” NBER Working Paper No. 8846, predictions, are put to the test. See K.H. we are to a hypothetical single market in March 2002, in Globalization in Historical O’Rourke, A.M. Taylor, and J.G. Williamson, goods and capital or how severe is “mar- Perspective, M. D. Bordo, A. M. Taylor, and “Factor Price Convergence in the Late Nineteenth ket failure.” These assessments also need J. G. Williamson, eds. (Chicago: University of Century,” NBER Historical Working Paper to take into account the wide ranges of Chicago Press, 2003); A.M.Taylor, “A Century No. 46, November 1996, and International policies and institutions that have operat- of Current Account Dynamics,” NBER Economic Review 37(1996), pp. 499–530; ed across time and space and which have Working Paper No. 8927, May 2002, and and K.H. O’Rourke and J.G. Williamson, Journal of International Money and Globalization and History: The Evolution encouraged or inhibited international Finance 21 (November 2002), pp. 725–48; of a Nineteenth-Century Atlantic Econo- trade and finance. A.M. Taylor, “A Century of Purchasing Power my (Cambridge: MIT Press, 1999). The latest research casts doubt on Parity,” NBER Working Paper No. 8012, 7 R.Glick and A.M.Taylor, “Collateral simple generalizations that globalization is November 2000, and Review of Economics Damage: Trade Disruption and the Economic always beneficial or always harmful; and Statistics 84 (February 2002), Impact of War,” NBER Working Paper No. rather, its benefits appear to be greater in pp.139–50; M. Obstfeld and A.M. Taylor, 11565, August 2005. 8 countries that climb up the ladder of “The Great Depression as a Watershed: See, for example, M. Obstfeld and K. Rogoff, institutional quality. In time we will devel- International Capital Mobility in the Long “The Six Major Puzzles in International Run,” NBER Working Paper No. 5960, May Finance: Is There a Common Cause?” NBER op a more detailed knowledge of how 1999, in The Defining Moment: The Macroeconomics Annual (2000), pp. globalization has worked in different con- Great Depression and the American 339–90; and J.E. Anderson and E. van texts. We will then be better placed to Economy in the Twentieth Century, M. D. Wincoop, “Trade Costs,” Journal of know whether the promises of prosperity Bordo, C. D. Goldin, and E. N. White, eds. Economic Literature, 42 (2004), pp. 691- held out by the process of globalization (Chicago: University of Chicago Press, 1998) 751. will be shared by only a few, or — as Kofi 2 See M.S. Feldstein and C.Y .Horioka, 9 In this area, one of the most promising avenues Annan and many others hope — by “Domestic Saving and International Capital appears to be the use of nonlinear models of price Flows,” Economic Journal 90 (1980), pp. adjustment to infer transaction costs. See E. many. 314–29. Canjels, G. Prakash-Canjels, and A.M. Taylor, 3 For an influential example, see B.J. Eichengreen, “Measuring Market Integration: Foreign 1 This research was recently published in book Globalizing Capital: A History of the Exchange Arbitrage and The Gold Standard, form: M. Obstfeld and A.M. Taylor, Global International Monetary System (Princeton, 1880–1913,” NBER Working Paper No. Capital Markets: Integration, Crisis, and N.J.: Princeton University Press, 1996). 10583, June 2004, and Review of Growth, Japan-U.S. Center Sanwa 4 The relevant published papers are as follows, and Economics and Statistics 86 (November Monographs on International Financial Markets all appeared first as NBER Working Papers: 2004), pp. 868–82; M.P. Taylor and A.M. (Cambridge: Cambridge University Press, A. Estevadeordal, B. Frantz, and A.M. Taylor, Taylor, “The Purchasing Power Parity Debate,” 2004). We gratefully acknowledge the financial “The Rise and Fall of World Trade, NBER Working Paper No. 10607, July 2004, support of the Sanwa Prize in International 1870–1939,” NBER Working Paper No. and Journal of Economic Perspectives 8 Economics and Financial Markets. The related 9318, November 2002, and Quarterly (Fall 2004), pp. 135–58; A.M. Taylor, background papers, all of which appeared as Journal of Economics 118 (May 2003), pp. “Potential Pitfalls for the Purchasing-Power NBER Working Papers, were published as fol- 359–407; A.Estevadeordal and A.M.Taylor, Parity Puzzle? Sampling and Specification Biases lows: M. Obstfeld, J.C. Shambaugh, and A.M. “Testing Trade Theory in Ohlin’s Time,” in Mean-Reversion Tests of the Law of One Taylor, “The Trilemma in History: Tradeoffs NBER Working Paper No. 8842, March Price,” NBER Working Paper No. 7577, among Exchange Rates, Monetary Policies, and 2002, in Bertil Ohlin: A Centennial March 2000, and Econometrica 69 (March Capital Mobility,” NBER Working Paper No. Celebration, 1899–1999, R. Findlay, L 2001), pp. 473–98; M.Obstfeld and A.M. 10396, March 2004, and Review of Jonung, and M. Lundahl, eds. (Cambridge: Taylor, “Nonlinear Aspects of Goods-Market Economics and Statistics 87 (August MIT Press, 2002); A. Estevadeordal and Arbitrage and Adjustment: Heckscher’s 2005), pp.423–38; “Monetary Sovereignty, A.M. Taylor, “A Century of Missing Trade?” Commodity Points Revisited,” NBER Working Exchange Rates, and Capital Controls: The NBER Working Paper No. 8301, May 2001, Paper No. 6053, June 1997, and Journal of Trilemma in the Interwar Period,” NBER and American Economic Review 92 the Japanese and International Working Paper No. 10393, March 2004, and (March 2002), pp.383–93. Economies 11 (December 1997), pp. 5 IMF Staff Papers 51 (Special Issue 2004): For a thorough survey of these models see R.C. 441–79.

------NBER Reporter Winter 2005/6 21 NBER Profile: James Adams

James Adams is a Research In recent years, Adams’s research Associate in the Productivity Program of interests have centered on the connec- the National Bureau of Economic tions between industrial productivity in Research and a professor of economics the output of goods and inventions, at Rensselaer Polytechnic Institute. He industrial R and D, and pre-technology received his Ph.D. degree from the science, especially university research. University of Chicago. Before moving to His interests also extend to the boundary Rensselaer he taught at the University of between technological change and labor Florida and Iowa State University. He has and public economics. also held visiting positions at the U.S. He is married to Jennifer Cobb Bureau of Labor Statistics, the Bureau of Adams, has three cats, and lives in the the Census, and the George J. Stigler scenic hill country near Rensselaer, New Center for the Study of the Economy York. He enjoys music, especially classi- and the State at the University of cal music from the seventeenth and eigh- Chicago. teenth centuries; reading of a wide-rang- Adams recently served on a National ing character; and nature studies. When Research Council panel on telecommuni- time and the weather permit, he likes to cations R and D. He also has advised the kayak and canoe the lakes of northern Bureau of the Census and the Advanced New York and Vermont. He also finds Technology Program of the National that bicycling, hiking, and puttering Institute of Standards and Technology around the house form useful breaks on the measurement of R and D. from research and teaching.

NBER Profile: Bruce A. Blonigen

Bruce A. Blonigen is a Research Blonigen has published numerous Associate in the NBER’s Program on articles on foreign direct investment, as International Trade and Investment and well as on the economics of trade poli- the Knight Professor of Social Science cy, particularly antidumping trade pro- in the economics department at the tection. He maintains a website that University of Oregon. He received his shares information and data on U.S. Ph.D. in economics from the University antidumping activity. He also serves on of California, Davis in 1995. While com- the editorial boards of the Journal of pleting his dissertation, he also worked as International Economics, Canadian Journal of an economist for the Research Division Economics, and the North American Journal of the Office of Economics at the U.S. of Economics and Finance, and is a International Trade Commission. Departmental Editor for the Journal of Blonigen teaches courses in interna- International Business Studies. tional economics, industrial organiza- Blonigen lives in Eugene, Oregon tion, and statistical methods at both the with his wife Denice Gray and their son, graduate and undergraduate level. He Ben. In his spare time, Blonigen is an won the University of Oregon’s Ersted avid golfer, and enjoys hikes and vaca- Award for Distinguished Teaching in tions with the family. 2003.

22 NBER Reporter Winter 2005/6 ------NBER Profile: William J. Collins

William J. Collins is a Research was the Model-Okun Fellow at the Associate in the NBER’s Development Brookings Institution. Collins serves on of the American Economy Program and the editorial boards of Social Science an Associate Professor of Economics at History, Historical Methods, and the Journal Vanderbilt University. Both his A.B. of Economic History. He also manages the (1993) and his Ph.D. (1998) degrees in book reviews on North American topics economics are from Harvard University. for the Journal of Economic History. Collins joined the Vanderbilt faculty Collins grew up near Reading, in 1998. In the same year, he received the Pennsylvania, and now lives in Nashville, Economic History Association’s Allan Tennessee. In his spare time, he reads Nevins Prize for the best dissertation on fiction and poetry, catches up on old a North American topic. movies, enjoys music around Nashville, During the 2001-2 academic year, and visits friends and family in Boston Collins was a National Fellow at the and elsewhere. NBER; in the 2003–4 academic year, he

NBER Profile: Alan M. Taylor

Alan M. Taylor is a Research Markets. Taylor has served as a consult- Associate in the NBER’s Programs on ant or visitor with the World Bank, the the Development of the American International Monetary Fund, the Inter- Economy, International Finance and American Development Bank, and the Macroeconomics, and International Federal Reserve Bank of San Francisco. Trade and Investment. He is also a He currently holds a John Simon Professor of Economics and a Guggenheim Memorial Fellowship and Chancellor’s Fellow at the University of has spent most of the last year on sab- California, Davis, and the Director of batical in Paris and London. the Center for the Evolution of the When not on leave, Taylor lives in Global Economy. He received his Ph.D. Davis, California, with his wife (who is a from Harvard University. professor of literature), daughter, and For his dissertation, Taylor was cat. When he can find the time, Taylor awarded the Alexander Gerschenkron enjoys listening to jazz and traveling to Prize by the Economic History either very urban or very remote places. Association. For their work on global He occasionally descends snow slopes capital markets, he and Maurice Obstfeld and stands in rivers, but any resem- were awarded the 1997 Sanwa Prize in blances to skiing and fly-fishing are pure- International Economics and Financial ly coincidental.

------NBER Reporter Winter 2005/6 23 Conferences

Risks of Financial Institutions

An NBER Conference on the Markus Brunnermeier, Princeton University of Pennsylvania and NBER, Risks of Financial Institutions was University, and Lasse Heje Pedersen, “Roughing It Up: Including Jump held in Cambridge on November 10. New York University, “Market Components in the Measurement, Mark Carey, Federal Reserve Board, Liquidity and Funding Liquidity” Modeling, and Forecasting of Return and Rene M. Stulz, NBER and Ohio Discussant: Jeremy C. Stein, Harvard Volatility”(NBER Working Paper No. State University, organized the meet- University and NBER 11775) ing, at which the following papers Discussant: Eric Ghysels, University were discussed: Gregory W. Brown, University of of North Carolina, Chapel Hill North Carolina, Chapel Hill; Söhnke Deborah Lucas, Northwestern M. Bartram, Lancaster University; and Samuel Hanson, Harvard University; University and NBER, and Robert John E. Hund, University of Texas at M. Hashem Pesaran, University of McDonald, Northwestern University, Austin, “Estimating Systemic Risk in Cambridge; and Til Schuermann, “An Options-Based Approach to the International Financial System” Federal Reserve Bank of New York, Evaluating the Risk of Fannie Mae and Discussant: Anthony Saunders, New “Firm Heterogeneity and Credit Risk Freddie Mac” York University Diversification” Discussant: David M. Lando, Copen- Diana Hancock and Wayne Viral Acharya and Timothy Johnson, hagen Business School Passmore, Federal Reserve Board, London Business School, “Insider “Understanding Market Discipline in Trading in Credit Derivatives” Sanjiv Das, Santa Clara University; the Presence of Implicit Government Discussant: Louis Scott, Morgan Darrell Duffie, Stanford University Guarantees: An Analysis of Subordi- Stanley and NBER; Nikunj Kapadia, nated Bond and Stock Returns for University of Massachusetts; and GSEs and for Bank Holding Com- Torben G. Andersen, Northwestern Leandro Saita, Stanford University, panies” University and NBER; Tim “Common Failings: How Corporate Discussant for both papers: Thomas Bollerslev, Duke University and Defaults Are Correlated” Wilson, ING NBER; and Francis X. Diebold, Discussant: David Li, Barclays Capital

Fannie Mae and Freddie Mac assume take an options pricing approach to eral credit guarantees. a significant amount of interest and pre- evaluating these costs and risks. Under When studying changes in the risks payment risk and all of the credit risk for the base case assumptions, the estimat- of large bank holding companies about half of the eight trillion dollar U.S. ed value of the guarantees is $7.9 billion (BHCs) and government-sponsored residential mortgage market. Their over ten years, with a combined 0.5 per- enterprises (GSEs), researchers routine- hybrid government-private status, and cent value at risk of $122 billion. The ly argue that changes in the responsive- the perception that they are too big to authors evaluate the sensitivity of these ness of stock and subordinated bond fail, make them a potentially large, but estimates to various modeling assump- returns to exogenous risk factors can be mainly unaccounted for, risk to the fed- tions, and also to the regulatory regime, interpreted as reflecting changes in eral government. Measuring the size including forbearance policies and capi- investors’ views about the firm’s expect- and risk of this liability is technically tal requirements. Their analysis high- ed losses. However, investors may per- difficult, but important for the debate lights the benefits, but also the chal- ceive that these large firms have sub- over the appropriate regulation of these lenges, of taking on options-based stantial implicit government guarantees. institutions. Lucas and McDonald approach to evaluating the value of fed- Hancock and Passmore show that

24 NBER Reporter Winter 2005/6 ------these guarantees can confound the to macroeconomic risk factors) when the likelihood of systemic failure. These interpretation of stock and bond return compared with BHCs. While their tech- estimates suggest that systemic risk is responsiveness, making changes in the nique does not identify the source of limited, even during major financial responsiveness of bond returns difficult this potentially greater risk, the authors crises. For example, maximum likeli- to interpret. They also show that note that financial theory would suggest hood estimation of bank failure proba- changes in the responsiveness of stock that GSEs might have greater risks bilities implied by equity prices suggests returns are almost impossible to inter- because they are less diversified and not the Asian crisis induced less than a sin- pret. These results suggest that implicit as well capitalized as BHCs. gle percent increase in the probability of guarantees can hide investors’ percep- Brunnermeier and Pedersen pro- systemic failure. Third, the authors esti- tions of changes in expected loss attrib- vide a model that links a security’s mar- mate systemic risk as implied by equity utable to important risk factors, thereby ket liquidity — that is, the ease of trad- option prices of U.S. and European confounding market and regulatory ing it — to traders’ funding liquidity — banks. The largest values are for the efforts to correctly price and manage that is, their availability of funds. Russian crisis and September 11; these risks. The authors provide conditions Traders provide market liquidity and show increases in estimated average under which bond returns can be use- their ability to do so depends on their default probabilities of only around 1-2 fully interpreted as reflecting expected funding — that is, their capital and the percent. Taken together, the results sug- losses and thus the relative riskiness of margins charged by their financiers. In gest statistically significant, but eco- firms. They consider the risk-sensitivity times of crisis, reductions in market liq- nomically small, increases in systemic of subordinated bond returns of highly uidity and funding liquidity are mutually risk around even the worst financial rated BHCs and of GSEs to macroeco- reinforcing, leading to a liquidity spiral. crises of the last ten years. Although nomic shocks during two periods: April The model here explains the empirical- policy responses are endogenous, the 1, 2001 to May 31, 2003 and June 1, ly documented features that market liq- low estimated probabilities suggest that 2003 to September 15, 2004. Although uidity: 1) can suddenly dry up (in other the distress of central bankers, regula- the GSEs (Fannie Mae and Freddie words, is fragile); 2) has commonality tors and politicians about the events Mac) and the largest U.S. bank holding across securities; 3) is related to volatili- they study may be overstated, and that companies may benefit substantially ty; 4) experiences “flight to liquidity” current policy responses to financial from a perceived implicit government events; and 5) co-moves with the mar- crises and the existing institutional guarantee of their liabilities, the political ket. Finally, the model shows how the framework may be adequate to handle support for government backing of the Fed can improve current market liquid- major macroeconomic events. GSEs seemed less certain to investors ity by committing to improving funding Insider trading in the credit deriva- in the later period, while there was no in a potential future crisis. tives market has become a significant news, or legislative developments, that With a unique and comprehensive concern for regulators and participants. likely would have changed the perceived dataset, Bartram, Brown, and Hund Acharya and Johnson attempt to implicit government guarantees for develop and use three distinct methods quantify the problem. Using news BHCs. The authors show that the to quantify the risk of a systemic failure reflected in the stock market as a bench- responsiveness of subordinated bond in the global banking system. They mark for public information, they returns to macroeconomic shocks dur- examine a sample of 334 banks (repre- report evidence of significant incre- ing the two sample periods indicate senting 80 percent of global bank equi- mental information revelation in the that: 1) BHCs’ bond returns across the ty) in 28 countries around six global credit default swap (CDS) market under two periods became less sensitive to financial crises (such as the Asian and circumstances consistent with the use changes in macroeconomic factors that Russian crises and September 11, 2001), of non-public information by informed affect credit risks but more sensitive to and show that these crises did not cre- banks. Specifically, the information rev- changes in macroeconomic factors that ate large probabilities of global financial elation occurs only for negative credit influence interest rate risks; 2) changes system failure. First, they show that news and for entities that subsequently in implicit guarantees made it difficult cumulative negative abnormal returns experience adverse shocks. Moreover, to interpret GSE bond returns across for the subset of banks not directly the degree of advance information rev- the two periods; and 3) bond investors exposed to a negative shock (unexposed elation increases with the number of generally believed that GSEs are at least banks) rarely exceed a few percent. banks that have lending/monitoring as risky, and maybe more risky, (that is, Second, they use structural models to relations with a given firm, and this their expected losses are more sensitive obtain more precise point estimates of effect is robust to controls for non-

------NBER Reporter Winter 2005/6 25 informational trade. They find no evi- casting model, the authors find that underestimation of expected losses. dence, however, that the degree of almost all of the predictability in daily, But, once expected losses are controlled asymmetric information adversely weekly, and monthly return volatilities for, neglecting parameter heterogeneity affects prices or liquidity in either the comes from the non-jump component. can lead to overestimation of risk, equity or credit markets. If anything, Their results thus set the stage for a whether measured by unexpected loss with regard to liquidity, the reverse number of interesting future economet- or value-at-risk. These results are con- appears to be true. ric developments and important finan- firmed empirically using returns and A rapidly growing literature has cial applications by separately modeling, credit ratings for firms in the United documented important improvements forecasting, and pricing the continuous States and Japan across seven sectors. in financial return volatility measure- and jump components of the total Ignoring parameter heterogeneity ment and forecasting through the use of return variation process. results in far riskier credit portfolios. realized variation measures constructed Hanson, Pesaran, and Schuer- Das, Duffie, Kapadia, and Saita from high-frequency returns, coupled mann consider a simple model of cred- develop, and apply to data on U.S. cor- with simple modeling procedures. it risk and derive the limit distribution porations from 1979-2004, tests of the Building on recent theoretical results in of losses under different assumptions standard doubly-stochastic assumption Barndorff-Nielsen and Shephard regarding the structure of systematic under which firms’ default times are (2004a, 2005) for related bi-power vari- and idiosyncratic risks and the nature of correlated only as implied by the corre- ation measures, Andersen, Bollerslev, firm heterogeneity. Their theoretical lation of factors determining their and Diebold provide a practical and results indicate that if firm-specific risk default intensities. This assumption is robust framework for non-parametri- exposures (including their default violated in the presence of contagion or cally measuring the jump component in thresholds) are heterogeneous but “frailty” (unobservable explanatory asset return volatility. In an application come from a common parameter distri- variables that are correlated across to the DM/$ exchange rate, the bution, then there is no scope for fur- firms). The tests here do not depend on S&P500 market index, and the 30-year ther risk reduction through active cred- the time-series properties of default U.S. Treasury bond yield, they find that it portfolio management for sufficiently intensities. The data do not support the jumps are both highly prevalent and dis- large portfolios. However, if the firm joint hypothesis of well-specified tinctly less persistent than the continu- risk exposures are drawn from different default intensities and the doubly-sto- ous sample path variation process. parameter distributions, say for differ- chastic assumption. There is also some Moreover, many jumps appear directly ent sectors or countries, then further evidence of default clustering in excess associated with specific macroeconom- risk reduction is possible, even asymp- of that implied by the doubly-stochastic ic news announcements. Separating totically, by changing the portfolio model with the given intensities. jump from non-jump movements in a weights. In either case, neglecting simple but sophisticated volatility fore- parameter heterogeneity can lead to

