NBER Reporter

NATIONAL BUREAU OF ECONOMIC RESEARCH

Reporter OnLine at: www.nber.org/reporter SUMMER 2005

Program Report IN THIS ISSUE Program Report: The Conference on Research in CRIW 1 Income and Wealth Research Summaries: How Private Health Insurance Pools Risk 5 . . . Emerging Market Economies 8 Charles R. Hulten* . . . Social Reform in the U.S. 11 Behavioral Finance 14

NBER Profiles 17 In 1935, the National Bureau of Economic Research invited a con- Conferences 18 sortium of universities to join with it in establishing a “program of Bureau News 31 cooperative research” in the area of economic measurement. The Bureau Books 53 Conference on Research in Income and Wealth (CRIW) was founded a Current Working Papers 54 year later to provide conceptual support for the establishment of the U.S. National Income and Product Accounts. It was a conference in the basic sense of the word — a group of experts drawn from academia, government statistical agencies, and the business and economic policy community, meeting at regular intervals to confer on issues of impor- tance for the development of the national accounts. While the CRIW has broadened its agenda to include measurement issues beyond the strict confines of the national accounts, the essential personality and objectives of the CRIW have remained unchanged up to the present. The CRIW has special status at the NBER, thus setting it apart from the Bureau’s regular programs. This status reflects the dual per- sonality of the organization: it is an activity within the NBER, which exercises oversight and provides administrative support, but at the same NBER Service Brings time, it has a separate membership, steering committee, and its own ded- icated funding from the government statistical agencies. Ongoing finan- You New Data for Free cial support is derived from the Bureau of the Census, the Bureau of Economic Analysis, the Bureau of Labor Statistics, the Board of A new, free NBER email serv- Governors of the System, and Statistics Canada. The ice gives you daily email links to all executive committee is drawn from these agencies, from academe, and U.S. government data releases, from program directors appointed by the NBER. The CRIW currently including unemployment, , has some 300 members. Election to membership in the CRIW does not interest rates, GDP, etc. We keep convey affiliation with the NBER, although there is considerable track of your preferences and overlap. email you the requested links when they are released. To sign up for any or all of the government releases, visit www.nber.org/releases * Hulten Chairs the Executive Committee of CRIW and is an NBER Research and register your choices. IT’S Associate in the Program on Productivity. He is also a Professor of Economics at the FREE!! University of Maryland. For links to the work discussed in this article, see www.nber.org/CRIW/.

NBER Reporter Summer 2005 1 The main activity of the CRIW is to hold annual conferences on subjects defined by the conference membership. The conference pro- NBER Reporter ceedings then appear in the series Studies in Income and Wealth, published by the NATIONAL BUREAU OF ECONOMIC RESEARCH University of Chicago Press for the NBER. This series now contains nearly 70 volumes. In recent years, the CRIW has joined with The National Bureau of Economic Research is a private, nonprofit research NBER’s Program on Productivity to hold organization founded in 1920 and devoted to objective quantitative analysis of the American economy. Its officers and board of directors are: workshops at the annual Summer Institutes, and this initiative has broadened to include President and Chief Executive Officer — Martin Feldstein other NBER programs. The conference vol- Vice President for Administration and Budget — Susan Colligan Controller — Kelly Horak umes published over the last decade include: The Economics of New Goods (1997), Geography BOARD OF DIRECTORS and Ownership as Bases for Economic Accounting Chairman — Michael H. Moskow (1998), Labor Statistics Measurement Issues (1998), Vice Chairman — Elizabeth E. Bailey International and Interarea Comparisons of Income, Treasurer — Robert Mednick Output, and Prices (1999), New Developments in DIRECTORS AT LARGE Productivity Analysis (2001), Medical Care Output Peter Aldrich Martin Feldstein Laurence H. Meyer and Productivity (2001), and Scanner Data and Price Elizabeth E. Bailey Jacob A. Frenkel Michael H. Moskow Indexes (2003). The last three years have been John Herron Biggs Judith M. Gueron Alicia Munnell particularly busy, with five regular CRIW con- Andrew Brimmer Robert S. Hamada Rudolph A. Oswald John S. Clarkeson George Hatsopoulos Robert T. Parry ferences whose volumes are not yet published: Don R. Conlan Karen N. Horn Richard N. Rosett Measuring Capital in the New Economy (2002), George Eads Judy Lewent Marina v. N. Whitman Jessica P. Einhorn John Lipsky Martin B. Zimmerman Hard-to-Measure Goods and Services: Essays in Memory of Zvi Griliches (2003), New Architecture DIRECTORS BY UNIVERSITY APPOINTMENT for the U.S. National Accounts (2004), Price Index , California, Berkeley Joel Mokyr, Northwestern Concepts and Measurement (2004), and Producer Jagdish W. Bhagwati, Columbia Andrew Postlewaite, Pennsylvania Dynamics: New Evidence from Micro Data (2005). Michael J. Brennan, California, Los Angeles Uwe Reinhardt, Princeton In addition, the CRIW has collaborated in Glen G. Cain, Wisconsin Nathan Rosenberg, Stanford Ray C. Fair, Yale Craig Swan, Minnesota three other conferences and five workshops. A Franklin Fisher, MIT David B. Yoffie, Harvard listing of recent conferences and programs Saul H. Hymans, Michigan Arnold Zellner, Chicago Marjorie B. McElroy, Duke and workshops is available on the CRIW web page (www.nber.org/CRIW/), along with a list DIRECTORS BY APPOINTMENT OF OTHER ORGANIZATIONS of volumes in the Studies in Income and Wealth series. Richard B. Berner, National Association for Business Economics Gail Fosler, The Conference Board This list of recent conference papers is Dr. Arthur Kennickell, American Statistical Association too long to describe fully in this brief report. Richard C. Green, American Finance Association Thea Lee, American Federation of Labor and Congress of However, there are several cross cutting Industrial Organizations themes: the importance of the revolution in Robert Mednick, American Institute of Certified Public Accountants information technology; the emergence of Angelo Melino, Canadian Economics Association Jeffrey M. Perloff, American Agricultural Economics Association new goods and services whose quality differ- John J. Siegfried, American Economic Association ences are important and changing; and the William W. Lewis, Committee for Economic Development globalization of economic activity. These Gavin Wright, Economic History Association dynamic issues provide a formidable challenge The NBER depends on funding from individuals, corporations, and private to economic measurement, since the national foundations to maintain its independence and its flexibility in choosing its research activities. Inquiries concerning contributions may be addressed to accounts are not a static structure, but rather Martin Feldstein, President & CEO, NBER 1050 Massachusetts Avenue, one that must evolve continuously to reflect Cambridge, MA 02138-5398. All contributions to the NBER are changes in the structure of the economy. deductible. The Reporter is issued for informational purposes and has not been reviewed by New Goods 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 Public Information Department with copies of anything reproduced. The 1980s and early 1990s were the peri- od in which personal computers, cell phones, Preparation of the NBER Reporter is under the editorial supervision of Donna Zerwitz. video games, ATMs, and then the Internet, became common items of daily life. The Requests for subscriptions, changes of address, and cancellations should be growing importance of these items, as well as sent to Reporter, National Bureau of Economic Research, Inc., 1050 Massachusetts Avenue, Cambridge, MA 02138-5398. Please include the cur- other aspects of the IT revolution, created a rent mailing label. disconnect between every-day experience and

2 NBER Reporter Summer 2005 macroeconomic statistics, prompting Measuring Capital Market Value: Building Up from Micro to remark in 1987 that, Evidence” by Abowd, Haltiwanger, “we see the computer revolution every- Technological innovation also Jarmin, Lane, Lengermann, McCue, where except in the statistics.” Federal causes equally serious problems for the McKinney, and Sandusky. Reserve Board Chairman measurement of capital, particularly for also noted this disconnect, and sug- the intangible “knowledge” assets like National Income Accounts gested that official statistics fail to cap- accumulated research and development ture the true dynamism of the econo- and human capital. Conventional accoun- New and future developments at my. He observed that official statistics ting practice treats R and D expendi- the BEA are also discussed in the vol- implied that productivity in the service tures as an intermediate expense that is ume New Architecture for the U.S. sector was virtually non-existent, used up in the same period in which it National Accounts. The paper by despite the widespread adoption of was produced, thereby ignoring the Landefeld and Jorgenson in this vol- cost-saving processes and new prod- large body of research that finds a large ume sets out a road map for the evolu- ucts based on the IT revolution. He rate of return to R and D as a capital tion of the U.S. NIPA in “Blueprint for also told the Senate Finance investment. This issue was taken up in an Expanded and Integrated Set of Committee in 1995 that he thought the 2003 conference Measuring Capital Accounts for the United States.” In a that the growth rate of the consumer in the New Economy. The paper by sense, this paper takes up the unfin- price index was biased upward by Corrado, Hulten, and Sichel presented ished business left over from the early between one-half to one and a half at this conference explores the conse- days of the national accounting move- percentage points per year because of quences for accounting practice of ment. A perusal of the early volumes mismeasurement. The Boskin Com- counting as investment a broad list of of the Studies in Income and Wealth mission, impaneled to investigate the intangibles – R and D, copyrights, films, reveals that the original intent of the matter, put the upward bias at around computerized databases, development “founding fathers” was to develop a one percent per year, and attributed of improved organizational structures, fully integrated income and wealth about half to the failure of the CPI to and brand equity. Business investment account, with allowance for non-mar- accurately reflect product innovation. in intangibles was found to be as large ket production. This vision was not The influence of this critique is as the spending on tangible capital – as realized when the national accounts evident in the CRIW conference, The much as $1 trillion in recent years – were first published in the late 1940s. Economics of New Goods. William and a significant portion of this These accounts emerged with a dis- Nordhaus, writing in this volume, amount is excluded from the existing tinctly Keynesian personality that traces the history of lighting and investment figures in the NIPA. emphasized short-run market flows, argues that, “by the very nature of In subsequent work reported at a without a balance sheet, a non-market their construction, price indexes miss CRIW/Productivity Program work- production, or a true production the most important technological rev- shop at last year’s NBER Summer account that allowed for capital servic- olutions in history (page 54).” Writing Institute, a follow-on study found that es as a productive input. Following the in the same volume, Jerry Hausman the rate of growth of measured labor paper by Landefeld and Jorgenson, the points out that the CPI did not include productivity in the non-farm business other contributions in New Architecture cellular telephones in the CPI market sector is strongly affected by the capi- for the U.S. National Accounts take up basket until some 10 years after their talization of this list of intangible these various issues. The issues in- introduction into the market place. He assets, and that the proportion of volved in developing a national balance proposes, in that paper, a procedure growth attributed to capital formation sheet and linking the BEA investment for handling new goods based on the is the dominant source of growth and capital stock estimates to the Hicksian reservation price. However, when intangible assets are included in FRB’s wealth and flow-of-funds esti- measurement practice has tended to the analysis. mates are described in “Integrated adopt the hedonic price model as the This practice of excluding intan- Macroeconomic Accounts for the appropriate framework. Price hedo- gibles is beginning to change at the United States: Draft SNA_USA,” by nics was used by the Bureau of macro level, with BEA’s decision to Teplin, Antoniewicz, McIntosh Economic Analysis in the mid 1980s to capitalize software expenditures and Palumbo, Solomon, Mead, Moses, and adjust prices and quantities in the the planned capitalization of R and D Moulton. Non-market accounting is national accounts for the rapid expenditures in a satellite account to discussed in “A Framework for Non- increase in computing power, and has the main NIPA. This issue is taken up Market Accounting,” Abraham and become the focus of attention at the in the paper by Fraumeni and Okubo, Mackie, and “Principles of National BLS as a means of improving the CPI “R&D in the National Income and Accounting for Non-Market Accounts,” program. The hedonic price model Product Accounts: A First Look at Its by Nordhaus. My paper, “Architecture has been given a great deal of attention Effect on GDP,” in Measuring Capital in of Capital Accounting: Basic Design at many recent CRIW conferences, the New Economy. The importance of Principles” deals with the problem of including Hard-to-Measure Goods and firm-specific investments in human developing a production account with Services: Essays in Memory of Zvi Griliches capital is discussed in “The Relation of capital service flows, linked to the cor- and Price Index Concepts and Measurement. Human Capital, Productivity, and responding stocks of capital and

NBER Reporter Summer 2005 3 wealth. This is also taken up in the ductivity after 1995 in the service sec- issues of importance to the CRIW. paper by Harper, Powers, Fraumeni, tors and its association with IT invest- Strengthening the linkage between the and Yuskavage, “An Integrated ments focused even more attention on aggregate data of the NIPA and it BEA/BLS Production Account: A First the measurement problem. It is there- microfoundations is another priority, Step and Theoretical Considerations.” fore not surprising, given this history, reflected in the most recent CRIW that the CRIW has devoted a great deal conference, Producer Dynamics: New Service Sector Output of attention to the subject, with the Evidence from Micro Data. The problems 1969 conference Production and posed by the globalization of econom- The problems associated with the Productivity in the Service Industries, the ic activity present yet another challenge measurement of service sector output 1992 Output Measurement in the Service for economic measurement. What represent another item of “old busi- Sectors, and more recently, Medical Care does it mean to have “national” ness” on the CRIW agenda. This sec- Output and Productivity (2001), Scanner accounts when capital and technology tor has grown steadily since the first Data and Price Indexes (2003), and the (and now labor as well) flow ever more publication of the U.S. national yet-to-be published proceedings Hard- freely across national boundaries? accounts, increasing from 36 percent to-Measure Goods and Services: Essays in Various aspects of this question are of GDP in 1947 to 56 percent by Memory of Zvi Griliches (2003), and Price examined in the CRIW volumes 1997. The failure to measure the out- Index Concepts and Measurement. Other International and Interarea Comparisons of put of services has been advanced as CRIW conferences and workshops Income, Output, and Prices and Geography an explanation of the productivity had papers on banking output, insur- and Ownership as Bases for Economic slowdown of the mid-1970s to the mid- ance, and educational output. Accounting, and will likely be the subject 1990s, and was given added salience by of the next CRIW conference in 2006. Greenspan’s skepticism about the slow Additional Work These problems and others will keep growth of productivity in a sector in the CRIW as busy in the next decade which the IT revolution was so appar- The topics reviewed thus far are as it has been in the past. ently important. The pick up in pro- by no means an exhaustive list of the

*

4 NBER Reporter Summer 2005 Research Summaries

How Private Health Insurance Pools Risk

Mark Pauly*

Introduction and Theory Richard Hirth, and I1 have shown, as full amount of additional risk that has John Cochrane,2 that there is a the- characterizes individuals and families. Most Americans obtain their oretical solution to this problem: mar- Moreover, in both markets the premi- health insurance in the private sector, kets can furnish incentive-compatible ums do not increase with the onset of in both group and non-group settings, insurance with “a longer time perspec- chronic conditions at rates close to the and from for-profit, non-profit, and tive,” in Kenneth Arrow’s words using expected expenses associated with self-insured arrangements. A matter of the policy provision labeled “guaran- those conditions Although risk pool- concern to both consumers and poli- teed renewability at class average pre- ing is present (but far from complete) cymakers is how the premiums that miums.” We also have been interested in employment-based group markets, will (in some fashion) be paid for this in pursuing another of Arrow’s even in individual insurance markets insurance — and even the availability insights: that institutional arrange- there is relatively little risk segmenta- of insurance at any premium — vary ments other than explicit and direct tion. The difference in the extent of with the insured unit’s risk level. There market transactions may emerge to risk pooling in the two types of mar- is tension here: to cover their costs and deal with this problem as with others kets is modest.3 to avoid adverse selection, insurers in health care and health insurance; the Our primary theoretical finding is need to collect premiums tailored to main candidate for this role is employ- that such risk pooling, far from being each buyer’s expected expense. But ment-based group health insurance. inconsistent with the operation of policymakers tend to regard payment We therefore have been investigating competitive insurance markets, is actu- of higher premiums by higher risks as the extent to which premiums paid ally what would be predicted in an unfair, and individual consumers real- vary with risk in competitive, largely incentive-compatible competitive equi- ize that their risk level may change over unregulated insurance markets, and the librium, according to models of insur- time, as chronic (though not necessar- underlying arrangements that support ance purchase in which buyers demand ily permanent) illnesses strike. risk pooling. This investigation has and insurers (implicitly or explicitly) Along with several colleagues, I largely been one of first discovering supply the policy feature of “guaran- have been investigating both theoreti- some empirical evidence and then teed renewability.”4 cal models of efficient insurance mar- working backwards to understand the These findings are most transpar- kets when risk varies (both across indi- supporting arrangement and the theo- ent in the individual or non-group viduals at a given point in time and for ry that explains their existence. insurance market (even though it is a given individual over time) and currently a small fraction of total pri- empirical evidence on how premiums Individual Insurance vate insurance). We assume that early and the securing of coverage vary with Markets in life people are generally of similar risk. The theoretical point of depar- risk (with only 3 percent of young ture is that, from a lifetime perspective, Our primary empirical finding is adults reporting the presence of the great bulk of people who are ini- that, in both group and non-group chronic conditions). To obtain insur- tially low risk will want protection insurance, premiums actually paid for ance for current coverage that also against large current costs and protec- insurance in the United States pool the protects against so-called “reclassifica- tion against “premium risk,” the risk risk to a very great extent. That is, in tion risk” — an increase in premiums that premiums will jump at the onset of both markets, actual premiums paid are if one contracts a chronic condition — chronic illness. Howard Kunreuther, not even close to being proportional to the great bulk of individual insurance risk. To be specific: although total pre- carries a provision promising “guaran- miums do rise with risk in both indi- teed renewability at class average * Pauly is a Research Associate in the vidual and group markets, primarily in rates.” This means that the insurer NBER’s Program on Health Care and a terms of reflecting the higher risk that promises to charge a premium — if professor at the University of Pennsylvania. accompanies older age (especially the person renews coverage in the next His profile appears later in this issue. above age 50), they do not reflect the period — which is the same for all of