26 NBER Reporter Winter 2005/6 ------IASE: Strengthening Global Financial Markets

The NBER and Pontificia Sebastian Edwards, “Financial Ana Carla A Costa, Banco Central do Universidade Catolica do Rio de Openness, Crises, and Output Losses” Brasil, and Joao Manoel Pinho de Janeiro (PUC-Rio) jointly sponsored a Comments: Edmar Bacha, Bank of Mello, PUC-Rio, “Judicial Risk and meeting of the Inter-American Semi- Italy, and Marcelo Muinhos, Banco Creditor Expropriation: Micro nar on Economics in Brazil on Central do Brasil Evidence from Brazilian Payroll December 2 and 3. This Seminar Loans” focused on “Strengthening Global Viviana Fernandez, Universidad de Comments: Renato Flores, EPGE/FGV, Financial Markets.” NBER Research Chile, “The International CAPM and a and Beny Parnes, PUC-Rio Associate Sebastian Edwards of Wavelet-based Decomposition of University of California, Los Angeles, Value at Risk” Eduardo Levy-Yeyati, Universidad and Marco Garcia of PUC-Rio, organ- Comments: Marcelo Medeiros, PUC- Torcuato di Tella, “Liquidity Insurance ized the following program: Rio, and Caio Ibsen, IBMEC Business in Financially Dollarized Economy” School-Rio Comments: Marco Bonomo, EPGE, Joshua Aizenman, University of and Alejandro Werner, Subsecretaria de California at Santa Cruz and NBER, Ross Levine, NBER and Brown Hacienda y Credito Publico do Mexico and Ilan Noy, University of Hawaii, University, and Sergio L. Schmukler, “Endogenous Financial and Trade The World Bank, “Internationalization Barry J. Eichengreen, University of Openness in a Volatile World” and Stock Market Liquidity” California at Berkeley and NBER; and Comments: Maria Cristina Terra, Comments: Ugo Panizza, Inter- Poonam Gupta and Ashoka Mody, Postgraduate School of Economics - American Development Bank, and IMF, “Sudden Stops and IMF Brazil (EPGE), and Thierry Verdier, Eduardo Loyo, IMF Programs” Centre for Economic Policy Research Comments: Ilan Goldfajn, PUC-Rio, Ricardo J. Caballero, NBER and and Affonso Celso Pastore, ACPastore Bernardo S. de M. Carvalho, Gávea MIT; Takeo Hoshi, NBER and & Associados Investments, and Marcio Garcia, University of California at San Diego; “Ineffective Controls on Capital and Anil K Kashyap, University of Inflows under Sophisticated Financial Chicago and NBER, “Zombie Lending Markets: Brazil in the Nineties” and Depressed Restructuring in Japan” Comments: Gustavo Franco and Comments: Vinicius Carrasco and Marcelo Abreu, PUC-Rio Walter Novaes, PUC-Rio

Aizenman and Noy study the cial openness has powerful effects on methodology. They find that almost all endogenous determination of financial future trade openness. While de-jure of the linear feedback between trade and trade openness when both are restrictions on capital mobility do not and financial openness can be account- volatile. First, they outline channels affect de-facto financial openness, de- ed for by G-causality from financial leading to two-way feedback between jure restrictions on the current account openness to trade openness (53 per- the different modes of openness; next, have large adverse effect on commercial cent) and from trade to financial open- they identify these feedbacks empirical- openness. This suggests that it is much ness (34 percent). They conclude that, ly. They find that a single standard devi- easier to overcome restrictions on capi- in an era of rapidly growing trade inte- ation increase in commercial openness tal account convertibility than restric- gration, countries cannot choose finan- is associated with a 9.5 percent (of tions on commercial trade. Having cial openness independent of their GDP) increase in de-facto financial established (Granger) causality, the degree of openness to trade —dealing openness, controlling for political econ- authors investigate the relative magni- with greater exposure to financial tur- omy and macroeconomic factors. tudes of these directions of causality bulence by curbing financial flows will Similarly, an increase in de-facto finan- using Geweke’s (1982) decomposition likely be ineffectual.

------NBER Reporter Winter 2005/6 27 Carvalho and Garcia analyze the incidence of crises, or to face a higher international markets. Second, there are Brazilian experience in the 1990s to probability of having a crisis, than coun- liquidity spillovers within markets. assess the effectiveness of controls on tries with lower mobility. His results do Aggregate domestic trading activity is capital inflows in restricting financial suggest, however, that once a crisis associated with the liquidity of individ- inflows and changing their composition occurs, countries with higher capital ual firms in the same market. The evi- towards long-term flows. Econometric mobility may face a higher cost, in terms dence is consistent with the view that exercises (VARs) lead them to conclude of growth decline. when firms cross-list or issue depositary that controls on capital inflows were Fernandez formulates a time-scale receipts in public international markets, effective in deterring financial inflows decomposition of an international ver- the domestic trading activity of their for only a brief period, from two to six sion of the Capital Asset Pricing Model shares falls, hurting the liquidity of the months. The hypothesis to explain the that accounts for both market and remaining firms in their home market. ineffectiveness of the controls is that exchange-rate risk. In addition, she Caballero, Hoshi, and Kashyap financial institutions performed several derives an analytical formula for time- proposes a bank-based explanation for operations aimed at avoiding capital scale value at risk and marginal value at the decade-long Japanese slowdown. controls. The authors conducted inter- risk (VaR) of a portfolio. She applies The starting point for their story is the views with market players in order to the methodology to stock indexes of well-known observation that most large provide several examples of the finan- seven emerging economies in Latin Japanese banks were only able to comply cial strategies that were used in this peri- America and Asia, for the sample peri- with capital standards because regulators od to invest in the Brazilian fixed od 1990-2004. Her main conclusions were lax in their inspections. To facilitate income market while bypassing capital are: 1) the estimation results hinge upon this forbearance, the banks often controls. Their main conclusion is that the choice of the world market portfo- engaged in sham loan restructuring that controls on capital inflows, while they lio. In particular, the stock markets of kept credit flowing to otherwise insol- may be desirable, are of very limited the sampled countries appear to be vent borrowers (“zombies”). Thus, the effectiveness under sophisticated finan- more integrated with other emerging normal competitive outcome, whereby cial markets. Therefore, policymakers countries than with developed ones. 2) the zombies would shed workers and should avoid spending the scarce Value at risk depends on the investor’s lose market share, was thwarted. The resources of bank supervision trying to time horizon. In the short run, potential authors’ model highlights the restructur- implement them and focus more in losses are greater than in the long run. ing implications of the zombie problem. improving economic policy. 3) Additional exposure to some specific The counterpart of the congestion cre- Edwards uses a broad multi-coun- stock indices will increase value at risk ated by the zombies is a reduction in try dataset to analyze the relationship to a greater extent, depending on the profits for potential new and more pro- between restrictions to capital mobility investment horizon. These results are in ductive entrants, which discourages their and external crises. The analysis focuses line with recent research in asset pricing entry. In this context, even solvent banks on two manifestations of external crises: that stresses the importance of hetero- will not find good lending opportunities. sudden stops of capital inflows; and cur- geneous investors. The authors confirm their story’s key rent account reversals. He deals with two What is the impact of internation- predictions — that zombie-dominated important policy-related issues: first, alization (firms raising capital and trad- industries exhibit more depressed job does the extent of capital mobility affect ing in international markets) on the liq- creation and destruction, lower produc- countries’ degree of vulnerability to uidity of the remaining firms in domes- tivity, and greater excess capacity. Most external crises; and second, does the tic markets? To address this question, importantly, they present firm-level extent of capital mobility determine the Levine and Schmukler assemble a regressions showing that the increase in depth of external crises — as measured panel database of more than 2,700 zombies has depressed the investment by the decline in growth — once the firms from 45 emerging economies and employment growth of non-zom- crises occur? Overall, his results cast over the period 1989-2000, constructed bies and has been associated with a some doubts on the assertion that from annual and daily data. First, they widening of the productivity gap increased capital mobility has caused find evidence of migration. There is a between zombies and non-zombies. The heightened macroeconomic vulnerabili- reduction in the domestic trading of evidence suggests that the healthiest ties. He finds no systematic evidence firms that cross-list or issue depositary non-zombies were harmed the most by suggesting that countries with higher receipts in foreign public exchanges as the zombies. capital mobility tend to have a higher trading migrates from domestic to A large body of literature has

28 NBER Reporter Winter 2005/6 ------stressed the institution-development Unlike the financial dollarization asset requirement (LAR) and an ex-ante nexus as critical in explaining differ- (FD) of external liabilities, the dollariza- suspension-of-convertibility clause or ences in countries’ economic perform- tion of domestic financial assets “circuit breaker” (CBR), as a way to ance. The empirical evidence, however, (domestic FD) has received compara- reduce self-insurance costs while limit- has been mainly on the aggregate level, tively less attention until very recently, ing bank losses in the event of a run. associating macro performance with when increasingly it has been seen as a Eichengreen, Gupta, and Mody measures of quality of institutions. A key source of real exchange rate expo- present evidence on the impact of IMF Costa and De Mello, by relating a judi- sure and balance sheet problems. Levy programs on sudden stops in capital cial decision on the legality of payroll -Yeyati focuses on an a complementary flows. Their results are consistent with debit loans in Brazil to bank-level deci- —and often overlooked— angle of the notion that IMF programs have sion variables, provide micro evidence domestic FD: the limit it imposes on some positive effect in reducing the on how creditor legal protection affects the central bank as domestic lender of incidence of these events. At the same market performance. Payroll debit loans last resort (LLR), and the resulting time, there is little evidence that larger are personal loans with principal and exposure to (dollar) liquidity runs. He Fund programs more effectively inocu- interests payments directly deducted addresses this issue in three steps. First, late countries against sudden stops. from the borrowers’ payroll check, he illustrates the incidence of FD on Newly-negotiated programs seem to be which, in practice, makes collateral out the propensity to suffer bank runs (and more effective in this regard than long- of future income. In June 2004, a high- the authorities’ belated reaction) by standing arrangements. It is tempting to level federal court upheld a regional means of two recent banking crises, interpret both observations as indicat- court ruling that had declared payroll Argentina 2001 and Uruguay 2002, and ing that the signaling effect of IMF pro- deduction illegal. Using personal loans shows that FD has been an important grams matters more than the emer- without payroll deduction as a control motive for self-liquidity insurance in the gency financial assistance. Finally, the group, the authors assess whether the form of reserve accumulation. Next, he authors are unable to identify evidence ruling had an impact on market per- explores the incentive problems associ- that IMF programs are more effective at formance. The evidence indicates that it ated with centralized self-insurance insulating countries from sudden stops had an adverse impact on banks’ risk (holdings of reserves at the central when they already have fundamentally perception, on interest rates, and on the bank). In this light, he argues for a com- strong policies in place. amount lent. bined scheme of decentralized liquid

------NBER Reporter Winter 2005/6 29 Structural Changes in the Global Economy

On December 9 and 10, an Discussant: James H. Stock, Harvard Hoyt Bleakley, University of Califor- NBER/Universities Research Confer- University and NBER nia, San Diego, and Kevin Cowan, ence on “Structural Changes in the Inter-American Development Bank, Global Economy: Implications for Sebnem Kalemli-ozcan, University of “Maturity Mismatch and Financial Monetary Policy and Financial Houston and NBER; Ariell Reshef, Crises: Evidence from Emerging Regulation” took place in Cambridge. New York University; and Bent E. Market Corporations” NBER Research Associates Andrew Sorensen, University of Houston, Discussant: Mark Aguiar, Federal B. Abel, The Wharton School, and “Productivity and Capital Flows: Reserve Bank of Boston Janice C. Eberly, Northwestern Uni- Evidence from U.S. States” versity’s Kellogg School of Manage- Discussant: John Coleman, Duke Prasanna Gai and Nicholas Vause, ment, organized this program: University Bank of England, and Peter Kondor, London School of Economics, F. Owen Irvine, Michigan State Charles A. Trzcinka and Andrey D. “Procyclicality, Collateral Values, and University, and Scott Schuh, Federal Ukhov, Indiana University, “Financial Financial Stability” Reserve Bank of Boston, “The Roles Globalization and Risk Sharing: Discussant: Adriano Rampini, North- of Comovement and Inventory Welfare Effects and the Optimality of western University Investment in the Reduction of Output Open Markets” Volatility” Discussant: Leonid Kogan, MIT and Söhnke Bartram, Lancaster Discussant: William Dupor, Ohio State NBER University; Gregory W. Brown, University University of North Carolina at Chapel Giovanni Olivei, Federal Reserve Hill; and John Hund, University of Stephen G. Cecchetti, Brandeis Bank of Boston, and Silvana Texas at Austin, “Estimating Systemic University and NBER; Alfonso Tenreyro, London School of Risk in the International Financial Flores-Lagunes, University of Ari- Economics, “The Timing of Monetary System” zona; and Stefan Krause, Emory Policy Shocks” Discussant: Craig Furfine, Federal University, “Assessing the Sources of Discussant: Marc Giannoni, Columbia Reserve Bank of Chicago Changes in the Volatility of Real University and NBER Growth”

Most of the reduction in GDP and inventory management techniques. responses to federal funds rate shocks, volatility since 1983 is accounted for by Using a HAVAR model (Fratantoni and thus suggesting that the case for “better a decline in the comovement of output Schuh 2003) with only two sectors — monetary policy” is complicated by among industries that hold inventories. manufacturing and trade — Irvine and changes in the real side of the economy. This decline is not simply a passive Schuh discover structural changes that In much of the world, growth is byproduct of reduced volatility in com- reduced the comovement of sales and more stable than it once was. Looking mon factors or shocks. Instead, struc- inventory investment both within and at a sample of 25 countries, Cecchetti, tural changes occurred in the long run between industries. As a result, the Flores-Lagunes, and Krause find that and there were dynamic relationships response of aggregate output to all in 16 of them, real GDP growth is less among industries’ sales and inventory types of shocks was dampened. volatile today than it was 20 years ago. investment behavior — especially in the Structural changes accounted for more And, these declines are large, averaging automobile and related industries, than 80 percent of the reduction in out- more than 50 percent. What accounts which are linked by supply and distribu- put volatility, thus weakening the case for the fact that real growth has been tion chains that feature new production for “good luck,” and altered industries’ more stable in recent years? The

30 NBER Reporter Winter 2005/6 ------authors survey the evidence and com- welfare effects of investment barriers very modest amount of uneven stagger- peting explanations and find support and the opening of markets to foreign- ing can generate differences in output for the view that improved inventory ers. They construct an equilibrium responses similar to those found in the management policies, coupled with model of international asset pricing data. financial innovation, adopting an infla- without agency costs that allows Substantial attention has been paid tion targeting scheme, and increased endogenous market participation in recent years to the risk of maturity central bank independence have all among heterogeneous agents. mismatch in emerging markets. been associated with more stable real Equilibrium prices and the set of par- Although this risk is microeconomic in growth. Furthermore, they find weak ticipating and non-participating agents nature, the evidence advanced thus far evidence suggesting that increased are jointly determined in equilibrium has taken the form of macro correla- commercial openness has coincided and the ability of agents to choose to tions. Bleakley and Cowan evaluate with increased output volatility. participate in the market affects prices this mechanism empirically at the micro Kalemli-Ozcan, Reshef, and of domestic and foreign assets. The level by using a database of over 3000 Sorensen study the determinants of authors examine the welfare effects of publicly traded firms from fifteen net capital income flows within the non-participation and find that when a emerging markets. They measure the United States where capital freely country moves from complete segmen- risk of short-term exposure by estimat- moves across state borders. They use a tation to open markets for foreigners, ing, at the firm level, the effect on simple neoclassical model in which total the cost of capital falls in the domestic investment of the interaction of short- factor productivity (TFP) varies across market. This is consistent with empiri- term exposure and aggregate capital states and over time and capital owner- cal findings in the international asset flows. This effect is (statistically) zero, ship is perfectly diversified across state pricing literature. Through the endoge- contrary to the prediction of the matu- borders. Capital will flow to states that nous participation mechanism, the rity-mismatch hypothesis. This conclu- experience an increase in TFP resulting model is able to capture sources of eco- sion is robust to using a variety of dif- in net cross-state investment positions. nomic growth. Contrary to previous ferent estimators, alternative measures Net ownership positions revert to zero models, however, this one shows that of capital flows, and controls for deval- over time in the absence of further TFP opening markets is not Pareto-optimal uation effects and access to internation- movements. States with increasing TFP and the authors identify a class of al capital. The authors do find evidence pay net capital income to states with domestic agents whose welfare is lower that short-term-exposed firms pay declining TFP relative to the U.S. aver- after the opening of markets. These higher financing costs and liquidate age. While TFP cannot be directly finding have political economy interpre- assets at resale prices, but not that this observed, the authors can identify states tations and policy implications. reduction in net worth translates into a with high TFP growth as states with A vast empirical literature has doc- drop in investment. high output growth. By comparing the umented delayed and persistent effects Gai, Kondor, and Vause analyze level of state personal income to state of monetary policy on output. Olivei how the risk-sharing capacity of the gross product, they construct indicators and Tenreyro show that this finding financial system varies over the business of net capital income flows. They then results from the aggregation of output cycle, leading to pro-cyclical fragility. examine empirically whether net capital impulse responses that differ sharply They show how financial imperfections income flows between states corre- depending on the timing of the shock: contribute to under-insurance by entre- spond to the predictions of the model When the monetary policy shock takes preneurs, generating a pecuniary exter- and whether net capital positions tend place in the first two quarters of the nality that leads to the build-up of sys- to converge to zero. The empirical find- year, the response of output is quick, tematic risk during upturns. Increased ings indicate persistent net capital sizable, and dies out at a relatively fast asset price uncertainty emerges as a income flows across states, which are an pace. In contrast, output responds very symptom of the sectoral concentration order of magnitude larger than the little when the shock takes place in the that builds up during booms. The liq- equivalent counterparts across coun- third or fourth quarter. The authors uidity of the collateral asset is shown to tries. Thus, the results imply that fric- propose a potential explanation for the play a key role in amplifying the finan- tions associated with borders are likely differential responses based on uneven cial cycle. The welfare costs of financial to be the main explanation for low staggering of wage contracts across stability, in terms of the efficiency costs international capital flows. quarters. Using a stylized dynamic gen- attributable to financial frictions and the Trzcinka and Ukhov study the eral equilibrium model, they show that a volatility costs attributable to amplifica-

------NBER Reporter Winter 2005/6 31 tion, also are illustrated. banks) rarely exceed a few percent. average default probabilities of only With a unique and comprehensive Second, they use structural models to around 1-2 percent. Taken together, the dataset, Bartram, Brown, and Hund obtain more precise point estimates of results suggest statistically significant, develop and use three distinct methods the likelihood of systemic failure. These but economically small, increases in sys- to quantify the risk of a systemic failure estimates suggest that systemic risk is temic risk around even the worst finan- in the global banking system. They limited even during major financial cial crises of the last ten years. Although examine a sample of 334 banks (repre- crises. For example, maximum likeli- policy responses are endogenous, the senting 80 percent of global bank equi- hood estimation of bank failure proba- low estimated probabilities suggest that ty) in 28 countries around 6 global bilities implied by equity prices suggests the distress of central bankers, regula- financial crises (such as the Asian and that the Asian crisis induced less than a tors, and politicians about the events Russian crises and September 11, 2001), 1 percent increase in the probability of studied here may be over-stated, and and show that these crises did not cre- systemic failure. Third, they obtain esti- that current policy responses to finan- ate large probabilities of global financial mates of systemic risk implied by equi- cial crises and the existing institutional system failure. First, they show that ty option prices of U.S. and European framework may be adequate to handle cumulative negative abnormal returns banks. The largest values are obtained major macroeconomic events. for the subset of banks not directly for the Russian crisis and September 11 exposed to a negative shock (unexposed and these show increases in estimated

32 NBER Reporter Winter 2005/6 ------18th Annual TRIO Conference

The Eighteenth Annual TRIO Shang-jin Wei, IMF and NBER, Paolo Pesenti, “Shocks, Reforms, and Conference, so-named because it is “Connecting Two Views on Financial Monetary Rules: A Scenario Analysis jointly sponsored by the NBER, the Globalization: Can We Make Further for Japan” Centre for Economic Policy Research Progress?” Discussants: Kazuo Ueda, University (CEPR), and the Tokyo Center for Discussants: Yuko Hashimoto, Toyo of Tokyo, and Tsutomu Watanabe, Economic Research (TCER), took University, and Elias Papaioannou, Hitotsubashi University place on December 9 and 10 in European Central Bank Tokyo. This year’s conference focused Richard Portes, London Business on “International Finance.” It was Takatoshi Ito, and Yuko Hashimoto, School and NBER; Elias organized by Shin-ichi Fukuda, “Intra-Day Seasonality in Activities of Papaioannou; and Gregorios University of Tokyo; Takeo Hoshi, the Foreign Exchange Markets: Siourounis, Barclays Capital, NBER and University of California, Evidence from the Electronic Broking “Optimal Currency Shares in San Diego; Takatoshi Ito, NBER and System” International Reserves: The Impact University of Tokyo; and Andrew K. Discussants: Robert F. Engle, Univer- of the Euro and the Prospects for the Rose, NBER and University of sity of California, San Diego and Dollar” California, Berkeley. The program NBER, and Paolo Pesenti, Federal Discussants: Takeo Hoshi and was: Reserve Bank of New York Takatoshi Ito Eiji Ogawa and Junko Shimizu, Shin-ichi Fukuda, and Masanori Mark Spiegel; Takeshi Kobayashi, Hitotsubashi University, “Stabilization Ono, Fukushima University, “On the Chukyo University; and Nobuyoshi of Effective Exchange Rates under a Determinants of Export Prices: Yamori, Nagoya University, Common Currency Basket System” History vs. Expectations” “Quantitative Easing and Japanese Discussants: Taizo Motonishi, Kansai Discussants: Andrew K. Rose, and Bank Equity Values” University, and Mark Spiegel, Federal Kiyotaka Sato, Yokohama National Discussants: Naohiko Baba, Bank of Reserve Bank of San Francisco University Japan, and Itsuhiro Fukao, Keio University Etsuro Shioji, Yokohama National Allan Drazen, University of Maryland University, “Invoicing Currency and and NBER, and Stefan Hubrich, T. Andrew K. Rose, “Size Really Doesn’t the Optimal Basket Peg for East Asia: Rowe Price, “A Simple Test of the Matter: In Search of a National Scale A New Open Economy Macroeco- Effect of Exchange Rate Defense” Effect” nomics Perspective” Discussants: Shin-ichi Fukuda, and Discussants: Allan Drazen and Etsuro Discussants: Kentaro Iwatsubo, Shigenori Shiratsuka, Bank of Japan Shioji Hitosubashi University, and Eiji Ogawa

Ogawa and Shimizu investigate stabilization effects of the common G3 exchange rates than a common G-3 cur- how effectively a common currency currency (the U.S. dollar, the Japanese rency basket peg system for four basket peg would stabilize the effective yen, and the euro) basket in the East (Indonesia, the Philippines, Singapore, exchange rates of East Asian curren- Asian countries (Williamson, 2005). and Thailand) of the seven countries cies. The authors use an Asian They find that the AMU peg system they study. These results suggest that Monetary Unit (AMU), which is a would be more effective in reducing the AMU basket peg would be useful weighted average of the ASEAN10 plus fluctuations of effective exchange rates for the East Asian countries whose 3 (Japan, China, and Korea) currencies, as more countries in East Asia applied intraregional trade weights are relatively as the common currency basket; they it. Further, the AMU peg system would higher than their trade weights with compare their results with others on the more effectively stabilize effective outsiders.

------NBER Reporter Winter 2005/6 33 Shioji analyzes the relationship financial development and the quality of quencies are high (low, respectively), between East Asia’s choice of currency bureaucratic institutions. Wei shows except for the beginning hour of Tokyo regime and the transmission of foreign that these effects can indeed be differ- (GMT 0), when the bid-ask spread is shocks to this area. He develops a three- ent. In particular, bad public institutions ideal despite high levels of activity. country model that consists of East ( reflected, for example, in a higher level Fukuda and Ono investigate the Asia, Japan, and the United States, in of bureaucratic corruption) strongly choice of invoice currency under the tradition of the “new open econo- discourage FDI, and possibly foreign exchange rate uncertainty. Their analysis my macroeconomics” literature. Using debt, in the shares of a country’s total is motivated by the fact that the U.S. numerical simulations, Shioji derives the foreign liabilities, but appear to encour- dollar has been the dominant vehicle optimal weight attached to the Japanese age the relative prominence of borrow- currency in developing countries. Their yen in East Asia’s currency basket, to ing from foreign banks. In comparison, theoretical analysis is based on an open- which this region pegs its own currency; low financial sector development dis- economy model of monopolistic com- optimality is defined with respect to sta- courages inward portfolio equity flows petition. The export prices are set bilization of its trade balance (or other but encourages inward FDI. Therefore, before exchange rates are known. When measures). In particular, this paper takes views on the connection between the market is competitive enough, the into account the reality that most inter- domestic institutions and the structure exporting firms tend to set their prices national transactions are invoiced in the of international capital flows must be so as not to deviate from those of the U.S. dollar, and asks how incorporating nuanced. To gain confidence that the competitors. As a result, a coordination that fact into the model changes the documented data patterns reflect causal failure can lead the third currency to be conclusion about the optimal basket relations, Wei uses instrumental vari- an equilibrium invoice currency. Since weights. ables for the institutional measures multiple equilibriums are Pareto ranked, For many developing countries, based on the economic histories of the this implies that the equilibrium choice financial globalization does not auto- countries in his sample (in particular, of the invoice currency may lead to a matically lead to improvement. the mortality rate of earlier European less efficient equilibrium. The role of According to the literature, there is a settlers and the origin of legal systems). expectations is important in the static threshold effect: only countries that The instrumental variables approach framework. However, in the staggered have met a minimum set of conditions, bolsters the case that bad institutions price-setting framework, history such as having attained reasonable con- are a cause of unfavorable composition becomes another key determinant of trol of corruption and a certain level of of capital inflows. the equilibrium currency pricing. The rule of law, can expect to benefit signif- Ito and Hashimoto examine intra- role of history becomes conspicuous icantly from financial globalization. day patterns of exchange rate behavior, when the firms discount future profits, And, there is a composition effect: for- using the “firm” bid-ask quotes and particularly in the competitive local eign direct investment (FDI), and per- transactions of JPY-USD and Euro- market. The result suggests that both haps portfolio inflows, are likely to be USD pairs recorded in the electronic history and expectations explain why more beneficial and less volatile than broking system of the spot foreign the firm tends to choose the U.S. dollar international bank lending, while total exchanges. First, activities of quotes as vehicle currency. capital flows — the sum of all types of and transaction volumes are high in the High interest rates used to defend capital flows — may not have a strong beginning hours of the three major cur- the exchange rate signal that a govern- positive effect on the recipient coun- rency markets — Tokyo, London, and ment is committed to fixed exchange tries’ rates of growth and their con- New York — and low during the Tokyo rates, but may also signal weak funda- sumption risk sharing. Further, the and London lunch hours and late after- mentals. Drazen and Hubrich test the threshold and composition effects can noon in New York. The U-shape of effectiveness of the interest rate be two sides of the same coin, as better intra-day activities only occurs among defense by disaggregating it into the institutional quality in a capital-import- Tokyo and London participants. effects on future interest rate differen- ing country may lead to a more favor- Second, activities do not increase tials, expectations of future exchange able composition of capital inflows for toward the end of business hours in the rates, and risk premiums. While much that country (Wei, 2000b, 2001; Wei and New York market, even on Fridays previous empirical work has been Wu, 2002; and Faria and Mauro, 2004). (ahead of weekend hours of non-trad- inconclusive because of offsetting But the earlier literature did not disen- ing). Third, an average bid-ask spread is effects, tests that “disaggregate” the tangle the possibly different effects of narrow (wide), when quote and deal fre- effects provide significant information.

34 NBER Reporter Winter 2005/6 ------Raising overnight interest rates stantial implications for the internation- credit extension by the troubled strengthens the exchange rate over the al financial system. Papaioannou, Japanese financial sector. Kobayashi, short-term, but also leads to an expect- Portes, and Siourounis use a mean- Spiegel, and Yamori conduct an event ed depreciation at a horizon of a year variance framework to estimate optimal study concerning the anticipated impact and longer and an increase in the risk weights among the main international of quantitative easing on the Japanese premium, consistent with the argument currencies and to assess how the euro banking sector by examining the impact that it also signals weak fundamentals. has changed this allocation over time. of the introduction and expansion of Using a two-country general equi- They also incorporate rebalancing costs, the policy on Japanese bank equity val- librium model calibrated to the Japanese which they proxy with (mean and ues. They find that excess returns of economy vis-à-vis the rest of the world, extreme) currency bid-ask spreads. The Japanese banks were greater when Pesenti simulates the macroeconomic results indicate that the recent drop in increases in the BOJ current account transmission of demand and supply euro spreads fully compensated for the balance target were accompanied by shocks contingent on whether or not diversification losses associated with “non-standard” expansionary policies, the zero interest floor (ZIF) is binding fewer currencies. The authors then per- such as raising the ceiling on BOJ pur- in monetary policy. First, he shows that form some simple simulations for the chases of long-term Japanese govern- negative demand shocks have more optimal currency allocation of four ment bonds. The authors also provide prolonged and startling effects on the large emerging market countries cross-sectional evidence that suggests economy when the ZIF is binding than (Russia, Brazil, China and India) incor- that the market perceived that the quan- during normal times when it is not porating a central bank’s desire to hold titative easing program would dispro- binding. Next, he illustrates how posi- a sizable portion of its portfolio in the portionately benefit financially weaker tive supply shocks that raise potential currencies of its foreign debt and inter- Japanese banks. output (such as structural reforms) can national trade. The constrained opti- Rose searches for a “scale” effect actually extend the period of time over mization suggests that the euro poten- in countries. He uses a panel data set which the ZIF may be expected to bind, tially rivals the dollar as an international that includes 200 countries over forty and therefore make the economy more reserve currency. Actual dollar alloca- years and links the population of a sensitive to negative demand shocks. tions are far greater than the optimizer country to a host of economic and Finally, he focuses on the problems implies, consistent with the current social phenomena. Using both graphical associated with inflation-targeting rules dominant role of the dollar as a reserve and statistical techniques, he searches and the advantages of policy rules that currency. But the increased tendency of for an impact of size on the level of include price-level-path targeting, both many developing countries to issue income, inflation, material well-being, in a deflationary environment and in euro-denominated assets and trade with health, education, the quality of a coun- normal times when the ZIF is not bind- the euro zone may shift this equilibrium try’s institutions, heterogeneity, and a ing. and put pressure on the dollar. number of different international Foreign exchange reserve accumu- One of the primary motivations indices and rankings. He has little suc- lation has risen dramatically over the offered by the Bank of Japan (BOJ) for cess; small countries are more open to past five years. The introduction of the its quantitative easing program — international trade than large countries, euro and the increased liquidity in other whereby it maintained a current account but are not systematically different major currencies has increased the pres- balance target in excess of required otherwise. sure on central banks to diversify away reserves, effectively pegging short-term from the dollar. This could have sub- interest rates at zero — was to maintain

------NBER Reporter Winter 2005/6 35 Bureau News

Science Watch Ranks NBER First in Citations Science Watch, the bimonthly per paper. These rankings are based on between 1995 and April 2005, the newsletter of Thomson Scientific, the number of published papers cited in NBER was the most mentioned institu- named the NBER as the institution nearly 200 journals in economics, busi- tion in the field of economics and busi- with the most cited publications in eco- ness, and accounting and management ness. Based on these figures, Science nomics and business. The NBER also that are indexed by Thomson Scientific, Watch delared that the NBER is one of ranked third in “citation impact” as and on citations per paper. the most influential institutions in busi- measured by the number of citations With more than 30,000 citations ness and economics.

Monetary Economics

NBER’s Program on Monetary “Three Great American Disinflations” Phillips Curve” Economics met in Cambridge on Discussant: Francois Velde, Federal Discussant: Jean Boivin, Columbia November 4. NBER Research Reserve Bank of Chicago University Associates Michael D. Bordo of Rutgers University and Julio J. Refet Gurkanak, Bilkent University; Igor Livshits and James Macgee, Rotemberg of MIT organized the Andrew Levin; and Eric Swanson, University of Western Ontario; and meeting. The following papers were Federal Reserve Bank of San Michele Tertilt, Stanford University, discussed: Francisco, “Does Inflation Targeting “Accounting for the Rise in Consumer Anchor Long-Run Inflation Expecta- Bankruptcies” Gauti Eggertsson, Federal Reserve tions? Evidence from Long-Term Discussant: Stephen Zeldes, Colum- Bank of New York, “Great Expecta- Bonds Yields in the U.S., U.K., and bia University tions and the End of the Great De- Sweden” pression” Discussant: Kenneth Kuttner, Ober- Lutz Kilian, University of Michigan, Discussant: Hugh Rockoff, Rutgers lin College “Exogenous Oil Supply Shocks: How University Big Are They and How Much Do Timothy Cogley, University of They Matter for the U.S. Economy?” Michael D. Bordo, Christopher California, Davis, and Argia Discussant: Ana Maria Herrera, Erceg, Andrew Levin, and Ryan Sbordone, Federal Reserve Bank of Michigan State University Michaels, Federal Reserve Board, New York, “A Search for a Structural

Eggertsson argues that the recov- side, FDR expanded government real lowing the Civil War, World War I, and ery from the Great Depression was and deficit spending, making his policy the Volcker disinflation of the early driven by a shift in expectations. This objective credible. Eggertsson evalu- 1980s. For each of these episodes, they shift was caused by President Franklin ates the economic consequences of derive measures of policy predictabili- Delano Roosevelt’s (FDR) policy FDR; he uses a dynamic stochastic ty that attempt to quantify the extent actions. On the monetary policy side, general equilibrium model, assuming to which each deflation was anticipated FDR abolished the gold standard and sticky prices and rational expectations. by economic agents. They use their — even more importantly — In their paper, Bordo and his co- measures to help account for the dis- announced an explicit policy objective authors examine three famous parate real effects observed across of inflating the price level to pre- episodes of disinflation (or deflation) episodes, and in turn relate them to the Depression levels. On the fiscal policy in U.S. history, including episodes fol- policy actions and communication

36 NBER Reporter Winter 2005/6 ------strategy of the monetary authority. ly helps to anchor the private sector’s there has been strong interest in the They then proceed to account for the views of the distribution of long-run question of how oil production short- salient features of each episode within inflation outcomes. falls caused by wars and other exoge- the context of a stylized model. Their The foundation of the New nous political events in OPEC countries simulations indicate how a more pre- Keynesian Phillips curve is a model of affect oil prices, U.S. real GDP growth, dictable policy of gradual deflation price setting with nominal rigidities that and U.S. CPI inflation. Kilian focuses could have helped avoid the sharp post- implies that the dynamics of inflation on the modern OPEC period since WWI depression. But the simulations are well explained by the evolution of 1973. His results differ along a number also suggest that securing the benefits real marginal costs. Cogley and of dimensions from the conventional of gradualism requires a supporting Sbordone attempt to analyze whether wisdom. First, he shows that under rea- institutional framework and communi- this is a structurally invariant relation- sonable assumptions, the timing, mag- cation strategy that allows the private ship. To assess this, they first estimate nitude, and even the sign of exogenous sector to make reliable inferences about an unrestricted time-series model for oil supply shocks may differ greatly the course of policy. inflation, unit labor costs, and other from current state-of-the-art estimates. Gürkaynak, Levin, and Swanson variables, and present evidence that Second, the common view — that the investigate the extent to which inflation their joint dynamics are well represent- case for the exogeneity of at least the targeting helps anchor long-run infla- ed by a vector autoregression with drift- major oil price shocks is strong — is tion expectations by comparing the ing coefficients and volatilities, as in supported by the data for the 1980–1 behavior of daily bond yield data in the Cogley and Sargent (2004). Then, fol- and 1990–1 oil price shocks, but not for United Kingdom and Sweden, both lowing Sbordone (2002, 2003), they other oil price shocks. Notably, statisti- inflation targeters, to that in the United apply a two-step minimum distance cal measures of the net oil price that States, a non-inflation-targeter. Using estimator to estimate deep parameters. increase relative to the recent past do the difference between far-ahead for- Based on their results, they argue that not represent the exogenous compo- ward rates on nominal and indexed the price-setting model is structurally nent of oil prices. In fact, only a small bonds as a measure of compensation invariant. fraction of the observed oil price for expected inflation and inflation risk Personal bankruptcies in the United increases during crisis periods can be at long horizons, the authors examine States have increased dramatically, rising attributed to exogenous oil production the extent to which far-ahead forward from 1.4 per thousand working age disruptions. Third, compared to previ- inflation compensation moves in population in 1970 to 8.5 in 2002. ous indirect estimates of the effects of response to macroeconomic data releas- Livshits, MacGee, and Tertilt use a exogenous supply disruptions on real es and monetary policy announcements. heterogeneous agent life-cycle model GDP growth that treated major oil In the United States, forward inflation with competitive financial intermedi- price increases as exogenous, the direct compensation exhibits highly signifi- aries who can observe households’ estimates that Kilian obtains suggest a cant responses to economic news. In earnings, age, and current asset holdings sharp drop after five quarters rather the United Kingdom, there is a level of to evaluate several commonly offered than an immediate and sustained reduc- sensitivity similar to that in the United explanations. They find that an increase tion in economic growth for a year. States prior to the Bank of England in uncertainty (income shocks, expense They also suggest a spike in CPI infla- gaining independence in 1997, but a uncertainty) cannot quantitatively tion three quarters after the exogenous striking absence of such sensitivity account for the rise in bankruptcies. oil supply shock rather than a sustained since the central bank became inde- Instead, stories related to a change in increase in inflation, as is sometimes pendent. In Sweden, inflation compen- the credit market environment are more conjectured. Finally, Kilian’s results put sation has been insensitive to economic plausible. In particular, a combination into perspective the importance of news over the whole period for which of a decrease in the credit market trans- exogenous oil production shortfalls in the authors have data. The authors actions cost together with a decline in the Middle East. He shows that exoge- show that these results also are matched “stigma” does a good job at accounting nous oil supply shocks made remark- by the times-series behavior of far- for the rise in consumer bankruptcy. ably little difference overall for the evo- ahead forward interest rates and infla- The authors also argue that the aboli- lution of U.S. real GDP growth and tion compensation over this period. All tion of usury laws and other legal CPI inflation since the 1970s, although of these findings suggest that a known changes have played little role. they did matter for some historical and credible inflation target significant- Since the oil crises of the 1970s, episodes.