NBER Reporter Summer 2005 5 those who initially bought insurance as We also examine the effects of requires states to have such a provision part of a “class.” Specifically, it means state regulation requiring community in their regulatory code, but leaves the that the insurer promises not to single rating, or putting bands on risk rating interpretation and the enforcement up out for above-average premiums those of premiums charged to newly cov- to the states. Patel and I find that who develop evidence of becoming ered individual insureds, and how dif- almost all state insurance departments higher risks. If the insurer raises the ferent types of insurers behave in this say that they have a guaranteed premium, it promises to do so for regard. Using data from the 1980s and renewability provision, but in some everyone uniformly. 1990s, we find little evidence that state states it is an explicit prohibition on The theoretical model shows that regulation led to less variation with risk changing rates if risk changed, where- this provision requires “frontloading” in premiums than did the absence of as in other states it was implicit in the of premiums: in the early years of the regulation. We do find that managed state’s power to forbid “arbitrary” time period over which a person plans care plans vary their premiums with insurance premium rates. to renew coverage from a given insur- risk less than conventional insurers do.6 er, the premium exceeds the expected In more recent work,7 Herring Group Insurance expense because it also includes a and I look at the effect of regulation charge to pay any above average costs on both differential premiums and dif- People who get their insurance for those who later become higher ferential purchases of individual insur- through their jobs are thought by risks. Recent work by Bradley Herring ance. We find that regulation modest- to pay in two ways. In most and me5 compares the estimated age- ly tempers the (already-small) relation- firms there is some explicit premium time path of premiums for individual ship of premium to risk, and leads to a charged to the employee and deducted insurance when risk changes because slight increase in the relative probability from take home pay on the person’s of the onset of chronic conditions. We that high-risk people will obtain indi- pay stub; where there are multiple plan consider the actual path exhibited in vidual coverage. However, we also find offerings, differences in these explicit the Medical Expenditure Panel Survey that the increase in overall premiums premiums affect which plan people (MEPS) data. We find that the actual from community rating slightly choose. The other way employees pay, path is qualitatively similar to what the reduces the total number of people and usually this applies for the much theory would predict (high front load- buying insurance. All of the effects of larger share of insured workers (about ing); the actual path is quantitatively regulation are quite small, though. We 75 percent on average), is in the form close to the path predicted even by our conjecture that the reason for the min- of lower money wages. This incidence crude risk prediction model; and imal impact is that guaranteed renewa- of health benefits (and other employee (because health risk rises with age, bility already accomplishes a large part benefits) on wages has strong theoreti- even for healthy people), younger buy- of effective risk averaging (without the cal and empirical support, but is often ers still pay substantially less in premi- regulatory burden), so additional regu- misunderstood by employers and poli- ums than older buyers for a given lation has little left to change. cymakers, as I explain in my book insurance policy. This occurs despite Guaranteed renewability is not Health Benefits at Work.10 relatively high dropout from individual the only factor affecting individual Herring and I examine employ- plans as people get jobs with coverage insurance premiums. Herring, David ment-based group insurance in the late and return to the group insurance mar- Song, and I8 find that higher risk indi- 1980s. We find that, although the ket. Although there have been some viduals engage in more aggressive explicit employee premium did not anecdotes about insurers slipping out search for lower premiums, with the vary with risk, it did have a huge vari- of their policy provision to renew cov- result (as before) that the difference ance across and even within firms.11 In erage at group average premiums for between premium and expected terms of explicit premiums, workers high risks by canceling the coverage expense was much smaller for higher were not all paying anything close to entirely, we conclude that on average risk people than lower risk people. We the average. We find that total premi- guaranteed renewability works in prac- also analyze Internet data on premi- ums varied little with health risk, and tice as it should in theory and provides ums and find that lower risk individu- that obtaining group insurance was a substantial amount of protection als (primarily young people) have rela- also largely independent of risk, except against high premiums to those high tively more to gain from using the for low-wage workers in small firms. risk individuals who bought insurance Internet to search for lower premiums What we also find, however, is that the before their risk levels changed. The than do higher risk people. But we rate at which money wages increased implication is that, although there are also find that the average gain in terms with experience or seniority was much some anecdotes about individual insur- of lower premiums from using the lower in firms offering health insur- ers trying to avoid covering people Internet to search for individual insur- ance than in otherwise similar firms who become high risk (for example, by ance premiums is zero or negative: if that did not. This finding is consistent canceling coverage for a whole class of there is a gain, it is primarily in reduced with larger differential incidence falling purchasers), the data on actual premi- search time, not lower prices. on older workers (reflecting their high- um-risk relationships strongly suggest Along with Vip Patel,9 I examine er expected losses). that such attempts to limit risk pooling state regulations regarding guaranteed Herring and I also examine strate- are the exception rather than the rule. renewability. The federal HIPAA law gies available to employers to limit

6 NBER Reporter Summer 2005 adverse selection when multiple plans ized or not), such workers tend to be Compatible Guaranteed Renewable with different levels of insurance cov- more likely to select coverage than they Health Insurance,” NBER Working erage are offered.12 We look at the spe- would have in the individual market, Paper No. 9888, August 2003. cific example of employers adding a and to choose coverage with fewer 5 See B. Herring and M. Pauly, high deductible (catastrophic) option, restrictions. There are also some dif- “Incentive Compatible Guaranteed often accompanied by a spending ferences in the effect of ethnicity Renewable Health Insurance.” account — similar to the Medical (Hispanic or Black) in individual versus 6 M. Pauly and B. Herring, “Premium Savings Account or Health Savings group markets. Variation in the Individual Health Account now enacted into federal law. Insurance Market,” International Journal We find that adverse selection by low Conclusion of Health Care Finance and Economics,1 risks is possible if employers follow (1) (2001), pp. 43-58. some type of contribution strategies Our overall conclusion is that pri- 7 M. Pauly and B. Herring, “Premiums (for example, fixed dollar contribu- vate health insurance in the United and Insurance Purchasing in Individual tions to all plans with premium differ- States involves a great deal of risk Health Insurance Markets,” Report to entials reflective of the average differ- pooling as long as individuals initially the Office of the Assistant Secretary ence in cost across plans), but we also obtain insurance (whether individual for Policy and Evaluation, Department develop a model of optimal employer or group) before they contract chronic of Health and Human Services, forth- contribution strategies which greatly illnesses and thus become high risks. coming. reduces the extent of adverse selection Both private guaranteed renewability 8 M. Pauly, B. Herring and D. Song, and, in some cases, permits both high provisions and state rate regulation “Health Insurance on the Internet and and low risk workers to gain from the encourage pooling, but guaranteed the Economics of Search,” NBER addition of such a plan. renewability seems to have fewer Working Paper No. 9299, October Finally, Herring and I examine adverse side effects in the sense of 2002. the differences between the amounts driving lower risks out of the market, 9 V. Patel and M. Pauly, “Guaranteed and types of insurance that people albeit with slight smaller efficacy. The Renewability and the Problem of Risk with different characteristics obtain, major difference between individual Variation in Individual Health depending on whether they use the and group insurance, then, is not the Insurance markets, “ Health Affairs Web individual or group market.13 We use price differences charged to high ver- Exclusive, August 28, 2002. observed demand in the individual sus low risks. Rather, the high loading 10 M. Pauly, Health Benefits at Work: An market as the “gold standard” of the in individual insurance means that it is Economic and Political Analysis of kind of insurance people would costly for people at all risk levels. Employment-based health Insurance (Ann choose if they had completely free Attention should be paid to ways to Arbor: University of Michigan Press, choice (though at much higher load- lower administrative costs across the 1997). ing). We find that most characteristics board. 11 See M. Pauly and B. Herring, Polling that predict both the choice of any Health Insurance Risks. insurance and the degree of restrictive- 1 M. Pauly, H. Kunreuther, and R. 12 M. Pauly and B. Herring, “An ness of the plan, such as income, edu- Hirth, “Guaranteed Renewability in Efficient Employer Strategy for cation, and location, are quite similar in Insurance,” Journal of Risk and Dealing with Adverse Selection in the individual and the group market. Uncertainty, 10 (2) (1995), pp. 143-56. Multiple-plan Offerings,” Journal of Most people do seem to get the kind 2 J. Cochrane, “Time-consistent Health Health Economics, 19 (4) (2000), pp. 513- of plan we would predict that they Insurance,” Journal of Political Economy, 28. would want, even within group set- 103 (3) (1995), pp. 445-73. 13 M. Pauly and B. Herring, “The tings. We find a major difference 3 M. Pauly and B. Herring, Polling Health Demand for Health Insurance in the among worker types who are heavily Insurance Risks (Washington: AEI Press, Group Setting: Can You Always Get unionized though: in group settings 1999). What You Want?” Journal of Risk and (whether the firms itself was union- 4 B. Herring and M. Pauly, “Incentive Insurance, forthcoming.