------NBER Reporter Winter 2005/6 37 Health Care Program Meeting

The NBER’s Health Care Mark Pauly, University of Dana Goldman, RAND Corporation Program met in Cambridge on Pennsylvania and NBER; Christy and NBER; Pinar Karaca-mandic November 4. NBER Research Thompson, Independence Blue and Geoffrey Joyce, RAND Associate David M. Cutler, of Cross; Thomas Abbott, Medstat, Inc.; Corporation; and Neeraj Sood, Harvard University, and Program William Sage, Columbia University; RAND Corporation and NBER, Director Alan M. Garber, of Stanford and James Margolis, Medical Group “Adverse Selection in Retiree Prescrip- University, organized the meeting. Management Association, “Who Pays: tion Drug Plans” These papers were presented: the Incidence of Higher Malpractice Premiums” Ernst R. Berndt, MIT and NBER; Nancy Beaulieu, Harvard University Alisa S. Busch and Sharon-lise and NBER; David M. Cutler; and Yu-chu Shen, Naval Postgraduate Normand, Harvard University; and Katherine Ho, Columbia University School and NBER; and Glenn Richard G. Frank, Harvard University and NBER, “The Business Case for Melnick, University of Southern and NBER, “Real Output in Mental Diabetes Disease Management at Two California, “Is Managed Care Still an Health Care During the 1990s” (NBER Managed Care Organizations” Effective Cost Containment Device?” Working Paper No. 11557)

Diabetes is the most common and and benefits worsen the health plan busi- physicians, public attitudes may be differ- costly of all chronic diseases. There is ness case. In addition, the payment sys- ent than if it is shared with the public. broad-based agreement on how to man- tems, from purchaser to health plan and Pauly, Thompson, Abbott, Sage, and age the disease, yet fewer than 40 percent health plan to provider, are very weakly Margolis report on a study of premium of diabetics receive guideline levels of connected to the quality of diabetes care incidence over the period 1994-2002, medical care. Beaulieu, Cutler, and Ho further weakening the business case. when the malpractice insurance system investigate the reasons for this phenome- Finally, overlapping provider networks again went into “crisis” as premiums rose non by examining the business case for create a public goods externality that lim- significantly in some geographic areas improved diabetes care from the per- its the health plan’s ability to privately and for some kinds of physicians. spective of a single health plan capture the benefits from its investments. Shen and Melnick take a historical (HealthPartners of Minnesota). The Malpractice premiums are higher in perspective in examining the effects of potential benefits accruing to a health some states than in others for apparently managed care growth and hospital com- plan from diabetes disease management similar physician practices. They are ris- petition on hospital cost and revenue include medical care cost savings and ing, and they are rising at different rates. growth. Looking at managed care’s boom higher premiums. The potential costs to Someone clearly is paying more into the period (1990-4), its mature period (1994- the health plan derive from disease man- health care or health insurance system, 8), and its backlash period (1998-2003), agement program costs and adverse but who? In the first instance, obviously they find that higher managed care pres- selection. The authors find that the physician practices pay the malpractice ence was indeed effective in slowing implementation of diabetes disease man- premium, but they may be able to shift down hospital cost and revenue growth agement coincided with large health some or all of high or growing premiums during the boom and the mature periods. improvements. Medical care cost savings onto insurers and patients. The question However, it lost its cost containment over several years were small in the closed of the “incidence” of premiums is an effect during the backlash period. This panel group practice but moderate for important part of understanding how the result persists under different estimation the health plan overall. The difference in system behaves and has been behaving methods designed to reduce biases that cost savings between these two patient over time. An answer to this question might result from omitted variable bias populations could be attributable to scale would also help in judging the distribu- and measurement errors. On the other or differences in the baseline health of tion of gains and losses from efforts to hand, competition effects appear to per- the populations. They find evidence that constrain premiums or damage awards. If sist throughout the three periods. Such adverse selection and the timing of cost all the gain from lower premiums goes to persistent competition effects were ini-

38 NBER Reporter Winter 2005/6 ------tially the result of aggressive selective and could lead to future health plan sav- services. In an era of consumer-directed contracting in the high managed care ings in terms of avoided health services health care and improved information markets, but were later dominated by the utilization. The authors model such a co- technology, tailoring co-payments to the less saturated, but growing, managed care payment scheme for one of the largest expected therapeutic benefit of a patient markets that seem to catch up with the classes of prescription drugs: cholesterol- can increase the clinical and economic more developed markets. lowering (CL) therapy. Using claims data efficacy of prescription medications. Rising co-payments for prescription from 88 health plans, they study 62,274 Health accounts document changes drugs, coupled with already low rates of patients aged 20 and older who initiated over time in the level and composition of compliance for chronic therapies, raise CL therapy between 1997 and 2001. health spending. There has been a con- concerns about the consequences of the They examine the association between tinued evolution in the ability to track design of pharmacy benefits. Goldman, co-payments and compliance in the year such outlays. Less rapid has been the abil- Karaca-mandic, Joyce, and Sood con- following initiation of therapy, and the ity to interpret changes in spending. sider one innovative such benefit by association between compliance and sub- Berndt, Busch, Frank and Normand which patients with the greatest thera- sequent hospital and emergency depart- apply quality-adjusted price indexes for peutic benefit from a prescription drug ment service use for up to four years fol- several major mental disorders to nation- have lower co-payments. Patients often lowing initiation. They use the results to al mental health account estimates to do not fully internalize future medical simulate the effects of co-payments that assess changes in real “output”. They benefits of a drug therapy and hence do vary depending on a patient’s risk of car- show that using the new price indexes not use the drug optimally. Better price diovascular events. They show that strate- reveals large gains in real output relative incentives (through lower co-payments) gically reducing co-payments for patients to application of BLS indexes. to patients with higher potential efficacy most at-risk can improve overall compli- would encourage optimal compliance ance and reduce use of other expensive

Macroeconomics and Individual Decisionmaking

The NBER’s Working Group on Groshen, Federal Reserve Bank of Forecasts, Uncertainty and Risk” Macroeconomics and Individual New York; Steinar Holden, Discussant: Paul Willen, Federal Decisionmaking met in Cambridge on University of Oslo; Julian Messina, Reserve Bank of Boston November 5. Working Group Jarkko Turunen, and Melanie Directors George A. Akerlof, Ward, European Central Bank; and Ricardo J. Caballero, MIT and University of California, Berkeley, and Mark E. Schweitzer, Federal NBER, and Arvind Krishnamurthy, Robert J. Shiller, NBER and Yale Reserve Bank of Cleveland, “The Northwestern University, “Financial University, set the following agenda: Interaction of Labor Markets and System Risk and Flight to Quality” Inflation: Analysis of Micro Data Discussant: Jon Faust, Federal Nabil Al-najjar, Sandeep Baliga, from the International Wage Flexibil- Reserve Board and David Besanko, Northwestern ity Project” University, “The Sunk Cost Bias and Discussant: Ricardo Reis, Princeton Annamaria Lusardi, Dartmouth Managerial Pricing Practices” University and NBER College and NBER, and Olivia S. Discussant: Truman Bewley, Yale Mitchell, University of Pennsylvania University Refet S. Gürkaynak, Bilkent Uni- and NBER, “Financial Literacy and versity, and Justin Wolfers, Univer- Planning: Implications for Retirement William T. Dickens, Brookings sity of Pennsylvania and NBER, Well-Being” Institution; Lorenz Goette, “Macroeconomic Derivatives: An Ini- Discussant: Andrew Caplin, New University of Zurich; Erica L. tial Analysis of Market-Based Macro York University and NBER

Al-Najjar, Baliga, and Besanko sunk cost bias persists among firms oligopoly. Firms experiment with cost provide an explanation for why the competing in a differentiated product methodologies that are consistent with

------NBER Reporter Winter 2005/6 39 real-world accounting practices, includ- substantial downward real wage rigidity. surprises in data. These markets also ing ones that allocate fixed and sunk At the same time, inflation can throw provide implied probabilities of the full costs to determine variable costs. These sand in the wheels of economic adjust- range of specific outcomes, allowing firms follow naive adaptive learning to ment by degrading the value of price the authors to measure uncertainty, adjust prices. Costing methodologies signals. Knowledge of which of these assess its driving forces, and compare that increase profits are reinforced. The imperfections dominates at different this measure of uncertainty with the authors show that all firms eventually levels of inflation and under different dispersion of point-estimates among display the sunk cost bias. For the institutional regimes can be valuable for individual forecasters (a measure of dis- canonical case of symmetric linear choosing an inflation target and for agreement). They also assess the accura- demand, they obtain comparative statics learning more about the economic envi- cy of market-generated probability den- results showing how the degree of the ronment in which monetary policy is sity forecasts. A consistent theme is that sunk cost bias changes with demand. conducted. The authors briefly review few of the behavioral anomalies present The adoption of explicit or implic- the empirical literature in order to moti- in surveys of professional forecasts sur- it inflation targets by many central vate the method used to distinguish vive in equilibrium, and that these mar- banks, and the low stable rates of infla- these three labor market imperfections. kets are remarkably well calibrated. tion that have ensued, raise the question Next, they describe the data and empir- Finally, they assess the role of risk, find- of how inflation affects market efficien- ical approach which applies a common ing little evidence that risk-aversion cy. Dickens, Goette, Groshen, protocol to 31 distinct panels of work- drives a wedge between market prices Holden, Messina, Schweitzer, ers wage changes. Then they establish and probabilities in this market. Turunen, and Ward study three market that wage changes show substantial dis- Caballero and Krishnamurthy imperfections that cause the rate of persion that rises with the rate of wage present a model of flight to quality inflation to affect labor market efficien- inflation, as predicted by grease and episodes that emphasizes financial sys- cy. First, the presence of substantial sand effects. To identify the three tem risk and the Knightian uncertainty resistance to nominal wage cuts in a low imperfections under consideration, they surrounding these episodes. In the inflation environment can slow the examine histograms of wage changes model, agents are uncertain about the adjustment of relative wages to labor (that are corrected for measurement probability distribution of shocks in market shocks and thus result in a mis- errors) for the particular asymmetries markets different from theirs, treating allocation of resources. Alternatively, to and spikes that are characteristic of such uncertainty as Knightian. Aversion the extent that the downward rigidity downward real and nominal wage rigid- to Knightian uncertainty generates prevents real wage cuts, rather than ity. This process yields estimates of the demand for safe financial claims. It also nominal wage cuts, inflation will not prevalence of real and nominal wage leads agents to require financial inter- improve efficiency. In this case, only rigidity for each dataset and year, which mediaries to lock-up capital to cover increases in real wages resulting from they then analyze for insight into the their own markets’ shocks in a manner productivity growth can reduce the mis- causes and consequences of wage that is robust to uncertainty over other allocation of resources caused by a real rigidities. Finally, they examine the link- markets, but is wasteful in the aggre- wage floor. Higher inflation is associat- age between estimates of true wage gate. Locked collateral cannot move ed with more frequent wage and price change dispersion and inflation for evi- across markets to offset negative shocks changes, higher search costs for goods dence of sand effects. and hence can trigger a financial accel- or jobs, and greater uncertainty about In September 2002, a new market erator. A lender of last resort can the future path of wages and prices. in “Economic Derivatives” was unlock private capital markets to stabi- These effects can lead to errors and launched allowing traders to take posi- lize the economy during these periods adjustment lags in wage setting and tions on future values of several macro- by committing to intervene should con- diminish the information value of economic data releases. Gürkaynak ditions worsen. observed wages. Thus, increased infla- and Wolfers provide an initial analysis Some recent evidence suggests that tion may also cause a misallocation of of the prices of these options. They many American households will not be resources. In short, inflation can grease find that market-based measures of able to maintain their lifestyles in retire- the wheels of economic adjustment in expectations are similar to survey-based ment. Little is known about why people the labor market by relieving the con- forecasts although the market-based fail to plan for retirement, and whether straint imposed by downward nominal measures somewhat more accurately planning and information costs might wage rigidity, but not if there is also predict financial market responses to affect retirement saving patterns. To

40 NBER Reporter Winter 2005/6 ------better understand these issues, Lusardi and only one-third understood these as respondents believed that they engaged and Mitchell devised and fielded a pur- well as stock market risk. Women, in successful retirement planning. The pose-built module on planning and minorities, and those without a college authors also find that financial knowl- financial literacy for the 2004 Health degree were particularly at risk of dis- edge and planning are clearly interrelat- and Retirement Study (HRS). This playing low financial knowledge. The ed: those who displayed financial module measures how workers make authors also evaluate whether people knowledge were more likely to plan and their saving decisions, how they collect tried to figure out how much they need to succeed in their planning. Moreover, the information for making these deci- to save for retirement, whether they those who did plan were more likely to sions, and whether they possess the devised a plan, and whether they suc- rely on formal planning methods such financial literacy needed to make these ceeded at the plan. In fact, these calcu- as retirement calculators, retirement decisions. The resulting analysis shows lations prove to be difficult: fewer than seminars, and financial experts, and less that financial illiteracy is widespread one-third of our age 50+ respondents likely to rely on family/relatives or co- among older Americans: only half of ever tried to devise a retirement plan, workers. Finally, Lusardi and Mitchell the age 50+ respondents could correct- and only two-thirds of those who tried, show that keeping track of spending ly answer two simple questions regard- actually claim to have succeeded. and budgeting habits appears conducive ing interest compounding and inflation, Overall, fewer than one-fifth of the to retirement saving.

Asset Pricing

The NBER’s Program on Asset Jun Pan, MIT and NBER, and University and NBER Pricing met in Cambridge on Kenneth Singleton, Stanford November 11. Jessica Wachter, University and NBER, “Default and Jacob Boudoukh, Matthew NBER and The Wharton School, and Recovery Implicit in the Term Richardson, and Robert Whitelaw, Luis M. Viceira, Harvard Business Structure of Sovereign CDS Spreads” New York University and NBER, “The School, organized the meeting. The Discussant: Francis Longstaff, Univer- Myth of Long-Horizon Predictability” following papers were discussed: sity of California, Los Angeles and NBER Amit Goyal, Emory University, and Bernard Dumas, INSEAD and Ivo Welch, Brown University and NBER, and Alexander Kurshev and Ravi Bansal and Ed Fang, Duke NBER, “A Comprehensive Look at the Raman Uppal, London Business University, and Amir Yaron, University Empirical Preformance of Equity School, “What Can Rational Investors of Pennsylvania and NBER, “Equity Premium Prediction” Do About Excessive Volatility and Capital: A Puzzle?” Sentiment Fluctuations?” Discussant: John Heaton, University of John Y. Campbell, Harvard Univer- Discussant: Leonid Kogan, MIT and Chicago and NBER sity and NBER, and Samuel B. NBER Thompson, Harvard University, “Pre- Borja Larrain, Federal Reserve Bank dicting the Equity Premium Out of Lubos Pastor and Pietro Veronesi, of Boston, and Motohiro Yogo, Sample: Can Anything Beat the His- University of Chicago and NBER, University of Pennsylvania, “Does torical Average?”(NBER Working Pa- “Technological Revolutions and Stock Firm Value Move Too Much to be per No. 11468) Prices” Justified By SubsequentChanges in Discussant for all three papers: John Discussant: Markus Brunnermeier, Cash Flow?” Cochrane, University of Chicago and Princeton University and NBER Discussant: Malcolm Baker, Harvard NBER

Dumas, Kurshev, and Uppal ana- “excessive” stock price volatility and sentiment. In it, there are two classes of lyze the trading strategy that would “sentiment” fluctuations. They con- agents; stock prices are excessively allow an investor to take advantage of struct a general equilibrium model of volatile because one class is overconfi-

------NBER Reporter Winter 2005/6 41 dent about a public signal. As a result, stock prices is observable ex post but sector, these inducements can be in the this class of irrational agents changes its unpredictable ex ante, and it is most pro- form of higher-mean or lower-volatility expectations too often, sometimes being nounced for technologies characterized assets. Bansal, Fang, and Yaron docu- excessively optimistic, sometimes being by high uncertainty and fast adoption. ment that shifts in sectoral financial excessively pessimistic. The authors The authors examine stock prices in the wealth have virtually no bearing on the determine and examine the trading strat- early days of American railroads, and subsequent mean and volatility of sec- egy of the rational investors who are not find evidence consistent with a large- toral returns. About 90 percent of the overconfident about the signal. They find scale adoption of the railroad technology wealth share fluctuations are attributable that, because irrational traders introduce by the late 1850s. to movements in net payout and 10 per- an additional source of risk, rational Pan and Singleton explore in depth cent to changes in expected returns. The investors reduce the proportion of the nature of the risk-neutral credit-event evidence shows that sectoral wealth and wealth invested into equity, except when intensities (λQ) that best describe the asset returns are not related — this leads they are extremely optimistic about term structures of sovereign CD spreads. to the equity capital puzzle. future growth. Moreover, their optimal They examine three distinct families of Through the flow-of-funds identity portfolio strategy is based not just on a stochastic processes: the square-root, log- and the capital accumulation equation, current price divergence but also on a normal, and three-halves processes. Their Larrain and Yo develop a present-value model of irrational behavior and a pre- models use different specifications of model that relates the market value of diction concerning the speed of conver- mean reversions and time-varying volatil- corporate assets to its expected future gence. Thus, the portfolio strategy ities to fit both the distributions of cash flow. The relevant measure of cash includes a protection in case there is a spreads and the variation over time in the flow is net payout, which is the sum of deviation from that prediction. The shapes of the term structures of spreads. dividends, interest, and net equity and authors find that long maturity bonds are They find that the models imply highly debt repurchases. A variance decomposi- an essential accompaniment of equity persistent λQ that are strongly correlated tion of the ratio of net payout to assets investment, as they serve to hedge this with measures of global credit event risks shows that 12 percent of its variation is “sentiment risk.” Even though rational and the VIX index of option-implied explained by asset returns, while 88 per- investors find it beneficial to trade on volatilities. Moreover, the correlations cent is explained by cash flow growth. their belief that the market is excessively across countries of the model-implied The constant discount rate present-value volatile, the answer to the question posed credit-event intensities are large, and model is adequate for valuing corporate in the title is: “There is little that rational change with credit-market conditions. assets, in contrast to its failure for valuing investors can do optimally to exploit, and There are substantial model-implied risk equity. hence eliminate, excessive volatility, premiums associated with unpredictable The prevailing view in finance is that except in the very long run.” future variation in λQ. The authors show the evidence for being able to predict During technological revolutions, that the term structure of CD spreads long-horizon stock returns is significant- stock prices of innovative firms tend to allows them to separately identify both ly stronger than for short-horizon exhibit high volatility and bubble-like pat- the loss rate in the event of default, LQ, returns. Boudoukh, Richardson, and terns, which are often attributed to and the parameters of the process, λQ. Whitelaw show that, for all practical pur- investor irrationality. Pastor and Unconstrained estimates of LQ are much poses, the estimators are almost perfectly Veronesi develop a general equilibrium lower than the values typically assumed in correlated across horizons under the null model that rationalizes the observed the financial industry. Finally, to shed hypothesis of no predictability. For price patterns. High volatility is a result of light on the economic consequences of example, for the persistence levels of div- high uncertainty about the average pro- differing levels of LQ or persistence in λQ, idend yields, the analytical correlation is ductivity of a new technology. Investors the authors explore the sensitivity of the 99 percent between the 1- and 2-year learn about this productivity before prices of options on CD contracts to horizon estimators and 94 percent deciding whether to adopt the technolo- alternative settings of the parameters between the 1- and 5-year horizons, gy on a large scale. For technologies that governing the default process. because of the combined effects of over- ultimately are adopted, the nature of In almost any equilibrium model, lapping returns and persistence of the uncertainty changes from idiosyncratic to shifts in sectoral wealth have direct impli- predictive variable. Common sampling systematic as the adoption becomes cations for asset returns, inducing error across equations leads to OLS coef- more likely; as a result, stock prices fall investors to hold more or less of their ficient estimates and R-squares that are after an initial run-up. This “bubble” in wealth in the sector. For an expanding roughly proportional to the horizon