NBER Reporter Summer 2005 7 Some Perspective on Capital Flows to Emerging Market Economies

Carmen M. Reinhart*

From Hume’s discussion of the decades was attributed initially to one is struck by an overwhelming specie-flow mechanism under the gold domestic developments, such as sound sense of “déjà vu.” It certainly seems a standard to the Keynes-Ohlin debate policies and stronger economic per- mystery why these wide swings in cap- on the transfer problem associated formance, implying both the good use ital flows recur, in spite of the major with German reparations after the of such funds in the recipient country costs associated with them. The com- WWI, understanding the flow of capi- and the informed judgment of mon theme is that investors enter each tal across national borders has been investors in the developed world.1 The episode of upsurge in capital flows central to international economics. My widespread nature of the phenome- confident in the belief that “this time it work on the topic has focused mainly non became clearer over time, though, is different” and look to international on the flow of funds between rich and as most developing countries — financial institutions to make them poor countries. Theory tells us that, whether they had improved, whole when they later learn that it real- for the recipient, foreign capital put to unchanged, or impaired macroeco- ly wasn’t different. good use can finance investment and nomic fundamentals — found them- stimulate economic growth. For the selves the destination of capital from Rich-to-Poor Capital Flows investor, capital flows can increase wel- global financial centers. The single fac- fare by enabling a smoother path of tor encouraging those flows was the To some, the mystery of cross- consumption over time and, through sustained decline in interest rates in the border flows is not these recurrent better risk sharing as a result of inter- industrial world.2 For example, short- cycles unanchored from country con- national diversification, a higher level term interest rates in the United States ditions but rather the restricted volume of consumption. declined steadily in the early 1990s and of these flows overall. Most famously, The reality is that the effects of by late 1992 were a their lowest level Robert Lucas argued that it was a puz- such flows — as seen from either since the early 1960s. Lower interest zle that more capital does not flow recent experience or the longer sweep rates in developed nations attracted from rich countries to poor countries, of history — do not fit neatly into investors to the high yields offered by given back-of-the envelope calcula- those theoretical presumptions. As a economies in Asia and Latin America. tions suggesting massive differences in result, my research has mostly been Given the high external debt burden of physical rates of return in favor of directed at shedding light on four many of these countries, low world capital poor countries.3 Lucas argued questions: interest rates also appeared to improve that the paucity of capital flows to 1. What motivates rich-to-poor capital their credit-worthiness and to reduce poor countries must be rooted in fun- flows? their default risk. Those improvements damental economic forces, such as 2. Why doesn’t capital flow more from were reflected in a marked rise in sec- externalities in human capital forma- rich to poor countries? ondary market prices of bank claims tion favoring further investment in 3. What are the consequences of a on most of the heavily indebted coun- already capital rich countries. My per- surge of capital inflows for an emerging tries and pronounced gains in equity spective, informed by work with market economy? values. Thus, the tightening of mone- and Miguel 4. How do policymakers typically tary policy in the United States and the Savastano, is quite different.4 respond to an incipient inflow of capital? resulting rise in interest rates made Throughout history, governments investment in Asia and Latin America have demonstrated that “serial default” The Causes of Capital relatively less attractive, triggering mar- is the rule, not the exception. ket corrections in several emerging Argentina has famously defaulted on Inflows stock markets and a decline in the five occasions since its birth in the The surge in capital inflows to prices of emerging market debt. 1820s. However, Argentina’s record is emerging market economies in the This experience strongly suggests surpassed by many countries in the early part of each of the past two multiple forms of investor myopia: New World (Brazil, Liberia, Mexico, The initial decision to invest seemed and Uruguay, Venezuela, and Ecuador) more motivated by reaching for yield and by almost as many in the Old * Reinhart is a Research Associate in the without an appropriate appreciation of World (France, Germany, Portugal, NBER's Program on International Finance risk, and the sudden withdrawal simi- Spain, and Turkey). Rogoff, Savastano, and Macroeconomics and a professor of eco- larly looked more like a quick dash for and I argue that this history of repeat- nomics at the University of Maryland. Her the exit door than a reasoned assess- ed defaults makes some countries less profile appears later in this issue. ment of fundamentals. Looking back, able to bear debt. These “debt intoler-

8 NBER Reporter Summer 2005 ant” countries typically have other import data are consistent with the that susceptibility to contagion is high- indicia of governmental failures, interpretation that the consumption ly nonlinear. A single country falling including bouts of high inflation, vari- boom is heavily driven by rising victim to a crisis is not a particularly able macroeconomic policies, and a imports of durable goods. (This held good predictor of crisis elsewhere, be weak rule of law. with particular force in the Latin it in the same region or in another part From this perspective, the key American experience of the early of the globe. However, if several explanation to the “paradox” of why 1990s.) In almost all countries, capital countries fall prey, then it is a different so little capital flows to poor countries inflows were accompanied by rapid story. That is, the probability of a may be quite simple—countries that growth in the money supply — both in domestic crisis rises sharply if a core do not repay their debts have a rela- real and nominal terms — and sharp group of countries are already infect- tively difficult time borrowing from increases in stock and prices. ed. Is the regional complexion of con- the rest of the world. The fact that so For example, during 1991, a major tagion attributable to trade links, as many poor countries are in default on equity index in Argentina posted a dol- some studies have suggested, or to their debts, that so little funds are lar return in excess of 400 percent, financial links — particularly through channeled through equity, and that while Chile and Mexico provided the role played by banks? Our results overall private lending rises more than returns of about 100 percent. suggest that it is difficult to distinguish proportionally with wealth, all strongly Then comes the crisis because the between the two, because most coun- support the view that credit markets surge in capital flows never proves tries linked in trade are also linked in and political risk are the main reasons durable. Unlike their more developed finance. In the Asian crises of 1997, why we do not see more capital flows counterparts, emerging market eco- Japanese banks played a similar role in to developing countries. If credit mar- nomies routinely lose access to interna- propagating disturbances to that ket imperfections abate over time tional capital markets. Furthermore, played by U.S. banks in the debt crisis because of better institutions, then given the common reliance on short- of the early 1980s. human capital externalities or other term debt financing, the public and I identify the links between these “new growth” elements may come to private sectors in these countries often episodes of currency crises and bank- play a larger role. But as long as the are asked to repay their existing debts ing crises in another paper with odds of non-repayment are as high as on short notice. Even with the recent Graciela Kaminsky.8 In particular, 65 percent for some low-income coun- large-scale rescue packages, official problems in the banking sector typical- tries, credit risk seems like a far more financing only makes up for part of ly precede a currency crisis, creating a compelling reason for the paucity of this shortfall. Hence, the need for vicious spiral in which the currency rich-poor capital flows. abrupt adjustment arises. strains then deepen the banking prob- More often than not, contagion lems. The anatomy of these episodes The Consequences of followed on the heels of the initial suggests that crises occur as the econ- shock. The capacity for a swift and omy enters a recession, following a Capital Inflows drastic reversal of capital flows — the prolonged boom in economic activity so-called “sudden stop” problem — that was fueled by credit, capital The experience of many emerg- 5 ing market economies is that attracting played a significant role. An analysis inflows, and accompanied by an over- global investors’ attention is a mixed of the experience of contagious finan- valued currency. blessing of macroeconomic imbal- cial crises over two centuries (with my ances and attendant financial crises. colleagues Graciela Kaminsky and The Policy Response As to the imbalances, a substantial por- Carlos Vègh) finds typically that the tion of the surge in capital inflows announcements that set off the chain Given this experience of wide reactions came as a surprise to finan- swings in foreign funding, it is not sur- tends to be channeled into foreign 6 exchange reserves. For instance, from cial markets. The distinction between prising that policymakers in many 1990 to 1994, the share devoted to anticipated and unanticipated events emerging market economies have reserve accumulation averaged 59 per- appears critical, because advance warn- come to fear large current account cent in Asia and 35 percent in Latin ing allows investors to adjust their deficits, irrespective of how they are America. Moreover, in most countries portfolios in anticipation of the event. financed, but particularly if they are the capital inflows were associated In all cases where there were signifi- financed by short-term debt. The cap- with widening current account deficits. cant immediate international repercus- ital inflow slowdown or reversal could This widening in the current sions, a leveraged common creditor push the country into insolvency or account deficit usually involves both was involved — be it commercial drastically lower the productivity of its an increase in national investment and banks, hedge funds, mutual funds, or existing capital stock. These multiple a fall in national saving. As one would individual bondholders — who helped concerns have produced multiple expect from the fall in national saving, to propagate the contagion across responses to capital inflows. private consumption spending typical- national borders. The policy of first recourse ly rises. While disaggregated data on Additional work with Graciela across countries and over time has Kaminsky indicates that contagion is been sterilized intervention.9 To avoid consumption are not available for all 7 emerging market economies, the more regional than global. We find some (or all) of the nominal exchange