42 NBER Reporter Winter 2005/6 ------under the null of no predictability. This is amine the performance of these vari- based on the historical average stock the precise pattern found in the data. ables, both in-sample and out-of-sample. return. In their paper, Campbell and They corroborate the asymptotic theory They find that most variables would not Thompson show that forecasting vari- and extend the analysis using extensive have helped an investor outpredict the ables with significant forecasting power simulation evidence. The authors then historical equity premium mean. Most in-sample generally have a better out-of- perform joint tests across horizons for a would have hurt outright. None deserves sample performance than a forecast variety of explanatory variables, and an unqualified endorsement. based on the historical average return, there is little or no evidence of pre- A number of variables are correlated once sensible restrictions are imposed on dictability in the data. with subsequent returns on the aggregate the signs of coefficients and return fore- Economists have suggested a whole U.S. stock market in the twentieth centu- casts. The out-of-sample predictive range of variables that predict the equity ry. Some of these variables are stock power is small, but they find that it is eco- premium: dividend price ratios, dividend market valuation ratios, others reflect pat- nomically meaningful. They also show yields, earnings-price ratios, dividend pay- terns in corporate finance or the levels of that a variable is quite likely to have poor out ratios, corporate or net issuing ratios, short- and long-term interest rates. Goyal out-of-sample performance for an book-market ratios, beta premia, interest and Welch (2004) have argued that in- extended period of time even when the rates (in various guises), and consump- sample correlations conceal a systematic variable genuinely predicts returns with a tion-based macroeconomic ratios (cay). failure of these variables out of sample: stable coefficient. Goyal and Welch comprehensively reex- none are able to beat a simple forecast

Corporate Finance

The NBER’s Program on the Dynamics of Corporate Owner- Massimo Massa, INSEAD, and Corporate Finance met at the ship” Andrei Simonov, Stockholm School Harvard Business School on Discussant: Randall Morck, NBER and of Economics, “Shareholder Homo- November 11. Heitor Almeida and University of Alberta geneity and Firm Value: The Disciplin- Daniel Wolfenzon, NBER and New ing Role of Non-Controlling Share- York University Stern School of Amir Sufi, University of Chicago and holders” Business, organized this program: NBER, “Bank Lines of Credit in Discussant: Martijn Cremers, Yale Corporate Finance: An Empirical University Viral Acharya, London Business Analysis” School, and Rangarajan Sundaram Discussant: Murillo Campello, Alexander Dyck, University of and Kose John, New York University, University of Illinois Toronto; Natalya Volchkova, “Cross-Country Variations in Capital Russian-European Center for Eco- Structures: The Role of Bankruptcy Marianne Bertrand, University of nomic Policy; and Luigi Zingales, Codes” Chicago and NBER, Francis Kramarz, Harvard University and NBER, “The Discussant: Matias Braun, University of CREST-ENSAE; Antoinette Schoar, Corporate Governance Role of the California, Los Angeles MIT and NBER; and David Thesmar, Media: Evidence from Russia” HEC, “Politically Connected CEOs Discussant: Stefano Dellavigna, UC, Jean Helwege, University of Arizona; and Corporate Outcomes: Evidence Berkeley Christo Pirinsky, Texas A&M Univer- from France” sity; and Rene Stulz, Ohio State Discussant: Mara Faccio, Vanderbilt University and NBER, “Why Do Firms University Become Widely Held? An Analysis of

Acharya, Sundaram, and John choices. They develop a theoretical code in determining optimal capital investigate the impact of bankruptcy model to identify how firm character- structures. A novel and sharp empirical codes on firms’ capital-structure istics may interact with the bankruptcy implication emerges from this model:

------NBER Reporter Winter 2005/6 43 the difference in leverage choices under among borrowers with access to a bank argue that the degree of shareholder a relatively equity-friendly bankruptcy line of credit, banks use strict covenants homogeneity affects firm value. code (such as the U.S. code) and one that on profitability, and the borrower loses Homogeneous shareholders act as a dis- is relatively more debt-friendly (such as access to the unused portion of the line ciplining device on managers, inducing the U.K. code) should be a decreasing of credit when it experiences a drop in higher profitability, higher stock price, function of the anticipated liquidation profitability. These findings identify a lower volatility and higher transparency. value of the firm’s assets. Using a large specific constraint (the inability to obtain Shareholder homogeneity represents an database of U.S. and U.K. firms over the a line of credit) that causes low prof- alternative and indirect source of corpo- period 1990 to 2002, they subject this itability firms to hold larger cash balances rate governance based on the stock mar- prediction to extensive empirical testing, in their liquidity management strategies. ket. The authors test this hypothesis by both parametric and non-parametric, A number of papers have docu- using a dataset containing information on using different proxies for liquidation val- mented that political leaders may use all the shareholders for each firm in ues and different measures of leverage. their power to grant favors to connected Sweden from 1995 to 2001. They con- They find strong support for the theory; private firms. In this paper, Bertrand, struct two novel proxies for shareholder that is, proxies for liquidation value are Kramarz, Schoar, and Thesmar investi- homogeneity: the first is based on the age both statistically and economically signif- gate the reverse perspective: they ask cohort of the shareholders, and the sec- icant in explaining leverage differences whether politically connected business ond on their degree of college interac- across the two countries. On the other leaders alter corporate decisions to tion. For each firm, they measure the hand, many of the other factors that are bestow “re-election favors” onto incum- degree of homogeneity of all sharehold- known to affect within-country leverage bent politicians. They study this question ers. Using this proxy, they show that (such as size) cannot explain across- in the context of France, where they doc- greater homogeneity increases firm prof- countries differences in leverage. ument a large overlap in educational and itability and returns, and reduces analyst Helwege, Pirinsky, and Stulz con- professional background between the error, analyst dispersion, and stock sider IPO firms from 1970 to 2001 and CEOs of publicly-traded firms and volatility. examine the evolution of their insider politicians: more than half of the assets Dyck, Volchkova, and Zingales ownership over time to understand better traded on the French stock markets are study the effect of media coverage on why and how U.S. firms that become managed by CEOS who were formerly corporate governance outcomes by widely held did so. In their sample, a civil servants. Overall, the results support focusing on Russia in the period 1999— majority of firms has insider ownership the hypothesis that connections between 2002. Russia provides multiple examples below 20 percent after ten years. The CEOs and politicians factor into corpo- of corporate governance abuses, where authors find that a firm’s stock market rate decisions on job creation and traditional corporate governance mecha- performance and trading play an destruction. Firms managed by connect- nisms are ineffective, and where they can extremely important role in its insider ed CEOs create more jobs (and destroy identify an exogenous source of news ownership dynamics. Firms that experi- fewer plants) in politically more contest- coverage arising from the presence of an ence large decreases in insider ownership ed areas, and especially around election investment fund, the Hermitage fund, and/or become widely held have high years. The authors find weak evidence that tried to shame companies by expos- valuations, good recent stock market per- that these networks between politicians ing their abuses in the international formance, and liquid markets for their and business executives follow partisan media. The authors find that the proba- stocks. In contrast and surprisingly, vari- lines. In return, “favors” extended by bility that a corporate governance abuse ables suggested by agency theory have connected CEOs to politicians seem to is reversed is affected by the coverage of limited success in explaining the evolu- be reciprocated through privileged access the news in the Anglo-American press. tion of insider ownership. to subsidy programs and lower taxes. The result is not attributable to the endo- Sufi uses novel data collected from Finally, the authors show that firms man- geneity of news reporting, since this annual 10-K SEC filings to conduct the aged by politically connected CEOs have result holds even when they instrument first large sample empirical examination lower performance than non-connected media coverage with the presence of the of the use of bank lines of credit by pub- firms, suggesting that political connec- Hermitage fund among its shareholders lic corporations. He finds that the supply tions might impose a cost on the firms. and the “natural” newsworthiness of the of lines of credit by banks to corporate Massa and Simonov study how the company involved. They confirm this borrowers is particularly sensitive to the shareholding structure of a firm affects evidence with a case study. borrowers’ historical profitability. Even its stock price and profitability. They

44 NBER Reporter Winter 2005/6 ------Behavioral Economics

The NBER’s Working Group on Discussant: Bhaskaran Swaminathan, Harrison Hong, Princeton University, Behavioral Economics, directed by Cornell University and Marcin Kacperczyk, University NBER Research Associates Robert J. of British Columbia, “The Price of Sin: Shiller of Yale University and Richard Markus Brunnermeier, Princeton The Effects of Social Norms on H. Thaler, University of Chicago, met University, and Christian Julliard, Markets” in Cambridge on November 12. The London School of Economics, Discussant: Owen Lamont, Yale following papers were discussed: “Money Illusion and Housing University and NBER Frenzies” Andrei Shleifer and Sendhil Discussant: Christopher J. Mayer, Luigi Guiso, University of Chicago; Mullainathan, Harvard University and Columbia University and NBER Paola Sapienza, Northwestern NBER, “Persuasion in Finance” University and NBER; and Luigi Discussant: Shane Frederick, MIT Yiming Qian and Matthew Billett, Zingales, Harvard University and University of Iowa, “Are Overconfi- NBER, “Trusting the Stock Jeffrey Wurgler, New York University dent Managers Born or Made? Market”(NBER Working Paper No. and NBER, and Malcolm Baker, Evidence of Self-Attribution Bias from 11648) Harvard University and NBER, Frequent Acquirers” Discussant: Joshua Coval, Harvard “Government Bonds and the Cross- Discussant: Ulrike Malmendier, Stan- University Section of Stock Returns” ford University and NBER

Persuasion is a fundamental part of drops in investor sentiment increase the mispricing; 2) the run-ups in housing social activity, yet it is rarely studied by demand for both government bonds and prices starting in the late 1990s can be economists. Mullainathan and Shleifer bond-like stocks. Consistent with both reconciled with the contemporaneous compare the traditional economic the required returns and sentiment chan- reduction in inflation and nominal inter- model, in which persuasion is an effort nels, the authors find a common pre- est rates; and 3) the tilt effect cannot to change the listener’s mind using infor- dictable component in bonds and bond- rationalize these findings. mation, with a behavioral model, in like stocks. Consistent with the senti- Billett and Qian explore the source which persuasion is an effort to fit the ment channel, they find that bonds and of managerial hubris in mergers and message into the audience’s already held bond-like stocks co-move with inflows acquisitions by examining the history of beliefs. They present a simple formaliza- into government bond and conservative deals made by individual acquirers. Their tion of the behavioral model, and com- stock mutual funds. study has three main findings: 1) compare the two models using data on finan- A reduction in inflation can fuel pared to their first deals, acquirers of cial advertising in Money and Business run-ups in housing prices if people suf- second and higher-order deals experi- Week magazines over the course the fer from money illusion. For example, ence significantly more negative internet bubble. The evidence on the basing the decision on whether to rent or announcement effects; 2) while acquisi- content of persuasive messages is broad- buy a house simply on monthly rent rel- tion likelihood increases in the perform- ly consistent with the behavioral model ative to current monthly mortgage pay- ance associated with previous acquisi- of persuasion. ments, agents do not properly take into tions, previous positive performance Baker and Wurgler document that account that inflation lowers future real does not curb the negative wealth effects U.S. government bonds co-move more mortgage payments, therefore systemati- associated with future deals; 3) top man- strongly with bond-like stocks: stocks of cally mis-evaluating real estate. After agement’s net purchase of stock is large, mature, low-volatility, profitable, empirically decomposing the price-rent greater preceding high order deals than it dividend-paying firms that are neither ratio into a rational component and an is for first deals. The authors interpret high growth nor distressed. This pattern implied mispricing, Brunnermeier and these results as consistent with self-attri- may be caused by common shocks to Julliard find that: 1) inflation and the bution bias leading to managerial over- real cash flows, rationally required nominal interest rate explain a large confidence. They also find evidence that returns, or flights to quality in which share of the time-series variation of the the market anticipates future deals based

------NBER Reporter Winter 2005/6 45 on an acquirer’s acquisition history and with them facing greater litigation risk factor in the risk of being cheated. The impounds such anticipation into stock and/or being neglected because of social perception of this risk is a function not prices. norms, they outperform the market even only of the objective characteristics of Hong and Kacperczyk provide after accounting for well-known return the stock, but also of the subjective evidence for the effects of social norms predictors. Corporate financing deci- characteristics of the investor. Less trust- on markets by studying “sin” stocks — sions and time-variation in norms for ing individuals are less likely to buy stock publicly-traded companies involved in tobacco also indicate that norms affect and, conditional on buying stock, they producing alcohol, tobacco, and gaming. stock prices. Finally, the authors gauge will buy less stock. The calibration of the They hypothesize that there is a societal the relative importance of litigation risk model shows that this problem is suffi- norm to not fund operations that pro- versus neglect for returns. Sin stock ciently severe to account for the lack of mote vice and that some investors, par- returns are not systematically related to participation of some of the richest ticularly institutions subject to norms, various proxies for litigation risk, but are investors in the United States as well as pay a financial cost in abstaining from weakly correlated with the demand for for differences in the rate of participa- these stocks. Consistent with this socially responsible investing, consistent tion across countries. The authors also hypothesis, sin stocks are less held by with them being neglected. find evidence consistent with these certain institutions, such as pension Guiso, Sapienza, and Zingales propositions in Dutch and Italian micro- plans (but not by mutual funds who are provide a new explanation for the limit- data, as well as in cross-country data. natural arbitrageurs), and less followed ed stock market participation puzzle. In by analysts than other stocks. Consistent deciding whether to buy stocks, investors

Higher Education

The NBER’s Working Group on University and NBER Discussant: Ronald G. Ehrenberg, Higher Education met in Cambridge Cornell University and NBER on November 17. Director Charles T. Devin Pope, University of Clotfelter, NBER and Duke California, Berkeley, and Jaren Pope, Scott Carrell, Dartmouth College, University, organized this program: North Carolina State University, and Frederick V. Malmstrom and “Understanding College Choice James E. West, U.S. Air Force Jesse Rothstein, Princeton Univer- Decisions: How Sports Success Academy, “Peer Effects in Academic sity and NBER, and Albert Yoon, Garners Attention and Provides Cheating” Northwestern University, “Mismatch Information” Discussant: David Zimmerman, in Law School” Discussant: Sarah Turner, University Williams College and NBER Discussant: Thomas J. Kane, Harvard of Virginia and NBER University and NBER Ofer Malamud, University of John J. Siegfried and T. Aldrich Chicago, “Breadth vs. Depth: The Christopher Cornwell and David Finegan, Vanderbilt University, and Effect of Academic Specialization on Mustard, University of Georgia, Wendy Stock, Montana State Unive- Labor Market Outcomes” “Merit Aid and Sorting: The Effects rsity, “Time-to-Degree for the Eco- Discussant: Bruce Sacerdote, of HOPE-Style Scholarships on nomics PhD Class of 2001–02” and Dartmouth College and NBER College Ability Stratification” “Attrition in Economics Ph.D. Pro- Discussant: Susan Dynarski, Harvard grams”

According to the “mismatch” where they are unable to compete with students. Students attending more hypothesis, affirmative action prefer- their more qualified white classmates. selective law schools earn substantially ences in admissions induce minority Rothstein and Yoon implement two lower grades than similarly-qualified students to attend selective schools tests of mismatch using data on law students at less selective schools, but

46 NBER Reporter Winter 2005/6 ------are no less likely to graduate or pass the measures of student quality and the many students may be more familiar bar exam, and obtain better jobs at homogeneity of students by ability. The with a school’s recent sports record higher salaries. The authors also com- lowest-quality institutions experience than its academic quality. Pope and pare black students to whites. In the no statistically significant effect from Pope develop a simple model of school upper four quintiles of the LSAT- HOPE on any measure of student qual- choice that incorporates the limited undergraduate GPA distribution, blacks ity. The authors conclude that state- awareness that high school students and whites graduate and pass the bar sponsored merit aid programs increase may have regarding the utility of attend- exam at similar rates, though blacks the retention of high ability students for ing different colleges. Their model pre- attend more selective schools and earn college and also increase the ability dicts that college sports success may lower grades; blacks also obtain better stratification of institutions within increase a school’s future applications post-law-school jobs. In the bottom states. They also examine two indirect both by making students more aware of quintile, black bar passage rates are measures of student selectivity: accept- that college and by increasing the utility lower. However, this cannot confidently ance and yield rates. HOPE decreases associated with attending that school. be attributed to mismatch, as many acceptance rates at all types of institu- Using an administrative dataset that more whites than blacks are unable to tions, but the percentage change is records where high school students sent gain admission to law school, introduc- largest at the universities, which are their SAT scores, the authors analyze ing the potential for sample selection most space constrained. HOPE increas- the effect of sports success on sent test bias. es yield rates for universities but not for scores for all 332 schools that partici- In the last 15 years there has been a any other institution categories. Put pate in NCAA Division I basketball or significant increase in merit aid. Since together, these results suggest that football. They show that sent test scores the early 1990s nearly 30 state-spon- HOPE substantially increased the selec- act as a reasonable proxy for sent appli- sored merit programs have been start- tivity at universities. In addition to cations. Their results indicate that ed, about half of which are based large- Georgia, five other states (Arkansas, sports success in a given year can ly on Georgia’s HOPE Scholarship. Florida, Kentucky, Louisiana, Maryland, increase the total number of students Coincident with this increase in merit and South Carolina) in the SREB start- that send their test scores the following aid has been increased attention to sort- ed large-scale merit aid programs during year by up to 10 percent. They also ing in various aspects of life, especially the sample period. The data show that show that certain demographic groups in education. Cornwell and Mustard in general universities in these states (males, blacks, and students who played examine the extent to which merit- experience similar gains in verbal and sports in high school) are significantly based aid exacerbates or ameliorates math SAT scores and the percentage of more influenced by sports success and sorting by ability in higher education. students who graduate in the top 10 that schools can expect changes in sent They use data from Peterson’s Guide to percent of their high school classes. test scores by up to 15–20 percent after Colleges and the Integrated Post-sec- There are two exceptions. Louisiana, a good sports year for these groups. ondary Education Data System which uses a relatively low threshold The authors conclude that the increase (IPEDS) to evaluate this relationship. criterion to qualify for its merit award, in sent test scores stems from both the From these sources they create a large experiences no statistically significant increased exposure/awareness that panel dataset of institutions of higher increase in SAT scores from its merit schools receive because of sports and learning in the Southern Regional program. Florida’s Bright Futures the increased utility that students associ- Education Board (SREB), and test how Scholarship appears to reduce the SAT ate with attending a school with a merit aid affects sorting between and scores of incoming students while strong sports program. within states. Their empirical strategy increasing the fraction of students who Stock and Siegfried use survey treats HOPE as a natural experiment graduated from the top 10 percent of responses from Ph.D. graduates and and contrasts the quality of freshmen at their high school classes. thesis advisors to estimate the time Georgia colleges to their out-of-state Deciding where to apply to college required for the class of 2001–2 to earn counterparts. The difference-in-differ- among the thousands of four-year a degree. Median time to earn the Ph.D. ences estimates show that HOPE schools in the United States is a daunt- is 5.5 years, up from 5.25 years for the increased the quality of entering fresh- ing task for most teenagers. High school class of 1996–7. The time required to men in Georgia institutions relative to students are typically not aware of all of write a dissertation is a little longer than their out-of-state peers. At the highest- the benefits that each school might the time required to complete compre- quality institutions HOPE raises all offer. In fact, observation suggests that hensive examinations and course work.

------NBER Reporter Winter 2005/6 47 Graduates who had their first child drawal. About 15 percent transfer to university, others allow students to while in a Ph.D. program are estimated other economics programs because postpone this choice. Malamud devel- to finish almost one year later than oth- they are dissatisfied with some aspect of ops a model in which individuals, by ers. Those with predominantly fellow- the particular program where they first taking courses in different fields of ship support finished about six months enrolled. study, accumulate field-specific skills faster than those funded predominantly Using self-reported academic honor and receive noisy signals of match qual- by a teaching assistantship, as did those violations from the classes of 1959 ity in these fields. With later specializa- whose dissertation was a set of essays through 2002 at the three major U.S. mil- tion, students have more time to learn rather than a single topic treatise. itary service academies (Air Force, Army, about match quality in each field but Americans who did their undergraduate and Navy), Carrell, Malmstrom, and less time to acquire specific skills once a work at either a Top-50 U.S. liberal arts West measure how peer honesty influ- field is chosen. Malamud derives com- or other U.S. college or university that ences individual cheating behavior. All parative static predictions between does not offer a Ph.D. in economics fin- else equal, they find higher levels of peer regimes with early and late specializa- ished faster than their counterparts who cheating result in a substantially tion, and tests these predictions across earned a bachelor’s degree from a U.S. increased probability that an individual British systems of higher education university that offers a Ph.D. in eco- will cheat. They identify through sepa- using university administrative data and nomics. International students from rate estimation procedures an exogenous survey data on 1980 university gradu- predominantly English speaking coun- (contextual or pre-treatment) peer effect ates. He finds that individuals in tries finished faster than other students and an endogenous (during treatment) Scotland, where specialization occurs studying in the United States on tempo- peer effect. Results for the (first-order) relatively late, are less likely to switch to rary visas. exogenous peer effect indicate that one an occupation that is unrelated to their Stock, Finegan, and Siegfried use additional high school cheater creates field of study compared to their information about 586 individuals who 0.33 to 0.48 new college cheaters. Results English counterparts who specialize matriculated into two economics Ph.D. for the (first-order) endogenous peer earlier. According to the model, the programs in Fall 2002 to estimate first- effect indicate that one additional college return to being well matched to an and second-year attrition rates. After cheater creates 0.61 to 0.86 new college occupational field is high relative to the two years, 26.5 percent of the initial cheaters. These results imply that, in return to specific skills and there may cohort had left, equally divided between equilibrium, the social multiplier for aca- therefore be benefits to later specializa- the first and second years. Attrition demic cheating ranges between 2.56 and tion. Malamud also finds strong evi- varies widely across individual pro- 3.97. dence in support of the prediction that grams. It is lower among the most high- Malamud examines the tradeoff individuals who switch to unrelated ly rated 15 programs, for students with between early and late specialization in occupations earn lower wages but no higher verbal and quantitative GRE the context of higher education. While evidence that the cost of switching dif- scores, and for those on a research some educational systems require stu- fers between those specializing early assistantship. Poor academic perform- dents to specialize early by choosing a and late. ance is the most cited reason for with- major field of study prior to entering

48 NBER Reporter Winter 2005/6 ------Education Program Meeting

The NBER’s Education Program Stephanie Riegg Cellini, University Time Discounting, and Human Capital met in Cambridge on November 18. of California, Los Angeles, “Funding Investment Decisions” Program Director Caroline M. Hoxby, Schools or Financing Students: Public NBER and Harvard University, chose Subsidies and the Market for Two- Nora Gordon, University of the following papers for discussion: Year College Education” California, San Diego and NBER; Elizabeth Cascio, University of David N. Figlio, University of Joshua Angrist, MIT and NBER, California, Davis and NBER; Sarah Florida and NBER, “Why Barbie Says and Aimee Chin, University of Reber, University of California, Los ‘Math is Hard’” Houston; and Ricardo Godoy, Angeles; and Ethan Lewis, Federal Brandeis University, “Is Spanish-Only Reserve Bank of Philadelphia, “ Margaret Ledyard, University of Schooling Responsible for the Puerto- Financial Incentives and the Deseg- Texas at Austin, “ Why are Private Rican Language Gap?” regation of Southern Public Schools” Schools Small? School Location, Returns to Scale, and Size” Lex Borghans and Bart Golsteyn University of Maastrich, “Imagination,

Figlio adopts a novel approach to ed to him by a large Florida school dis- school choice through private school discerning one pathway through which trict. He then relates the name that a vouchers could lead to larger private family and cultural expectations and given high-achieving sister has to her high schools. This would be a demand resultant identity-formation could influ- propensity to take calculus and physics effect — private schools are smaller ence young women’s choices about in high school, as compared with her because fewer people want to go to studies and potential future careers. He high-achieving sisters with different them. Using cost estimates for public posits that a girl with a more feminine names. Parents often give pairs of sis- schools from Ledyard (2004), Ledyard name may be treated systematically dif- ters very different names in terms of can account for approximately 70 per- ferently by parents, teachers, and peers, their femininity, offering the opportuni- cent of the difference in size between or may herself relate to more feminine ty to directly test the presumption that a public and Catholic high schools. She stereotypes. In such a circumstance, name can have causal influences on a holds costs fixed, and uses data on pri- girls with more feminine names may be girl’s academic development. Figlio vate school location, size, and affiliation more likely to select coursework that is finds that girls with more feminine to predict the size of Catholic schools. more “traditionally female”— such as names are less likely to self-select into Both public and private two-year the humanities and foreign languages advanced mathematics and science colleges rely on public subsidies to — and shy away from coursework that classes in high school, holding constant make their education affordable for stu- is more “traditionally male” — such as family fixed effects and prior achieve- dents. Public community colleges advanced math and science. Of course, ment. These results suggest that envi- receive government support directly in names are not exogenously given to ronmental factors play a large role in the form of subsidies, while private for- girls. Parents often pay great attention determining whether women choose profit colleges or proprietary schools to the names they give their children, mathematics and science as potential receive government support indirectly and parents with different proclivities careers. in the form of grants or vouchers given toward mathematics and science, say, Private schools are, on average, a to students. Cellini analyzes the impact may systematically select different third of the size of public schools. But of these two funding schemes on the names for their daughters. In order to why are they small? Two possible expla- entry decisions of proprietary schools avoid confounding unmeasured family- nations are differences in demand and and enrollments in community colleges. specific factors with causal effects of differences in production. If the returns She uses a new administrative dataset of names, Figlio uses a unique dataset of to scale are similar for private and pub- for-profit colleges in California, panel pairs of highly-achieving sisters provid- lic high schools, then an increase in data methods, and a unique regression

------NBER Reporter Winter 2005/6 49 discontinuity design. She finds that an best way to raise English proficiency. It ical results support the model, so the increase in public funding for a local suggests that increased emphasis on main conclusion is that a lack of imagi- community college diverts students using English as the language of nation induces students to stay longer in from the private to the public sector instruction may do little to benefit education while it reduces the efficiency and causes a corresponding decline in Puerto Ricans who remain on the island of this investment. the number of proprietary schools in today. Cascio, Gordon, Lewis, and the county. Raising student financial aid While economic theory regards Reber examine whether the financial awards, on the other hand, expands the education as an important investment, incentives put in place by two pieces of overall pool of sub-baccalaureate stu- the reality of students’ behavior does federal legislation — the Civil Rights dents and causes proprietary schools to not always seem to support this view. Act of 1964 and the Elementary and enter the market. This effect is particu- Borghans and Golsteyn aim to analyze Secondary Education Act of 1965 — larly strong in counties with high pover- the behavior of students at college from played a causal role in desegregating ty rates where more students are eligible an investment perspective. They pro- southern schools. The latter targeted a for aid. vide robust paradoxical findings that large federal education program toward Between 1898 and 1948, English college students with higher discount the South, while the former tied the was the language of instruction for rates stay longer in education. The receipt of funds under this new pro- most post-primary grades in Puerto explanation they pursue is that a higher gram to nondiscrimination. Using a Rican public schools. Since 1949, the discount rate can partly be a conse- newly collected dataset on school language of instruction in all grades has quence of a lack of imagination about desegregation and school finance for been Spanish. Angrist, Chin, and the future work life. If so, the discount the 1960s, the authors find that districts Godoy use this policy change to esti- rate will be very high at moments when with relatively more to lose under feder- mate the effect of English-intensive there are major changes in circum- al funding allocation rules engaged in instruction on the English-language stances, in this case when students go more student desegregation, were more skills of Puerto Ricans. Although naïve from college to work. This provides likely to have desegregated their facul- estimates suggest that English instruc- incentives for students who lack a clear ties, and were more likely to have tion increased English-speaking ability picture about their future work-finding received their federal funding by the fall among Puerto Rican natives, estimates life to stay in education. To test this of 1967. Qualitatively similar results are that allow for education-specific cohort model, the authors measure the crucial found for the fall of 1966. These results trends show no effect. This result is sur- individual attributes, ask students about suggest that legislative and executive prising in light of the strong presump- the way they made their choices, and enforcement efforts — not just the tion by American policymakers at the present them other choices that reveal courts — contributed to the desegrega- time that English instruction was the the nature of their behavior. The empir- tion of southern education.

50 NBER Reporter Winter 2005/6 ------Political Economy

The NBER’s Working Group on Allan Drazen, University of Maryland UC, Los Angeles, “Inequality and the Political Economy, directed by NBER and NBER, and Marcela Eslava, Evolution of Institutions of Taxation: Research Associate Alberto F. Alesina Universidad de los Andes, “Pork Barrel Evidence from the Economic History of Harvard University, met in Cycles” of the Americas” Cambridge on November 19. The fol- Discussant: Alessandro Lizzeri, New Discussant: William Easterly, New lowing papers were discussed: York University York University

Timothy Besley, London School of John N. Friedman and Richard T. Per Pettersson-Lidbom, Stockholm Economics; Torsten Persson, Stock- Holden, Harvard University, “Optimal University, and Matz Dahlberg, holm University and NBER; and Gerrymandering” Uppsala University, “An Empirical Daniel Sturm, University of Munich, Discussant: Roland G. Fryer, Harvard Approach for Estimating the Causal “Political Competition and Economic University Effect of Soft Budget Constraints on Performance: Theory and Evidence Economic Outcomes” from the United States” Kenneth L. Sokoloff, UC, Los Discussant: Antonio Merlo, University Discussant: Roberto Perotti, Universita Angeles and NBER, and Eric M. Zolt, of Pennsylvania Bocconi and NBER

Besley, Persson, and Sturm for- that voter prefers. The authors show the solution to a problem in which the ger- mulate a model to explain why the lack existence of a political equilibrium in rymanderer observes a noisy signal of of political competition may stifle eco- which rational voters support an incum- voter preferences from a continuous dis- nomic performance; they use the United bent who targets them with spending tribution and creates N districts of equal States as a testing ground for the model’s before the election even though they size to maximize the expected number of predictions, exploiting the 1965 Voting know it may be electorally motivated. In districts that she wins. They show that Rights Act which helped to break the equilibrium, voters in the more “swing” “throwing away” districts is not general- near monopoly on political power of the regions are targeted at the expense of ly optimal, nor is “smoothing.” The opti- Democrats in southern states. They find types of spending not favored by these mal solution involves creating a district that changes in political competition voters. This will be true even if they that matches extreme “Republicans” have quantitatively important effects on know they live in swing regions. with extreme “Democrats,” then contin- state income growth, state policies, and However, the responsiveness of these uing to match toward the center of the quality of Governors. By their bottom- voters to electoral manipulation depends signal distribution. The value to being line estimate, the increase in political on whether they face some degree of the gerrymanderer increases with the competition triggered by the Voting uncertainty about the electoral impor- extremity of voter preferences, the qual- Rights Act raised long-run per capita tance of the group they are in. Use of ity of the signal, and the number of dis- income in the average affected state by targeted spending also implies voters can tricts. about 20 percent. be influenced without election-year Sokoloff and Zolt turn to history to Drazen and Eslava present a model deficits, consistent with recent finding gain a better perspective on how and of political budget cycles in which for established democracies. why tax systems vary. They focus on the incumbents influence voters by targeting Standard intuitions for optimal ger- societies of the Americas over the nine- government spending to specific groups rymandering involve concentrating one’s teenth and twentieth centuries, for two of voters at the expense of other voters extreme opponents in “unwinnable” dis- major reasons. First, despite the region or other expenditures. Each voter faces a tricts (“throwing away”) and spreading having the most extreme inequality in the signal extraction problem: being targeted one’s supporters evenly over “winnable” world, the tax structures of Latin with expenditure before the election may districts (“smoothing”). These intuitions America are generally recognized as reflect opportunistic manipulation, but are not robust and depend crucially on among the most regressive, even by may also reflect a sincere preference of arbitrary modeling assumptions. developing country standards. Second, as the incumbent for the types of spending Friedman and Holden characterize the has come to be widely appreciated, the

------NBER Reporter Winter 2005/6 51 colonization and development of the ments in North America, accounted for problem of the soft budget constraint is Americas constitute a natural experiment a radical divergence in the overall struc- a problem of credibility; that is, inability of sorts that students of economic and ture of taxation. Tapping these progres- of a supporting organization to commit social development can exploit. sive sources of government revenue, itself not to extend more resources to a Beginning more than 500 years ago, a state and local governments in the budget-constrained organization (in small number of European countries United States and Canada, even before other words, bailouts) ex post than it was established colonies in diverse environ- independence, began directing substan- prepared to provide ex ante. This means ments across the hemisphere. The differ- tial resources toward public schools, that current economic behavior of a ent circumstances meant that largely improvements in infrastructure involv- budget-constrained organization will exogenous differences existed across ing transportation and health, and other depend upon its expectations of being these societies, not only in national her- social programs. In contrast, the societies bailed out in the future. Thus, to estimate itage, but also in the extent of inequality. of Latin America, which had come to be the causal effect of soft budget con- The principal concern in this paper is characterized soon after initial settlement straints (that is, bailout expectations) on with how the extent of inequality may by rather extreme inequality in wealth, economic outcomes, one has to measure influence the design and implementation human capital, and political influence, these expectations and link them to the of tax systems. Several salient patterns tended to adopt tax structures that were current behavior of the budget-con- emerge. The United States and Canada significantly less progressive in incidence strained organization. The authors argue (like Britain, France, Germany, and even and manifested greater reluctance or that one can use information about real- Spain) were much more inclined to tax inability to impose local taxes to fund ized bailouts to construct credible meas- wealth and income during their early local public investments and services. ures of bailout expectations. They apply stages of growth — and into the twenti- These patterns persisted, moreover, well an empirical framework to Swedish local eth century — than developing countries into the twentieth century – indeed up to governments, which provide an attrac- are today. Although the U.S. and the present day. The apparent association tive testing ground for the soft budget Canadian federal governments were sim- between initial inequality and the institu- constraint since the central government ilar to those of their counterparts in tions of taxation and public finance is all has extended a total of 1,697 bailouts Latin America in relying primarily on the the more intriguing in that Sokoloff and over the period 1974 to 1992. The taxation of foreign trade (overwhelming- Zolt find corresponding patterns across authors find that bailout expectations ly tariffs) and excise taxes, the greater different regions of the United States have a causal effect on economic behav- success or inclination of state (provin- and across different countries of Latin ior. The estimated effect is quite sizeable: cial) and local governments in North America. on average, a local government increases America to tax wealth (primarily in the Pettersson-Lidbom and Dahlberg its debt by 30 percent if it is certain of form of property or estate taxes) and develop an empirical framework for esti- being bailed versus when it is certain of income (primarily in the form of busi- mating the causal effect of soft budget not being bailed out. ness taxes), as well as the much larger rel- constraints on economic outcomes. ative sizes of these sub-national govern- Their point of departure is that the

52 NBER Reporter Winter 2005/6 ------International Trade and Investment

The NBER’s Program on Pol Antras, Mihir A. Desai, and C. Diego Puga and Daniel Trefler, International Trade and Investment Fritz Foley, Harvard University and University of Toronto and NBER, met at the Bureau’s California office NBER, “FDI Flows and Multi- “Wake Up and Smell the Ginseng: on December 2 and 3. Program national Firm Activity” The Rise of Incremental Innovation Director Robert C. Feenstra of in Low-Wage Countries” University of California, Davis, Lee J. Branstetter and Raymond organized the meeting. The following Fisman, Columbia University and Svetlana Demidova, Pennsylvania papers were discussed: NBER; C. Fritz Foley; and Kamal State University; Hiau Looi Kee, The Saggi, Southern Methodist Univer- World Bank; and Kala Krishna, Carsten Eckel, University of sity, “Intellectual Property Rights, Pennsylvania State University and Goettingen, and J. Peter Neary, Imitation, and Foreign Direct NBER, “Rules of Origin and Firm University College Dublin, “Multi- Investment: Theory and Evidence” Heterogeneity” Product Firms and Flexible Manufacturing in the Global Andrew B. Bernard, Dartmouth Christian Broda, University of Economy” College and NBER; J. Bradford Chicago; Nuno Limão, University of Jensen, Institute for International Maryland; and David E. Weinstein, Volker Nocke, University of Economics; and Peter Schott, Yale Columbia University and NBER, Pennsylvania, and Stephen Yeaple, University and NBER, “Transfer “Optimal Tariffs: The Evidence” University of Pennsylvania and Pricing by U.S.-Based Multinational NBER, “Endogenizing Firm Scope: Firms” Multiproduct Firms in International Trade”

Eckel and Neary present a new number of its product lines. The more al environments? As developers of model of multi-product firms (MPFs) product lines a firm manages, the high- technologies, MNCs have long been and flexible manufacturing and explore er are its unit costs, but this trade-off is characterized as having comparative its implications in partial and general less severe for firms with greater orga- advantage in monitoring the deploy- equilibrium. International trade integra- nizational capabilities. Paradoxically, ment of their technology. Antras, tion affects the scale and scope of more efficient firms optimally increase Desai, and Foley show that, in a setting MPFs through a competition effect and their scope to such an extent that their of non-contractible monitoring and a demand effect. The authors demon- unit costs are higher than those of less financial frictions, this comparative strate how MPFs adjust in the presence efficient firms. The model thus explains advantage endogenously gives rise to of single-product firms and in hetero- the empirical puzzle that there is a neg- MNC activity and FDI flows. The geneous industries. Their results are in ative relationship between firm size and mechanism generating MNC activity is line with recent empirical evidence and Tobin’s Q. Positive industry shocks — not the risk of technological expropria- suggest that MPFs in conjunction with such as those caused by trade liberaliza- tion by local partners but the demands flexible manufacturing play an impor- tion — induce a merger wave that alters of external funders who require MNC tant role in the impact of international the intra-industry dispersion of participation to ensure value maximiza- trade on product diversity – that is, the observed productivity as high-Q firms tion by local entrepreneurs. The model range of products produced by all buy or sell product lines with low-Q delivers distinctive predictions for the firms. firms. impact of weak institutions on patterns Nocke and Yeaple develop a theo- How are foreign direct investment of MNC activity and FDI flows, with ry of multi-product firms that differ in (FDI) flows and patterns of multina- weak institutional environments limit- their organizational capabilities. In the tional firm (MNC) activity determined ing the scale of multinational firm activ- model, a firm’s unit cost is the endoge- in a world with frictions in financial ity but increasing the share of that nous outcome of its choice of the contracting and variations in institution- activity that is financed by multinational