NBER Reporter Summer 2005 9 rate appreciation that would have sector, usually with the aim of deter- servative and zealous resulted from the capital inflow, mone- ring short-term inflows.11 (Sometimes supervision of the domestic financial tary authorities have tended to inter- these controls take the form of pru- sector are essential at all times, espe- vene in the foreign exchange market dential measures to curb the exposure cially when expectations are buoyant. and accumulate foreign exchange of the domestic banking sector to the reserves. To offset some or all of the vagaries of real estate prices and equi- 1 The experience of the early 1990s is associated monetary expansion, central ty markets.) One main finding of my discussed in G.A. Calvo, L. Leiderman, banks have most often opted to sell paper with Todd Smith, however, is and C.M. Reinhart, “Inflows of Capital Treasury bills or central bank paper. that the on capital inflows to Developing Countries in the 1990s,” Central banks also have tools to neu- must be very high in order to have Journal of Economic Perspectives, 10 (2) tralize the effects on the money stock much effect on the capital account bal- (Spring 1996), pp. 123-39. of their foreign exchange operations ance.12 For instance, a reduction in the Comparisons between that episode beyond offsetting domestic open mar- capital account balance by 5 percent of and the Asian crises are drawn in G. L. ket transactions. Importantly, the GDP would require a tax rate on net Kaminsky and C.M. Reinhart, effect of the sale (purchase) of domes- interest payments on foreign-held debt “Financial Crises in Asia and Latin tic currency can be offset by raising on the order of 85 percent for one America: Then and Now,” American (lowering) reserve requirements to year or 60 percent for two years. Economic Review, (May 1998), pp. 444-8. keep the money stock constant.10 Allowing the nominal exchange 2 Regression evidence on the determi- However, as long as domestic reserves rate to appreciate (or be revalued in nants of regional capital flows is pro- do not pay a competitive interest rate, cases where the exchange arrangement vided in C.M. Reinhart and V. R. reserve requirements are a tax on the is less flexible) also has to be consid- Reinhart, “What Hurts Most: G-3 banking system. Changes in the tax can ered as part of the menu of viable pol- Exchange Rate or Interest Rate have real effects, including on the icy responses, particularly as inflows Volatility?” in S. Edwards and J. A. exchange value of the currency. become persistent. As noted in my Frankel, eds. Preventing Currency Crises in Moreover, depending on the incidence paper with Reinhart, long-lived attempts Emerging Markets, Chicago: University of the reserve tax, domestic spending to avoid a nominal appreciation via of Chicago Press, 2001, pp. 73-99. and production may change as well. unsterilized foreign exchange market 3 R. Lucas, “Why Doesn’t Capital Flow Fiscal austerity measures, particu- intervention will fuel a monetary from Rich to Poor Countries,” larly on the spending side, have been expansion (owing to the accumulation American Economic Review, (May 1990), used to alleviate some of the pressures of foreign exchange reserves), which pp. 92-6. on the real exchange rate and to cool may prove inflationary. While sterilized 4 C.M. Reinhart and K.S. Rogoff, down overheating in the economy. intervention may curtail the monetary “Serial Default and the ‘Paradox’ of Furthermore, fiscal surpluses deposit- expansion, it can become both increas- Rich-to-Poor Capital Flows,” NBER ed at the central bank have helped to ingly difficult to implement and costly Working Paper No. 10296, February “sterilize” the expansive monetary over time. In some cases, the authori- 2004, and American Economic Review,94 effects of foreign exchange purchases. ties reached the conclusion that, if an (2), (May 2004), pp. 53-9, and C.M. The process of trade liberaliza- appreciation of the real exchange rate Reinhart, K.S. Rogoff, and M. tion has been accelerated in some was inevitable, it was better that it Savastano, “Debt Intolerance,” NBER cases, in the hope that productivity occur through a change in the nominal Working Paper No. 9908, August 2003, gains in the nontraded sector could exchange rate than through a pick-up in and Brookings Papers on Economic Activity, dampen pressures on the real exchange domestic inflation. (Spring 2003), pp. 1-74. rate. Moreover, reducing distortions Often, policymakers have resort- 5 A term coined by Guillermo A. Calvo associated with controls on trade may ed to some combination of these poli- (see the discussion in G.A. Calvo and temporarily widen the current account cies. A repeated lesson is that the law of C.M. Reinhart, “When Capital Inflows deficit—effectively absorbing some of unintended consequences has not been Come to a Sudden Stop: Consequences the inflows without boosting domestic repealed. Multiple policy responses to and Policy Options” in P. Kenen and A. demand. capital inflows have tended to interact Swoboda, eds., Reforming the International Liberalization of capital outflows in ways that were probably not antici- Monetary and Financial System, Washington also has been a popular response to pated by the framers of such policies. DC: International Monetary Fund, rising capital inflows. By permitting Most likely, even the best policy mix 2000, pp.175-201.) domestic residents to hold foreign cannot altogether avoid the eventual 6 G.L. Kaminsky, C.M. Reinhart, and assets, the conventional wisdom holds, reversal of capital flows, given that they C.A. Vègh, “The Unholy Trinity of gross outflows would increase—there- are so sensitive to the behavior of Financial Contagion,” NBER Working by reducing net. investors in financial centers. The Paper No. 10061, November 2003, and Various forms of controls on appropriate policy mix may dampen Journal of Economic Perspectives, 17 (4), capital inflows—whether in the form the amplitude of the swings in capital (Fall 2003), pp. 51-74. of , quantitative restrictions, or in flows, thus ensuring a softer landing 7 G.L. Kaminsky and C.M. Reinhart, the guise of “prudential measures”— when international investors retrench. “The Center and the Periphery: The have been imposed on the financial The strongest policy lesson is that con- Globalization of Financial Shocks,”

10 NBER Reporter Summer 2005 NBER Working Paper No. 9479, responses in C.M. Reinhart and V.R. (January 1999), pp. 27-54. February 2003, forthcoming in C.M. Reinhart, “Some Lessons for Policy 11 A survey of the literature on capital Reinhart, C. A. Végh and A. Velasco, Makers Who Deal with the Mixed controls is provided in N. Magud and eds. Capital Flows, Crisis, and Blessing of Capital Inflows,” in M. C.M. Reinhart, “Capital Controls: An Stabilization: Essays in Honor of Guillermo Kahler, ed., Capital Flows and Financial Evaluation” in S. Edwards, ed., A. Calvo. Crises, Council on Foreign Relations International Capital Flows, forthcoming 8 For a fuller discussion, see G.L. Book, Ithaca, NY: Cornell University from the University of Chicago Press. Kaminksy and C.M. Reinhart, “The Press, 1998, pp. 93-127. 12 C.M. Reinhart and R.T. Smith, Twin Crises: The Causes of Banking 10 This is discussed in C.M. Reinhart “Temporary Controls on Capital and Balance-of-Payment Problems,” and V.R. Reinhart, “On the Use of Inflows,” NBER Working Paper No. American Economic Review, 89 (3), (June Reserve Requirements in Dealing with 8422, August 2001, and Journal of 1999), pp. 473-500. the Capital-Flow Problem,” International International Economics, 57 (2), August 9 I have documented the typical policy Journal of Finance and Economics, 4 (1), 2002, pp. 327-51.

Prospects for Social Security Reform in the United States

Kent Smetters*

One of the most far-reaching typically already fairly portable across of idiosyncratic risks, traditional shifts in fiscal policy around the world employers. With a few exceptions, the defined-benefit systems more easily during the past two decades has been public pension benefit formula in most allow for sharing wage and longevity the fundamental restructuring of pub- countries is dependent on the wages of uncertainty.2 Relatively larger transac- lic pension systems. Over two dozen the worker regardless of the actual tion costs in defined contribution countries across five continents have employer. plans, as well as problems associated converted at least part of their pay-as- In fact, the public plan conver- with financial literacy, moral hazard, you-go, defined-benefit, public pen- sions seem fairly puzzling at first. To and adverse selection, only seem to sion systems into systems based on be sure, many common arguments buttress the case for the traditional funded, defined-contribution accounts. have been put forth in favor of “per- design. So why are more and more Several additional countries are cur- sonal accounts” including the potential countries abandoning the traditional rently in the process of conversion, to earn higher rates of return in equi- design for the funded, defined contri- and even more countries, including the ties, increased national savings, as well bution model that, in theory, is no bet- United States, are debating it. as greater bequeath-ability. However, ter than the traditional design and The shift toward defined-contri- even if we believed that a portion of potentially even worse? bution plans in the private sector has the equity premium were a “freebie” Adding to the puzzle is that these increased the mobility of workers, and not just a compensation for risk, reforms have taken on numerous because traditional defined-benefit higher returns could be earned by a shapes and sizes across the world. plans have a “lock-in effect” that dis- public pension system by investing in While, politically, the adoption of per- courages employees from switching equities, which has the added benefit sonal accounts are often linked to employers. Therefore, the conversion of potentially improving risk sharing demographic concerns (for example, of public pension plans to the defined- across generations.1 National saving retirement of baby boomers in the contribution model sometimes has also could be increased by pre-funding United States), the actual evidence been lumped together as part of this the traditional pension system. The does not seem to support this motive same “modernization” movement. traditional pension scheme also could for personal accounts. Indeed, the However, that explanation is problem- be complemented with a life insurance largest reforms occurred in less devel- atic, because public pension plans are payment upon death that replicated oped countries where future demo- the bequeath-ability aspect of personal graphic problems are projected to be * Smetters is a Research Associate in the accounts. the least severe, including Chile (1981), NBER’s Program on Public Economics and Indeed, in a deterministic setting, Columbia (1993), Peru (1993), Mexico an associate professor at The Wharton School the traditional public defined benefit (1997), Bolivia (1997), El Salvador at the University of Pennsylvania. His pro- pension systems in theory could (1998), and Kazakhstan (1998). In each file appears later in this issue. achieve the same economic objectives of these countries, the vast majority of as personal accounts. In the presence the final expected retirement benefit is