------NBER Reporter Winter 2005/6 53 parents through FDI flows. In addition strongly on patented intellectual prop- low-wage countries and even across to accounting for distinctions between erty as part of their global business firms within each low-wage country. patterns of MNC activity and FDI strategy. Data tracking industry level They then draw out implications for the flows, the model can help explain sub- value-added in the reforming countries location of production, trade, capital stantial two-way FDI flows between point to an overall expansion of indus- flows, earnings and living standards. countries with high levels of financial trial activity after IPR reform. Finally, Demidova, Kee, and Krishna development and small and unbalanced evidence from highly disaggregated develop a heterogeneous firm model to FDI flows between countries with dif- trade data also suggests that the expan- study the effects of trade policy, trade ferent levels of financial development. sion of multinational activity leads to a preferences, and the rules of origin The main predictions of the model are higher net level of production shifting (ROOs) needed to obtain them. They tested and confirmed using firm-level to developing countries, more than off- apply their model to Bangladeshi gar- data on U.S. outbound FDI. setting any possible decline in the imita- ment exports to the United States and Does the adoption of stronger tive activity of indigenous firms. European Union. There are differences intellectual property rights (IPR) in Bernard, Jensen, and Schott across products and export destinations developing countries enhance or retard examine how prices set by multination- that make for an interesting natural their industrial development? How does al firms vary across arm’s-length and experiment. These differences generate such a policy shift affect industrial activ- related-party customers. They find that differences in the composition of ity in the developed countries, where arm’s length prices are substantially and exporters and productivity. The authors most innovative activity is concentrat- significantly higher than related party use data on Bangladeshi garment ed? Branstetter, Fisman, Foley, and prices for U.S.-based multinational exporters to construct firm-level total Saggi address these questions both the- exporters. The price difference is large factor productivity (TFP) estimates. oretically and empirically. On the theo- even when comparing the export of the They then test the predictions of the retical side, they develop a North-South same good by the same firm to the model on the relationship between the product cycle model in which Northern same destination country in the same distributions of TFP of various groups innovation, Southern imitation, and month by the same mode of transport. of firms. They show that the facts FDI are all endogenous. This model The price wedge is smaller for com- match the predictions of the model. predicts that IPR reform in the South modities than for differentiated goods The theoretical debate over leads to increased FDI from the North, and is increasing in firm size and firm whether countries can and should set as Northern firms shift production to export share. The difference between tariffs in response to export elasticities Southern affiliates. This increased FDI arm’s length and related party prices is goes back over a century to the writings drives an acceleration of Southern also significantly greater for goods sent of Edgeworth (1894) and Bickerdike industrial development, as the South’s to countries with lower taxes and high- (1907). Despite the optimal tariff argu- share of global manufacturing and the er tariffs. Changes in exchange rates ment’s centrality in debates over com- pace at which production of more have differential effects on arm’s length mercial policy, there exists no evidence recently invented goods shifts to the and related party prices; an appreciation about whether countries actually use it South both increase. The model also of the dollar strongly reduces the dif- in setting tariffs. Broda, Limão, and predicts that as production shifts to the ference between the prices Weinstein estimate disaggregate export South, Northern resources will be real- Increasingly, a small number of elasticities and find that countries that located to R and D, driving an increase low-wage countries such as China and are not members of the World Trade in the global rate of innovation. The India are involved in innovation — not Organization systematically set higher authors confront the theoretical model “big ideas” innovation, but the constant tariffs on goods that are supplied inelas- with evidence on the response of U.S. incremental innovations needed to stay tically. The result is robust to the inclu- multinationals to a series of well-docu- ahead in business. Puga and Trefler sion of political economy variables and mented IPR reforms by developing provide some evidence of this new phe- a variety of model specifications. countries in the 1980s and 1990s. Their nomenon and develop a model in which Moreover, they find that countries with results indicate that U.S.-based MNCs there is a transition from old-style prod- higher aggregate market power have on expand the scale of their activities in uct-cycle trade to trade involving incre- higher average tariffs. In short, there is reforming countries after IPR reform, mental innovation in low-wage coun- strong evidence in favor of the optimal and this effect is disproportionately tries. They explain why levels of tariff argument. strong for affiliates whose parents rely involvement in innovation vary across

54 NBER Reporter Winter 2005/6 ------Productivity Program Meeting

The NBER’s Program on Evidence from Professional Tran- Discussant: Margaret Kyle, Duke Productivity met in Cambridge on sitions for the Superstars of Medicine” University and NBER December 2. Program Director Ernst Discussant: Manuel Trajtenberg, Tel R. Berndt of MIT and Pierre Azoulay, Aviv University and NBER Nick Bloom, Stanford University, and NBER and Columbia University, John Van Reenan, London School of organized the meeting. The agenda Laura Schultz and Sumiye Okubo, Economics, “Measuring and Explain- was: Bureau of Economic Analysis — ing Management Practices Across Report on the BEA/NSF R and D Firms and Countries” Amy Finkelstein and Daron Satellite Accounts: Estimating the Discussant: Richard B. Freeman, Acemoglu, MIT and NBER, “Input Returns and Spillovers from Business R Harvard University and NBER and Technology Choices in Regulated and D Industries: Evidence from the Health James D. Adams, Renssalaer Care Sector” Yonghong Wu, University of Illinois; Polytechnic University and NBER, and Discussant: David M. Cutler, Harvard David Popp, Syracuse University and J. Roger Clemmons, University of University and NBER NBER; and Stuart Bretschneider, Florida, “Industrial Scientific Discov- Syracuse University, “The Effects of ery” Pierre Azoulay, and Joshua Graff Innovation Policies on Business R&D: Discussant: Scott Stern, Northwestern Zivin, Columbia University and A Cross-National Empirical Study” University and NBER NBER, “Peer Effects in the Workplace:

Acemoglu and Finkelstein exam- with PPS. Most interestingly, they find away from collaboration with other sci- ine the implications of regulatory that the PPS reform seems to have entists cancels a significant portion of change for the input mix and technolo- encouraged the adoption of a range of the benefits of exposure to superstar gy choices of regulated industries. They new medical technologies. They also talent. The authors also find that the present a simple neoclassical framework show that the reform was associated location spillovers declined markedly in that emphasizes changes in relative fac- with an increase in the skill composition the 1990s. tor prices faced by regulated firms of these hospitals, which is consistent Wu, Popp, and Bretschneider under different regimes, and investigate with technology-skill or capital-skill examine the effect of three major how this might affect technology choic- complementarities. national innovation policies (patent pro- es through substitution of (capital Azoulay and Zivin estimate the tection, R and D tax incentives, and embodied) technologies for tasks previ- magnitude of knowledge spillovers gen- government funding of business R and ously performed by labor. They empiri- erated by 4,764 academic superstars in D) on business R and D spending. cally examine some of the implications the life sciences onto their coauthors’ Unlike previous work, their study con- of the framework by studying the research productivity. Using matched siders the effect of openness to interna- change from full cost to partial cost employee-employer data, the authors tional trade. They use data from nine reimbursement under the Medicare measure how scientic output (grants, OECD countries (Australia, Canada, Prospective Payment System (PPS) publications, and patents) for a coau- France, Germany, Italy, Japan, Spain, reform, which increased the relative thor changes when the superstar moves United Kingdom, and United States) in price of labor faced by U.S. hospitals. to or from a different institution. 1985–95. Their results show that all Using the interaction of hospitals’ pre- Preliminary results indicate that super- three innovation policies play a signifi- PPS Medicare share of patient days stars generate substantial spillovers cant role in stimulating business funded with the introduction of these regulato- through two independent channels: and performed R and D. Among the ry changes, they document a substantial location and co-authorship. Location components of patent rights, enforce- increase in capital-labor ratios and a spillovers decline more than linearly ment of patent legal regime and dura- large decline in labor inputs associated with geographic distance. Substitution tion of protection term consistently

------NBER Reporter Winter 2005/6 55 have a positive effect on business R and while French and British firms also make significant contributions to scien- D decisions. In addition, R and D per- report substantially higher levels of tific discovery. The authors also uncov- formed by the government has a posi- primo geniture because of the influence of er a host of potential biases. First, the tive effect on business R and D, while R Norman legal origin. These two factors response of discovery to the firms’ own and D by the higher education sector explain up to two thirds of the average R and D is biased upward by the failure has a negative impact on business R and U.S.-Europe management gap. to include science spillovers from uni- D. The authors also find modest empir- Adams and Clemmons estimate versities and other firms. Second, the ical support for the positive role of science production functions for top R university citation spillover is biased openness to international trade in busi- and D firms in the United States. Their upward by the failure to include collab- ness R and D investment. data include estimated flows of basic oration between firms and universities. Bloom and Reenen use an innova- science from universities to firms, from Third, the effects of spillovers and spill- tive survey tool to collect management firms to other firms, and within firms. backs are biased downward when practice data from 732 medium-sized The underlying evidence consists of zeroes of the spillovers and spillbacks manufacturing firms in the United papers and citations from the Institute are not considered by the estimation States, France, Germany, and the for Scientific Information (ISI) in procedure. The elasticity of firms’ sci- United Kingdom. These measures of Philadelphia, Pennsylvania. The data ence output with respect to university managerial practice are strongly associ- cover the top 200 R and D firms and citation spillovers is consistently larger ated with productivity, profitability, the top 110 universities during 1981–99. than the firm spillover elasticity. In Tobin’s Q, sales growth, and survival These account for most U.S. scientific addition, the marginal product of uni- rates. Management practices also dis- research during this period. Their versity spillovers exceeds the marginal play significant cross-country differ- empirical estimates are based on a panel product of firm spillovers, so that addi- ences, with U.S. firms on average better of firms, science fields, and years that is tional science output per dollar of uni- managed than European firms, and sig- an extract from the papers and citations versity R and D is several times larger nificant within-country differences with data. Using this panel, they find that sci- than additional output per dollar of a long tail of extremely badly managed ence spillovers from universities and firm R and D. University collaboration firms. The authors find this is attributa- other firms occur primarily within only serves to increase this productivity ble to: different levels of product mar- fields. Industry is much less of a barrier advantage of universities. Since univer- ket competition, associated with better and, in fact, most knowledge flows sity R and D is primarily funded by gov- management; and family firms passing occur between, rather than within ernment, this potency of university management control down to the eldest industries. Citation and collaboration spillovers appears to reassert the role of sons (primo geniture), associated with spillovers from universities, citation publicly funded science in propagating worse management. European firms spillovers from other firms, and citation knowledge externalities throughout the report lower levels of competition, spillbacks from firms’ past research all U.S. economy.

56 NBER Reporter Winter 2005/6 ------Bureau Books Corruption and Reform: Lessons from America’s Economic History Corruption and Reform: Lessons from contributors explore this shadowy peri- eye over government and business. America’s Economic History, edited by od of U. S. history in search of better These and other facets of American Edward L.Glaeser and Claudia Goldin, methods of fighting corruption world- history analyzed in this volume make it is available from the University of wide today. indispensable as background for anyone Chicago Press. This NBER Conference The chapters in this volume address interested in corruption today. Report may be purchased for $75.00 the measurement and consequences of Glaeser is a Research Associate in from: University of Chicago Press, fraud and corruption and the forces the NBER’s Programs on Aging, Law Order Department, 11030 South that ultimately led to their decline with- and Economics, and Economic Langley Avenue, Chicago, IL 60628- in the United States. They show that Fluctuations and Growth. He is also the 2215; 1-800-621-2736. Academic dis- various approaches to reducing corrup- Fred and Eleanor Glimp Professor of counts are available. tion, such as deregulation and in partic- Economics at Harvard’s Kennedy Despite recent corporate scandals, ular “free banking” in the 1830s, have School of Government. the United States is among the world’s met with success. In the 1930s, corrup- Goldin directs the NBER’s least corrupt nations. But in the nine- tion was kept in check when new feder- Program on the Development of the teenth century, municipal governments al bureaucracies replaced local adminis- American Economy and is the Henry and robber barons alike found new trations in doling out relief. Another Lee Professor of Economics at ways to steal from taxpayers and swin- deterrent to corruption was the inde- Harvard. dle investors. In Corruption and Reform, pendent press, which kept a watchful

The Democratization of Invention: Patents and Copyrights in American Economic Development, 1790-1920 The Democratization of Invention: property rights and the operation of Term Factors in Economic Patents and Copyrights in American efficient markets. When markets Development, include: patent laws and Economic Development, 1790-1920, by B. expanded, these inventors contributed litigation; women inventors in America; Zorina Khan, is available from to the proliferation of new technolo- patentees and married women’s proper- Cambridge University Press for $60.00. gies and improvements. In contrast to ty rights; and intellectual property and An examination of the evolution its leadership in the area of patents, the economic development. and impact of American intellectual U.S. copyright regime was among the Khan is a Faculty Research Fellow property rights during the “long nine- weakest in the world, in keeping with its in the NBER’s Program on the teenth century,” this monograph com- utilitarian objective of promoting the Development of the American pares the American system to that of general welfare. American patent and Economy and a member of the eco- the more oligarchic societies of France copyright institutions promoted a nomics faculty at Bowdoin College. and Britain. The United States created process of democratization that not Orders for the book should be sent the first modern patent system and its only furthered economic and techno- to the Press at: 100 Brook Hill policies toward inventors were the logical progress but also provided a Drive,West Nyack, NY 10994-2133. most liberal in the world. Individuals conduit for the creativity and achieve- Or, by phone: 800-872-7423 (U.S. and who did not have the resources to ments of disadvantaged groups. Canada); 95-800-010-0200 (Mexico); or directly exploit their inventions benefit- The topics discussed in this book, 845-353-7500. ed disproportionately from secure part of the NBER’s series on Long-

------NBER Reporter Winter 2005/6 57 Current Working Papers

NBER Working Papers On-Line A complete list of all NBER Working Papers with searchable abstracts, and the full texts of Working Papers (issued since November 1994) are available at http://www.nber.org/wwp.html to anyone located at a university or other organization that subscribes to the (hard copy) Working Paper series. If you believe that your organization subscribes, but you cannot access the online Working Paper service, please e-mail the NBER at [email protected] for more information and assistance. * Individual copies of NBER Working Papers, Historical Factors in Long-Run Growth Papers, and Technical Papers are available free of charge to Corporate Associates. For all others, there is a charge of $10.00 per hardcopy or $5.00 per downloaded paper. (Outside the United States, add $10.00 per order for postage and handling.) Advance payment is required on all orders. To order, call the Publications Department at (617)868-3900 or visit www.nber.org/papers. Please have ready the number(s) of any Working Paper(s) you wish to order. Subscriptions to the full NBER Working Paper series include all 700 or more papers published each year. Subscriptions are free to Corporate Associates. For others within the United States, the standard rate for a full subscription is $2525; for academic libraries and faculty members, $1475. Higher rates apply for foreign orders. The on-line standard rate for a full subscription is $1750 and the on-line academic rate is $725. Partial Working Paper subscriptions, delineated by program, are also available. For further information, see our Web site, or please write: National Bureau of Economic Research, 1050 Massachusetts Avenue, Cambridge, MA 02138-5398. * Titles of all papers issued since November 2005 are presented below. For previous papers, see past issues of the NBER Reporter. Working Papers are intended to make results of NBER research available to other economists in preliminary form to encourage discussion and suggestions for revision before final publication. They are not reviewed by the Board of Directors of the NBER.

NBER Working Papers Paper Author(s) Title

11752 Yongmin Chen Buyer Investment, Product Variety, and Intrafirm Trade Robert C. Feenstra

11753 Benjamin A. Olken Monitoring Corruption: Evidence from a Field Experiment in Indonesia

11754 Trevon D. Logan The Transformation of Hunger: The Demand for Calories Past and Present

11755 Xavier Gabaix Shrouded Attributes, Consumer Myopia, and David Laibson Information Suppression in Competitive Markets

11756 Clemens Sialm Tax Changes and Asset Pricing: Time-Series Evidence

11757 Itay Goldstein An Information-Based Trade Off between Foreign Assaf Razin Direct Investment and Foreign Portfolio Investment

11758 Bruce N. Lehmann The Role of Beliefs in Inference for Rational Expectations Models

58 NBER Reporter Winter 2005/6 ------Paper Author(s) Title

11759 John F. Helliwell How’s the Job? Well-Being and Social Capital Haifang Huang in the Workplace

11760 Raj Chetty Why do Unemployment Benefits Raise Unemployment Durations? The Role of Borrowing Constraints and Income Effects

11761 Menzie D. Chinn Current Account Balances, Financial Development, Hiro Ito and Institutions: Assaying the World “Savings Glut”

11762 Andrew K. Rose Cities and Countries

11763 Li Gan A Simple Test of Adverse Events and Strategic Tarun Sabarwal Timing Theories of Consumer Bankruptcy

11764 Fernando Alvarez General Equilibrium Analysis of the Eaton-Kortum Robert E. Lucas Model of International Trade

11765 Benjamin Edelman Internet Advertising and the Generalized Second Michael Ostrovsky Price Auction: Selling Billions of Dollars Worth of Keywords Michael Schwarz

11766 Marcin Kacperczyk Unobserved Actions of Mutual Funds Clemens Sialm Lu Zheng

11767 Wojciech Kopczuk To Leave or Not To Leave: The Distribution of Joseph P. Lupton Bequest Motives

11768 Wojciech Kopczuk Electronic Filing, Tax Preparers, and Participation Cristian Pop-Eleches in the Earned Income Tax Credit

11769 Kathryn Dominguez What Defines “News” in Foreign Exchange Markets? Freyan Panthaki

11770 Toshiaki Iizuka Drug Advertising and Health Habits Ginger Zhe Jin

11771 Michael P. Dooley Interest Rates, Exchange Rates, and David Folkerts-Landau International Adjustment Peter M. Garber

11772 Joseph E. Stiglitz The Creation of the Rule of Law and the Karla Hoff Legitimacy of Property: The Political and Economic Consequences of a Corrupt Privatization

------NBER Reporter Winter 2005/6 59 Paper Author(s) Title

11773 Ann P. Bartel How Does Information Technology Really Casey Ichniowski Affect Productivity? Plant-Level Comparisons of Kathryn L. Shaw Product Innovation, Process Improvement, and Worker Skills

11774 Bee Yan Aw The Complementary Role of Exports and R&D Mark J. Roberts Investments as Sources of Productivity Growth Tor Winston

11775 Torben G. Anderson Roughing it Up: Including Jump Components Tim Bollerslev in the Measurement, Modeling, and Forecasting Francis X. Diebold of Return Volatility

11776 Robert J. Gordon A Century of Housing Shelter Prices: Is there Todd vanGoethem a Downfall Bias in the CPI?

11777 Robert J. Gordon What Caused the Decline in U.S. Business Cycle Volatility?

11778 Robert J. Gordon The 1920s and the 1990s in Mutual Reflection

11779 Roberto Chang Financial Crises and Political Crises

11780 A. Mitchell Polinsky The Theory of Public Law Enforcement Steven Shavell

11781 Steven Shavell Liability for Accidents

11782 David Neumark The Effects of Wal-Mart on Local Labor Markets Junfu Zhang Stephen Ciccarella

11783 Farley Grubb The U.S. Constitution and Monetary Powers

11784 Farley Grubb Two Theories of Money Reconciled: The Colonial Puzzle Revisited with New Evidence

11785 Assaf Razin Evaluation of Currency Regimes: The Unique Role Yona Rubinstein of Sudden Stops

11786 Gary Charness Pay Inequality, Pay Secrecy, and Effort: Peter Kuhn Theory and Evidence

11787 Roberto Chang Openness Can be Good for Growth: Linda Kaltani The Role of Policy Complementarities Norman Loayza

60 NBER Reporter Winter 2005/6 ------Paper Author(s) Title

11788 Jordi Gali Robustness of the Estimates of the Mark Gertler Hybrid New Keynesian Phillips Curve J. David Lopez-Salido