NBER Reporter Summer 2005 11 derived from income produced by of reform but simply a necessary by about 1.5 percent of wages. assets held in the new defined-contri- byproduct of securing a safer retire- These calculations, however, bution accounts. In contrast, more ment income independent of the sub- unrealistically and optimistically modest reforms occurred in countries stantial level of trust required by a pay- assume that people continue to work with higher per-capita incomes where as-you-go financed scheme. Personal and earn just as much as before the demographic problems are more retirement accounts in these countries change in fiscal policy. New empirical severe, including Switzerland (1985), probably would have been created evidence by Prescott (2004), though, the United Kingdom (1986), Denmark even without demographic concerns. suggests otherwise: he attributes the (1990), Australia (1992), Argentina In contrast, in developed coun- sizable difference in the average num- (1994), China (1995), Uruguay (1996), tries, where reforms have been smaller, ber of working hours per worker in the Hungary (1998), Sweden (1998), and previous downward benefit adjust- United States versus many European Poland (1999). These countries have ments and inequities, while existing, countries to the differences in tax adopted systems that blend defined have been less important. Instead, the rates, largely used to finance state- contribution accounts with a defined primary objective in these countries is based retirement benefits.5 Even using benefit. Some countries with the most to pre-fund future benefits since many a much smaller labor supply elasticity serious demographic problems, includ- of these countries face more severe than implied by his study, raising pay- ing Germany and Japan, have passed demographic problems. Partial pre- roll taxes could substantially reduce only minor reforms. So, why are the funding of future benefits avoids more household savings and output relative largest reforms appearing in countries drastic changes, whether benefit cuts to controlling the growth rate of Social with the least amount of demographic or tax increases, down the road. Security benefits.6 problems? However, the governments in these Personal accounts themselves, One central theme appears to countries are not trusted enough to however, don’t improve or worsen emerge that might help to explain properly save or invest the required Social Security’s financial outlook. these puzzles: the public pension con- additional resources. So, in developed Contrary to some proponents, person- versions appear to represent a funda- countries, the creation of personal al accounts don’t offer a “free lunch” mental distrust in the ability of the accounts is a byproduct of attempts to by reducing Social Security’s shortfall. government to provide secure retire- increase funding. But the reforms are At the same time, contrary to many ment resources.3 The exact nature of smaller because they are mainly moti- opponents, personal accounts, when the distrust, though, differs between vated by demographics. If these coun- designed similarly to President Bush’s developing and developed countries, tries faced no demographic pressures, recent proposal, don’t add an addition- consistent with the differences in the the incentive to partially privatize al burden to the Social Security system. magnitudes of pension reforms. would be greatly reduced. The appearance of “transition costs” Traditional pay-as-you-go pen- from creating personal accounts sim- sion systems require a significant The United States ply reflects the fact that the federal amount of trust between workers of cash-flow budget system fails to different generations — the so-called According to the 2005 Social account for the long-term liabilities in “social contract.” The median voter Security Trustees’ Report, the U. S. the nation’s entitlement programs.7 who supports a pay-as-you-go pension Social Security system currently faces a While the diversion of payroll taxes to system is typically a middle-aged work- shortfall equal to about $11.1 trillion, personal accounts increases the er: she has no incentive to support a which is equal to the present value of amount of debt held by the public, the system that might not be viable when all future projected benefits minus the unfunded obligations in the Social she is retired. present value of all future projected Security program are reduced by an In developing countries, where payroll taxes after subtracting the value equal amount in present value. The reforms have been the largest, the dis- of the trust fund.4 This shortfall is federal budget tracks the increase in trust in the government provision of equal to about 3.5 percent of future the debt but not the concomitant public pensions appears to be condi- payroll. The shortfall in the Medicare reduction in obligations. tioned on past downward movements program, including the commitments in the real value of benefits (often on general revenue, is about seven The Social Security Trust caused by inflation), misuse of retire- times larger. Absent benefit cuts, plac- ment resources, and other risks and ing Social Security and Medicare on a Fund inequities in the pre-reform public permanently sustainable course could Currently, the Social Security sys- pension systems. Workers in develop- hypothetically be achieved by increas- tem collects more in revenue than it ing countries don’t trust the govern- ing payroll taxes from their current rate pays in benefits. The excess is placed ment to run even a strict pay-as-you-go of 15.3 percent of wages to about 36.1 into the Social Security Trust Fund. system. Funded defined-contribution percent of wages — immediately and The Trustees currently project that accounts, therefore, give workers an forever. For each 5 years in which outlays will begin to exceed revenue independence from the government. action were delayed, the required around 2017, at which time the Trust The concomitant increase in the level immediate and permanent Fund will be tapped in order to pay of funding is not the primary objective hike increases by about 10 percent, or

12 NBER Reporter Summer 2005 benefits. The Trust Fund is projected time toward the size of government though, because, ironically, they pro- to become depleted around 2041, would tend to create an upward bias vide “transition relief ” in an effort to after which only about 74 percent of (positive correlation). When controls protect the value of previous contribu- benefits can be paid. for these factors are added to the tions. Partial transition relief, however, Central to the debate on how to regression model, the correlation turns can lead to Pareto gains. reform Social Security is whether the negative and statistically significant Upon extending the life-cycle Social Security trust fund really repre- even at the 2 percent level over the model to three or more periods, a sents a “store of value.” To be sure, entire sample period.8 In fact, the household’s accrued benefit — which the assets in the Trust Fund are “real” results suggest that the entire Trust is observable by the government — in the sense that the U.S. Treasury will Fund buildup since 1983 has been becomes a source for an efficient honor the claims made by the Social used to hide additional spending or tax implicit lump sum tax that can be used Security Administration. But the rele- reductions elsewhere in the budget. It to replace some future revenue that vant question is whether Congress appears that Trust Fund surpluses would have been collected from that really uses Trust Fund surpluses to failed to reduce the debt held by the household using a distorting labor reduce the amount of debt held public, public! . Equivalently, this implicit thereby increasing the government’s Probably the most suggestive wealth levy can afford participants a ability to pay future benefits. Or, does indication that Congress has routinely higher return on their future contribu- Congress simply use Trust Fund sur- used Trust Fund surpluses to mask tions, thereby reducing the effective pluses to hide additional spending or tax larger non-Social Security deficits is tax rate on their labor supply. A back- reductions elsewhere in the federal the evidence of a structural break of-the-envelope calculation suggests budget? Congress is potentially able to when unified budget accounting was that the efficiency gains for the United mask larger non-Social Security adopted in 1970. Before 1970, Trust States could exceed $1 trillion, deficits because the federal unified Fund surpluses certainly existed, although this calculation should be budget combines the Trust Fund sur- although they dwindled over time as interpreted with some caution because pluses with non-Social Security deficits Congress increased the generosity of it ignores the transaction and other even though, as a pure technicality, benefits during the 1970s. But there is costs associated with personal Social Security is officially “off budget.” no evidence that Congress used Trust accounts. If Trust Fund surpluses are part- Fund surpluses to hide larger deficits ly hiding fiscal looseness elsewhere in elsewhere in the budget before 1970. 1 If Social Security purchased equities the budget then diverting future Social The evidence only appears after the with its excess contributions, risk shar- Security surpluses to personal accounts adoption of the unified budget ing could be potentially improved might require more fiscal restraint on accounting scheme. across generations. See H. Bohn, policymakers in the future, that is, per- “Should the Social Security Trust Fund sonal accounts may be a superior “lock Social Security hold Equities? An Intergenerational box.” Keeping the money away from Welfare Analysis,” Review of Economic politicians, though, one risk for Privatization with Second Dynamics, 2 (3) (July 1999), pp. 666-97. another, namely, keeping the money Best Taxes It is possible to replicate this same away from pensioners before they retire. effect using a capital income tax with- Nonetheless, personal accounts would It is generally believed that allow- out direct government ownership. See likely improve the informational con- ing workers to divert a portion of their K. Smetters, “The Equivalence of the tent of government spending. pay-as-you-go payroll taxes to private Social Security’s Trust Fund Portfolio It is not hard to find Republican accounts would require levying a new Allocation and Capital Income Tax and Democrats who believe that tax in order to continue to pay the Policy,” NBER Working Paper No. Congress routinely “spends the Trust Social Security benefits of those 8259, April 2001. Fund” whenever Congress runs a uni- retired at the time of privatization. In 2 S. Nishiyama and K. Smetters, “Does fied deficit. But if Congress would other words, privatization simply sub- Social Security Privatization Produce have run the same “on budget” deficits stitutes one distorting tax for another, Efficiency Gains?” NBER Working (excluding Social Security) without producing no efficiency gains. This Paper, forthcoming. Trust Fund surpluses then the Trust conventional wisdom, though, is based 3 K. Smetters and C. Park, A Matter of Fund surpluses were actually saved. on the standard life-cycle model with Trust: Understanding A Matter of Trust: Empirically, over the passed six just two periods. Understanding Worldwide Public Pension decades, “on budget” (non-Social Recent research, using a multi- Conversion, in progress. Security) and “off budget” (Social period life-cycle model, has shown 4 The 2005 OASDI Trustees Report, Security) surpluses have been uncorre- how to privatize in a way that reduces Table IV, B6. lated, which appears to suggest that labor supply distortions to current and 5 E. Prescott, “Why Do Americans Trust Fund surpluses have not been future generations without hurting ini- Work More than Europeans?” Federal used to hide non-Social Security tial retirees — a Pareto improvement.9 Reserve Bank of Minneapolis Quarterly deficits. However, macroeconomic The two most common methods of Review, 28 (1) (July 2004), pp. 2-13. shocks and changes in attitudes over privatization fail to reduce distortions, 6 L.J. Kotlikoff, K. Smetters, and J.