11789 Edward L. Glaeser Paternalism and Psychology

11790 Michael Greenstone Does Hazardous Waste Matter? Evidence Justin Gallagher from the Housing Market and the Superfund Program

11791 Amartya Lahiri Output Costs, Currency Crises, and the Carlos A. Vegh Interest Rate Defense of a Peg

11792 Linda S. Goldberg Establishing Credibility: Evolving Perceptions Michael W. Klein of the European Central Bank

11793 David G. Blanchflower An Analysis of the Impact of Affirmative Action Jon Wainwright Programs on Self-Employment in the Construction Industry

11794 Justine S. Hastings Economic Outcomes and the Decision to Vote: Thomas J. Kane The Effect of Randomized School Admission Douglas O. Staiger on Voter Participation Jeffrey M. Weinstein

11795 Natalia Chernyshoff Stuck on Gold: Real Exchange Rate Volatility David S. Jacks and the Rise and Fall of the Gold Standard Alan M. Taylor

11796 Sandra E. Black From the Cradle to the Labor Market? Paul J. Devereux The Effect of Birth Weight on Adult Outcomes Kjell Salvanes

11797 Woojin Kim Motivations for Public Equity Offers: Michael S. Weisbach An International Perspective

11798 Bruce A. Blonigen Foreign Subsidization and the Excess Wesley W. Wilson Capacity Hypothesis

11799 Bruce A. Weinberg Creative Careers: The Life Cycles of David W. Galenson Nobel Laureates in Economics

11800 Dale W. Jorgenson The Industry Origins of Japanese Economic Growth Koji Nomura

11801 Dale W. Jorgenson Information Technology and the Japanese Economy Kazuyuki Motohashi

------NBER Reporter Winter 2005/6 61 Paper Author(s) Title

11802 Patrick Bayer Choice and Competition in Local Education Markets Robert McMillan

11803 Bernard Dumas What Can Rational Investors Do About Excessive Alaxender Kurshev Volatility and Sentiment Fluctuations? Raman Uppal

11804 Irene Brambilla Farm Productivity and Market Structure: Guido Porto Evidence from Cotton Reforms in Zambia

11805 Justine S. Hastings Parental Preferences and School Competition: Thomas J. Kane Evidence from a Public School Choice Program Douglas O. Staiger

11806 Olivier Blanchard Real Wage Rigidities and the New Keynesian Model Jordi Gali

11807 John F. Helliwell Well-Being, Social Capital, and Public Policy: What’s New?

11808 Leora Friedberg Searching for Better Prospects: Endogenizing Michael T. Owyang Failing Job Tenure and Private Pension Coverage Tara M. Sinclair

11809 Jerry Hausman Consumer Benefits from Increased Competition in Ephraim Leibtag Shopping Outlets: Measuring the Effect of Wal-Mart

11810 Thomas J. Philipson Who Benefits from New Medical Technologies? Anupam B. Jena Estimates of Consumer and Prodicer Surpluses for HIV/AIDS Drugs

11811 Mark Duggan Aching to Retire? The Rise in the Full Retirement Age Perry Singleton and its Impact on the Disability Rolls Jae Song

11812 Susanna Loeb How Much is Too Much? The Influence of Preschool Margaret Bridges Centers on Children’s Social and Cognitive Development Daphna Bassok Bruce Fuller Russ Rumberger

11813 Patrick Bayer Racial Sorting and Neighborhood Quality Robert McMillan

11814 Frederic S. Mishkin How Big a Problem is Too Big to Fail?

62 NBER Reporter Winter 2005/6 ------Paper Author(s) Title

11815 Jordi Gali Optimal Monetary and Fiscal Policy in a Currency Union Tommaso Monacelli

11816 Tano Santos Cash-Flow Risk, Discount Risk, and the Value Premium Pietro Veronesi

11817 Paul Oyer Salary or Benefits?

11818 Dana Goldman Medical Expenditure Risk and Nicole Maestas Household Portfolio Choice

11819 Katherine Ho The Welfare Effects of Restricted Hospital Choice in the US Medical Care Market

11820 Ricardo Reis Inattentive Producers

11821 Paul R. Bergin Towards a Theory of Firm Entry and Stabilization Policy Giancarlo Corsetti

11822 Katherine Ho Insurer-Provider Networks in the Medical Care Market

11823 Caroline Freund Current Account Deficits in Industrial Countries: Frank Warnock The Bigger they are the Harder they Fall?

11824 Andrew Ang Downside Risk Joseph Chen Yuhang Xing

11825 Dora L. Costa Surviving Andersonville: The Benefits of Matthew E. Kahn Social Networks in POW Camps

11826 George-Marios Angeletos Efficiency and Welfare with Complementarities Alessandro Pavan and Asymmetric Information

11827 James Markusen Modeling the Offshoring of White-Collar Services: From Comparative Advantage to the New Theories of Trade and FDI

11828 David Hummels Trade in Ideal Varieties: Theory and Evidence Volodymyr Lugovskyy

11829 David Card The Effect of Firm-Level Contracts on the Structure Sara de la Rica of Wages: Evidence from Matched Employer-Employee Data

11830 Stephen G. Cecchetti Do Capital Adequacy Requirements Lianfi Li Matter for Monetary Policy?

------NBER Reporter Winter 2005/6 63 Paper Author(s) Title

11831 Laurence J. Kotlikoff Would the FairTax Raise or Lower Marginal David Rapson and Average Tax Rates

11832 Michael Baker Universal Childcare, Maternal Labor Supply, and Jonathan Gruber Family Well-Being Kevin Milligan

11833 Laurence J. Kotlikoff Who’s Going Broke? Comparing Growth in Healthcare Christian Hagist Costs in Ten OECD Countries

11834 Ricardo J. Caballero Financial System Risk and Flight to Quality Arvind Krishnamurthy

11835 Joshua D. Angrist New Evidence on the Causal Link Between the Victor Lavy Quantity and Quality of Children Analia Schlosser

11836 Chulhee Lee Rising Family Income Inequality in the United States, 1968-2000: Impacts of Changing Labor Supply, Wages, and Family Structure

11837 Dan T. Rosenbaum The Cost of Caring for Young Children Christopher J. Ruhm

11838 Sendhil Mullainathan Persuasion in Finance Andrei Shleifer

11839 Esteban Rossi-Hansberg Firm Fragmentation and Urban Patterns Pierre-Daniel Sarte Raymond Owens III

11840 Jacob Boudoukh The Information in Long-Maturity Forward Rates: Matthew Richardson Implications for Exchange Rates and the Forward Robert Whitelaw Premium Anomaly

11841 Jacob Boudoukh The Myth of Long-Horizon Predictability Matthew Richardson Robert Whitelaw

11842 Ian Dew-Becker Where did the Productivity Growth Go? Inflation Robert J. Gordon Dynamics and the Distribution of Income

11843 Nicolae Garleanu Demand-Based Option Pricing Lasse Heje Pedersen Allen M. Poteshman

64 NBER Reporter Winter 2005/6 ------Paper Author(s) Title

11844 Donald Boyd How Changes in Entry Requirements Alter the Pamela Grossman Teacher Workforce and Affect Student Achievement Hamilton Lankford Susanna Loeb James Wyckoff

11845 Daron Acemoglu Equilibrium Bias of Technology

11846 Richard J. Murnane Improving the Performance of the Education Sector: Richard R. Nelson The Valuable, Challenging, and Limited Role of Random Assignment Evaluations

11847 Jeffrey R. Campbell Competition in Large Markets

11848 Charles F. Manski Search Profiling with Partial Knowledge of Deterrence

11849 Rebecca Henderson Inertia and Incentives: Bridging Organizational Sarah Kaplan Economics and Organizational Theory

11850 Axel Boersch-Supan Aging, Pension Reform, and Capital Flows: Alexander Ludwig A Multi-Country Simulation Model Joachim Winter

11851 Xavier Gabaix Limits of Arbitrage: Theory and Evidence from the Arvind Krishnamurthy Mortgage-Backed Securities Market Olivier Vigneron

11852 Annabelle Gawer Platform Owner Entry and Innovation in Rebecca Henderson Complementary Markets: Evidence from Intel

11853 Menzie D. Chinn Three Current Account Balances: Jaewoo Lee A “Semi-Structuralist” Interpretation

11854 Stephanie Schmitt-Grohe Optimal Inflation Stabilization in a Martin Uribe Medium-Scale Macroeconomic Model

11855 David Neumark The Impact of Provider Choice on Workers’ Peter S. Barth Compensation Costs and Outcomes Richard Victor

11856 Martin Feldstein Monetary Policy in a Changing International Environment: The Role of Capital Flows

11857 Edward L. Glaeser Myths and Realities of American Political Geography Bryce A. Ward

------NBER Reporter Winter 2005/6 65 Paper Author(s) Title

11858 Laurence J. Kotlikoff Simulating the Dynamic Macroeconomic and Sabine Jokisch Microeconomic Effects of the FairTax

11859 Joni Hersch The Generational Divide in Support for Environmental W. Kip Viscusi Policies: European Evidence

11860 Edward J. Kane Can the European Community Afford to Neglect the Need for More Accountable Safety-Net Management?

11861 Luca Benzoni Can Standard Preferences Explain the Prices of Pierre Collin-Dufresne Out-of-the-Money S&P 500 Put Options Robert S. Goldstein

11862 Adi Brender How Do Budget Deficits and Economic Growth Affect Allan Drazen Reelection Prospects? Evidence from a Large Cross-Section of Countries

11863 W. Kip Viscusi The Perception and Valuation of the Risks of Richard J. Zeckhauser Climate Change: A Rational and Behavioral Blend

11864 Jaime Casassus Equilibrium Commodity Prices with Irreversible Pierre-Collin Dufresne Investment and Non-Linear Technology Bryan R. Routledge

11865 Edward J. Kane Confronting Divergent Interests in Cross-Country Regulatory Arrangements

11866 Timothy J. Hatton A Dual Policy Paradox: Why Have Trade and Jeffrey G. Williamson Immigration Policies Always Differed in Labor-Scarce Economies?

11867 Patrick Bolton Thinking Ahead: The Decision Problem Antoine Faure-Grimaud

11868 Farley Grubb The Net Asset Position of the U.S. National Government, 1784-1802: Hamilton’s Blessing or the Spoils of War?

11869 David B. Gordon Are Countercyclical Fiscal Policies Counterproductive? Eric M. Leeper

11870 Adam Copeland The Response of Prices, Sales, and Output George Hall to Temporary Changes in Demand

11871 Harsha Thirumurthy The Economic Impact of AIDS Treatment: Joshua Graff Zivin Labor Supply in Western Kenya Markus Goldstein

66 NBER Reporter Winter 2005/6 ------Paper Author(s) Title

11872 Lingxin Hao Games Parents and Adolescents Play: Risky Behaviors, V. Joseph Hotz Parental Reputation, and Strategic Transfers Ginger Z. Jin

11873 V. Joseph Hotz The Impact of Minimum Quality Standards on Firm Entry, Mo Xiao Exit, and Product Quality: the Case of the Child Care Market

11874 Troy Davig Generalizing the Taylor Principle Eric M. Leeper

11875 John A. Vernon The Future of Drug Development: W. Keener Hughen The Economics of Phamacogenomics

11876 Lubos Pastor Technological Revolution and Stock Prices Pietro Veronesi

11877 Peter Diamond Pensions for an Aging Population

11878 Ann Huff Stevens The More Things Change, the More They Stay the Same: Trends in Long-term Employment in the United States, 1969-2002

11879 Shin-Yi Chou Fast-Food Restaurant Advertising on Television Inas Rashad and its Influence on Childhood Obesity Michael Grossman

11880 Esther Duflo Monitoring Work: Getting Teachers to Come to School Rema Hanna

11881 Henry Saffer The Demand for Social Interaction

11882 Owen A. Lamont Investor Sentiment and Corporate Finance: Jeremy C. Stein Micro and Macro

11883 Simeon Djankov The Law and Economics of Self-Dealing Rafael La Porta Florencio Lopez-de-Silanes Andrei Shleifer

11884 Kevin H. O’Rourke Did Vasco da Gama Matter for European Markets? Jeffrey G. Williamson Testing Frederick Lane’s Hypotheses Fifty Years Later

11885 Robert C. Feenstra Contractual Versus Generic Outsourcing: Barbara J. Spencer The Role of Proximity

11886 Lucian Bebchuk Firm Expansion and CEO Pay Yaniv Grinstein

------NBER Reporter Winter 2005/6 67 Paper Author(s) Title

11887 Amil Petrin Measuring Aggregate Productivity Growth Using James Levinsohn Plant-Level Data

11888 Robert Shimer Mismatch

11889 William D. Nordhaus Life After Kyoto: Alternative Approaches to Global Warming

11890 James E. Rauch Neckties in the Tropics: A Model of International Vitor Trindade Trade and Cultural Diversity

11891 Frederic S. Mishkin Is Financial Globalization Beneficial?

11892 Marianne Bertrand What’s Psychology Worth? A Field Experiment Dean Karlan in the Consumer Credit Market Sendhil Mullainathan Eldar Shafir Jonathan Zinman

11893 Pinka Chatterji Psychiatric Disorders and Labor Market Outcomes: Margarita Alegria Evidence from the National Latino and Asian American Study Mingshan Lu David Takeuchi

11894 Ross Levine Internationalization and Stock Market Liquidity Sergio Schmukler

11895 Peter Kuhn The Expanding Workweek? Understanding Trends Fernando Lozano in Long Work Hours Among U.S. Men, 1979-2004

11896 Michael Woodford Robustly Optimal Monetary Policy with Near Rational Expectations

11897 Michael D. Bordo The Role of Foreign Currency Debt in Financial Crises: Christopher M. Meissner 1880-1913 vs. 1972-1997

11898 Michael Woodford Central Bank Communication and Policy Effectiveness

11899 David W. Galenson The Greatest Artists of the Twentieth Century

11900 Gabriel Weintraub Markov Perfect Industry Dynamics with Many Firms C. Lanier Benkard Ben Van Roy

11901 Laura Alfaro Why Doesn’t Capital Flow from Rich to Poor Countries? Sebnem Kalemli-Ozcan An Empirical Investigation Vadym Volosovych

68 NBER Reporter Winter 2005/6 ------Paper Author(s) Title

11902 Philip J. Cook The Myth of the Drinker’s Bonus

11903 Andrew Ang CAPM Over the Long Run: 1926-2001

11904 Abhijit Banerjee Remedying Education: Evidence from Two Randomized Shawn Cole Experiments in India Esther Duflo Leigh Linden

11905 Ricardo Hausmann What You Export Matters Jason Hwang Dani Rodrik

11906 Geert Bekaert International Stock Return Comovements Robert J. Hodrick Xiaoyan Zhang

11907 Lucian A. Bebchuk Executive Pensions Robert J. Jackson

11908 Lee. J. Alston Who Should Govern Congress? Jeffrey A. Jenkins Access to Power and the Salary Grab of 1873 Tomas Nonnenmacher

11909 Lisa Sanbonmatsu Neighborhoods and Academic Achievement: Jeffrey R. Kling Results from the Moving to Opportunity Experiment Greg J. Duncan Jeanne Brooks-Gunn

11910 Sendhil Mullainathan Sticking with Your Vote: Cognitive Dissonance and Voting Ebonya Washington

11911 David Bell The Scots May be Brave but They are Neither Healthy David G. Blanchflower Nor Happy

11912 Rene M. Stulz Financial Globalization, Corporate Governance, and Eastern Europe

11913 Martin Uribe On Overborrowing

11914 A. Burak Guner The Impact of Boards with Financial Expertise Ulrike Malmendier on Corporate Policies Geoffrey Tate

11915 Ebonya Washington How Black Candidates Affect Voter Turnout

------NBER Reporter Winter 2005/6 69 Paper Author(s) Title

11916 Orley Ashenfleter Measuring the Value of a Statistical Life: Problems and Prospects

11917 Pierre Azoulay The Impact of Academic Patenting on the Rate, Quality, Waverly Ding and Direction of (Public) Research Output Toby Stuart

11918 David Card When to Start a Fight and When to Fight Back: Liability Brian P. McCall Disputes in the Workers Compensation System

11919 Sung Won Kang Capitalizing Patriotism: The Liberty Loans of World War I Hugh Rockoff

11920 John Beshears Early Decisions: A Regulatory Framework James J. Choi David Laibson Brigitte C. Madrian

11921 Charles Engel The U.S. Current Account Deficit and the John H. Rogers Expected Share of World Output

11922 Pierpaolo Benigno Overconfidence, Subjective Perception, Anastasios G. Karantounias and Pricing Behavior

11923 Lawrence H. Goulder The Economics of Climate Change William A. Pizer

11924 Ebonya Washington Female Socialization: How Daughters Affect Their Legislator Fathers’ Voting on Women’s Issues

11925 Andrew Leigh Happiness and the Human Development Index: Justin Wolfers Australia is Not a Paradox

11926 Mary Amiti Service Offshoring and Productivity: Shang-Jin Wei Evidence from the United States

11927 Zeynep Hansen The Political Economy of “Truth-in-Advertising” Marc T. Law Regulation During the Progressive Era

11928 Diego Comin Five Facts You Need to Know About Technology Diffusion Bart Hobijn Emilie Rovito

11929 Refet Gurkaynak Macroeconomic Derivatives: An Initial Analysis of Justin Wolfers Market-Based Macro Forecasts, Uncertainty, and Risk

70 NBER Reporter Winter 2005/6 ------Paper Author(s) Title

11930 Tomas Philipson IP & External Consumption Effects: Generalizations Stephanie Mechoulan from Health Care Markets Anupam B. Jena

11931 Jeremy Atack Steam Power, Establishment Size, and Labor Fred Bateman Productivity Growth in Nineteenth Century Robert Margo American Manufacturing

11932 Peter Kuhn The Long-Term Effects of a Generous Income Chris Riddell Support Program: Unemployment Insurance in New Brunswick and Maine, 1940-1991

11933 Joshua Aizenman Globalization and Developing Countries — Yothin Jinjarek A Shrinking Tax Base?

11934 W. Kip Viscusi Regulation of Health, Safety, and Environmental Risks

11935 W. Kip Viscusi Hyperbolic Discounting of Public Goods Joel Huber

11936 Charles T. Clotfelter Teacher-Student Matching and the Assessment of Helen F. Ladd Teacher Effectiveness Jacob L. Vigdor

11937 V. Joseph Hotz Strategic Information Disclosure: the Case of Mo Xiao Multi-Attribute Products with Heterogenous Consumers

11938 Rosemary Avery Private Profits and Public Health: Does Advertising Donald Kenkel Smoking Cessation Products Encourage Smokers to Quit? Dean R. Dillard Alan Mathios

11939 V. Joseph Hotz Evaluating the Differential Effects of Alternative Guido W. Imbens Welfare-to-Work Training Components: A Re-Analysis Jacob A. Klerman of the California GAIN Program

------NBER Reporter Winter 2005/6 71 Nonprofit Org. U.S. Postage NBERReporter PAID National Bureau of NATIONAL BUREAU OF ECONOMIC RESEARCH Economic Research

1050 Massachusetts Avenue Cambridge, Massachusetts 02138-5398 (617) 868-3900

Change Service Requested