NBER Reporter Summer 2005 13 Walliser, “Finding a Way Out of 2005. Trust Funds?” NBER Working Paper America’s Demographic Dilemma,” 8 K. Smetters, “Is the Social Security No. 10953, December 2004. NBER Working Paper No. 8258, April Trust Fund Worth Anything?” 9 K. Smetters, “Social Security 2001. American Economic Review (Papers and Privatization with Elastic Labor Supply 7 See the discussion in J. Gokhale and Proceedings), 94 (2) (May 2004) pp. and Second-Best Taxes,” NBER K. Smetters, “Measuring Social 176–81. Also see, S. Nataraj and J.B. Working Paper No. 11101, February Security’s Financial Problems,” NBER Shoven, “Has the Unified Budget 2005. Working Paper No. 11060, January Undermined the Federal Government

Behavioral Finance

Jeremy C. Stein*

Much of my research over the who examines a firm’s annual report price, she will tend to favor the growth last several years has been in the broad may focus on earnings per share, while strategy when the stock market is fol- area of behavioral finance. Some of ignoring much of the other material in, lowing a valuation model that pays this work investigates the beliefs of say, the footnotes. more attention to performance on the less-than-fully rational investors — the Of course, even people who use growth dimension. Conversely, it can valuation models they use, and the par- very simple models are likely to give up be rational for the stock market to ticular sources of information that on these models when they fare poor- weight observed growth measures they pay attention to. Another part ly — as the honest-accounting model more heavily when it is known that the focuses on the constraints that profes- is likely to have done in recent years — firm is following a growth strategy. sional arbitrageurs face because of the and to move on to alternatives. This two-way feedback between firms’ agency problems inherent in delegated Motivated by this idea, Harrison Hong business strategies and the market’s money management. Finally, a third and I study the implications of learn- valuation model can lead to purely strand explores the connection ing in an environment where the true intrinsic fluctuations in sales and out- between investor irrationality and cor- model of the world is multivariate, but put, creating excess volatility in these porate-finance outcomes. in which agents update only over the real variables even in the absence of class of simple univariate models.1 If a any external source of shocks. Simple Models particular simple model does a poor job of forecasting over a period of Local and Social Influences In attempting to make even the time, it is eventually discarded in favor most basic kinds of forecasts, we can of an alternative — yet equally simple on Investment Decisions find ourselves inundated with a stag- — model that would have done better A number of recent papers show gering amount of potentially relevant over the same period. This theory that investors tend to have a strong raw data. A large literature in psychol- makes several distinctive predictions. local bias in their portfolio choices. ogy suggests that people simplify such For example, it suggests that a high- This bias shows up not only as a pref- forecasting problems by focusing their priced glamour stock has particularly erence for domestic as opposed to for- attention on a small subset of the low conditional expected returns, and eign stocks, but perhaps more striking- available data. One powerful way to particularly high conditional volatility, ly as a preference for those domestic simplify is with the aid of a theoretical in the wake of recent bad news about stocks that are headquartered close by.3 model. A parsimonious model will fundamentals, because this high- While the existence of within-country focus the user’s attention on those price/bad-news configuration suggests local bias now seems to be incontro- pieces of information deemed to be that the potential for a “paradigm vertible, there is little evidence to date particularly relevant for the forecast at shift” among investors is elevated. regarding its equilibrium asset-pricing hand; the user will disregard the rest. In a related vein, Philippe Aghion implications. For example, an investor with an “hon- and I examine a setting in which a firm Hong, Jeffrey Kubik, and I est-accounting” model of the world can devote its efforts either to increas- explore these asset-pricing effects.4 We ing sales growth or to improving per- begin by constructing a variable we call unit profit margins, for example by * Stein is a Research Associate in NBER's 2 RATIO which, for any given region at Program on Corporate Finance and a profes- cutting costs. If the firm’s manager is any point in time, is equal to the aggre- sor of economics at . concerned with the current stock gate book value of all firms headquar-

14 NBER Reporter Summer 2005 tered in the region divided by the effects story, the impact of sociability the survival-of-the-fittest view.8 While aggregate income of all households is stronger in states where stock-mar- maintaining the premise that agency living in the region. Intuitively, RATIO ket participation rates are higher. considerations play a crucial role in the measures the supply of shares in a In the context of professional decision to open-end, I show that region relative to the potential demand money managers, we show that a given when there is also asymmetric infor- for these shares, so we expect it to mutual fund manager is more likely to mation about managerial quality, the have a negative impact on stock prices. buy (or sell) a particular stock in any end result may be a degree of open- The data support this hypothesis. For quarter if other managers in the same ending that is socially excessive. This is example, if one goes from the Census city are buying (or selling) that same because any one high-quality manger region with the highest value of stock.6 This pattern shows up even will be tempted to go open-end to sig- RATIO (the Middle Atlantic), to the when the fund manager and the stock nal confidence in his ability, and there- region with the lowest value (the Deep in question are located far apart, so it is by lure assets under management away South), holding all else equal, the distinct from anything having to do from his competitors. This business- implied increase in the stock price is on with local preference. This evidence stealing effect sets off a counterpro- the order of 8 percent. For smaller-cap- can be interpreted in terms of an epi- ductive race to be open-ended, with italization companies, the correspon- demic model in which investors spread both high-quality and low-quality man- ding number is roughly 15 percent. information about stocks to one agers ultimately being forced to open- Digging deeper, we find that our another by word of mouth. end in order not to lose their investors. results are intimately connected to Owen Lamont and I present regional variation in population densi- Limited Arbitrage by some empirical evidence which is con- ty. That is, regions with low population sistent with the idea that the open-end density — of which the Deep South is Open-End Funds nature of professional arbitrage firms an example — tend to have low values makes it difficult for them to buck The vast majority of profession- 9 of RATIO, which are associated with ally-managed investment vehicles aggregate mispricings. We examine higher stock prices. This is because (mutual funds, hedge funds, and the some basic data on the evolution of most of the variation in RATIO across like) are structured on an open-end (as aggregate short interest, both during regions is driven by the book value opposed to closed-end) basis, making the dot-com era and at other times in component, which is very sensitive to it possible for their clients to liquidate history. In a striking contrast to the population density. In other words, in shares on demand. Both theory and patterns seen in the cross-section of spite of low per-capita income, the evidence suggest that the open-end stocks, total short interest moves in a Deep South is associated with higher form imposes serious constraints on countercyclical fashion. For example, stock prices because of an “only- arbitrageurs, since they are exposed to short interest in NASDAQ stocks game- in-town” effect: any one com- the risk of withdrawals if they perform actually declines as the NASDAQ pany headquartered there faces rela- poorly in the short run. This risk can index approaches its peak. Moreover, tively little competition for local make it dangerous for them to put on this decline does not seem to reflect a investors’ dollars, because so few other trades that are attractive in a long-run substitution away from outright short- companies are headquartered there. sense, but where convergence to fun- selling and towards put options: the In related work, Hong, Kubik, damentals is unlikely to be either ratio of put-to-call volume displays the and I examine social influences on smooth or rapid.7 For example, open- same countercyclical tendency. One investor behavior, at the level of both end funds are unlikely to want to bet possible interpretation is that as stock individuals and professional money heavily against something like the dot- prices rise, funds specializing in short- managers. With respect to individuals, com bubble of the late 1990s. selling realize negative returns, experi- we hypothesize that their decision Given the obvious drawbacks, ence outflows of assets under manage- about whether to participate in the why is the open-end form so domi- ment, and thus have to cut back their stock market is influenced by social nant? One answer, in a survival-of-the- positions. More generally, the evidence interaction: in our model, any given fittest spirit, might be that open-end- also suggests that short-selling does “social” investor finds the market ing is an optimal response to agency not play a particularly helpful role in more attractive when more of his stabilizing the overall stock market. 5 problems. If a fund is set up on a peers participate. We test this theory closed-end basis, dispersed investors using data from the Health and have no recourse in the face of mana- Corporate Finance Retirement Study and find that social gerial misbehavior, and may see their households — those who interact with entire investment slowly eaten away. In Stock-market inefficiencies like their neighbors, or attend church — contrast, if the fund is open-end, the dot-com bubble are of particular are indeed substantially more likely to investors can liquidate their positions interest to economists to the extent invest in the market than non-social at the first sign of trouble, thereby that they influence the allocation of households, controlling for wealth, avoiding large losses attributable to real resources, such as corporate race, education, and risk tolerance. mismanagement or theft. investment. Malcolm Baker, Jeffrey Moreover, consistent with a peer- In a recent paper, I take issue with Wurgler, and I attempt to address this

NBER Reporter Summer 2005 15 question.10 We use a simple model to tion of investor inertia is that it “The Only Game in Town: Stock-Price outline the conditions under which improves the terms for the acquiring Consequences of Local Bias,” working corporate investment is in fact sensi- firm in a stock-for-stock merger, since paper. tive to non-fundamental movements in acquirer shares are placed in the hands 5 H. Hong, J.D. Kubik, and J. C. Stein, stock prices. The key prediction of the of investors who, independent of their “Social Interaction and Stock-Market model is that stock prices have a beliefs, do not resell these shares on Participation,” NBER Working Paper stronger impact on the investment of the open market. In the presence of a No. 8358, July 2001, and Journal of “equity dependent” firms — those downward-sloping demand curve, this Finance, (February 2004), pp. 137-63. firms that need external equity to leads to a reduction in price pressure, 6 H. Hong, J.D. Kubik, and J. C. Stein, finance marginal investments. Using an and hence to cheaper equity financing. “Thy Neighbor’s Portfolio: Word-of- index of equity dependence based on We develop a simple model to Mouth Effects in the Holdings and the work of Kaplan and Zingales illustrate this idea, and present sup- Trades of Money Managers,” NBER (KZ), we find strong support for this porting empirical evidence. Both indi- Working Paper No. 9711, May 2003, hypothesis.11 In particular, firms that vidual and institutional investors tend and Journal of Finance, forthcoming. rank in the top quintile of the KZ to hang on to shares granted them in 7 See A. Shleifer and R.W. Vishny, index have investment that is almost mergers, with this tendency being “The Limits of Arbitrage,” Journal of three times as sensitive to stock prices much stronger for individuals. Finance, (March 1997), pp. 35-53. as firms in the bottom quintile. Consistent with the model and with 8 J.C. Stein, “Why Are Most Funds In another corporate-finance this cross-sectional pattern in inertia, Open End? Competition and the application, Baker and I try to under- acquirers targeting firms with high Limits of Arbitrage,” NBER Working stand how non-financial managers institutional ownership experience Paper No. 10259, January 2004, and might be able to successfully time the more negative announcement effects Quarterly Journal of Economics, (February market for seasoned equity offerings and greater announcement volume. 2005), pp. 247-72. (SEOs) even without access to any pri- Moreover, the results are strongest 9 O.A. Lamont and J. C. Stein, vate information or special insight when the overlap in target and acquir- “Aggregate Short Interest and Market about future stock returns.12 We build a er institutional ownership is low and Valuations,” NBER Working Paper model in which increases in liquidity - when the demand curve for the acquir- No. 10218, January 2004, and American such as lower bid-ask spreads, a lower er’s shares appears to be steep. Overall, Economic Review, (May 2004), pp. 29-32. price impact of trade, or higher turn- this framework may be helpful in 10 M. Baker, J. C. Stein, and J. Wurgler, over - predict lower subsequent returns explaining why stock-for-stock merg- “When Does the Market Matter? Stock in both firm-level and aggregate data. ers are a more significant source of Prices and the Investment of Equity- The model features a class of irrational equity finance for many firms than Dependent Firms,” NBER Working investors, who underreact to the infor- SEOs: with SEOs, unlike with merg- Paper No. 8750, February 2002, and mation contained in order flow, thereby ers, there is no scope for placing shares Quarterly Journal of Economics, (August boosting liquidity. In the presence of with inertial investors, so the adverse 2003), pp. 969-1005. short-sales constraints, high liquidity is price impact associated with issuance is 11 S. Kaplan and L. Zingales, “Do a symptom of the fact that the market likely to be more pronounced. Investment-Cash Flow Sensitivities is dominated by these irrational Provide Useful Measures of Financing investors, and hence is overvalued. If 1 H. Hong and J. C. Stein, “Simple Constraints?” Quarterly Journal of managers do nothing more than simply Forecasts and Paradigm Shifts,” NBER Economics, (February 1997), pp. 169- follow a rule of thumb that involves Working Paper No. 10013, October 215. issuing when the SEO market is partic- 2003. 12 M. Baker and J.C. Stein, “Market ularly liquid — perhaps because their 2 P. Aghion and J. C. Stein, “Growth vs. Liquidity as a Sentiment Indicator,” investment bankers find it easier to Margins: Destabilizing Consequences NBER Working Paper No. 8816, underwrite issues at such times — then of Giving the Stock Market What it February 2002, and Journal of Financial their financing decisions will tend to Wants,” NBER Working Paper No. Markets (June 2004), pp. 271-99. forecast aggregate stock returns. 10999, December 2004. 13 M. Baker, J. Coval, and J.C. Stein, Baker, Joshua Coval, and I argue 3 See, for example, J. Coval and T. “Corporate Financing Decisions that a bias toward inertial behavior on Moskowitz, “Home Bias at Home: When Investors Take the Path of the part of investors — a tendency to Local Equity Preference in Domestic Least Resistance,” NBER Working take the path of least resistance — can Portfolios,” Journal of Finance, Paper No. 10998, December 2004. have significant consequences for cor- (December 1999), pp. 2045-73. porate financial policy.13 One implica- 4 H. Hong, J.D. Kubik and J. C. Stein,

16 NBER Reporter Summer 2005 NBER Profile: Mark V. Pauly

Mark V. Pauly is a Research the influences that determine whether Associate in the NBER’s Program on insurance coverage is available and, Health Care and the Bendheim through several cost effectiveness stud- Professor in the Department of Health ies, the influence of medical care and Care Systems at the Wharton School of health practices on health outcomes and the University of Pennsylvania.. Also at cost. His interests in health policy deal the Wharton School, Pauly is a with ways to reduce the number of Professor of Health Care Systems, uninsured through tax credits for public Insurance and Risk Management, and and private insurance, and appropriate Business and Public Policy; he is a design for Medicare in a budget-con- Professor of Economics in the School strained environment. of Arts and Sciences at the University of Pauly is co-editor-in-chief of the Pennsylvania as well. Pauly holds a International Journal of Health Care Finance Ph.D. in economics from the University and Economics and an associate editor of of Virginia. the Journal of Risk and Uncertainty.He A former member of the Physician has served on Institute of Medicine Payment Review Commission and an panels on public accountability for active member of the Institute of health insurers under Medicare and on Medicine, Pauly’s work on the econom- improving the financing supply of vac- ics of moral hazard was the first to cines. He is also a former member of point out how health insurance coverage the advisory committee to the Agency may affect patients’ use of medical serv- for Health Care Research and Quality, ices. His subsequent work has explored and most recently a member of the the impact of conventional insurance Medicare Technical Advisory Panel. coverage on preventive care, on outpa- Pauly’s major hobby is “keeping the tient care, and on prescription drug use rate of renovation of a 120-year-old in managed care. He is currently studying house ahead of the rate of deterioration.” the effect of poor health on worker pro- ductivity. In addition, he has explored

NBER Profile: Carmen M. Reinhart

Carmen M. Reinhart is a Research contagion. Her work has been published Associate in the NBER’s Program on in leading scholarly journals, including International Finance and Macro- the American Economic Review, the Journal economics and a Professor of Econo- of Political Economy, the Quarterly Journal mics at the School of Public Policy and of Economics, and the Journal of Economic the Economics Department of the Perspectives and featured in the financial University of Maryland. She received her press, including The , The Ph.D. from in 1988. Financial Times, The Washington Times, and Reinhart held positions as Vice The Wall Street Journal. She also serves on President at the investment bank Bear the editorial board of the American Stearns and, more recently, as Deputy Economic Review. Director at the Research Department of Reinhart lives in Washington, D.C. the International Monetary Fund. She with her husband, an economist who is has written and published on a variety of Director of the Division of Monetary topics in macroeconomics and interna- Affairs at the Federal Reserve Board, tional finance and trade including: capi- and teenage son. Her hobbies include tal flows to developing countries; capital cooking, watching old movies, and won- controls; inflation stabilization; balance dering why her family ever adopted a of payments and banking crises; and basset hound.

NBER Reporter Summer 2005 17 NBER Profile: Kent Smetters

Kent Smetters is a Research academic year. Associate in the NBER’s Programs on Smetters was appointed Deputy Public Economics and Aging and an Assistant Secretary for Economic Policy associate professor at The Wharton of the U.S. Treasury on July 3, 2001, and School at the University of served in that position until August 30, Pennsylvania. He received his Ph.D. in 2002. He remains active in Washington, economics in 1995 from Harvard DC, and recently served as a member of University, and then worked for the U.S. the Blue Ribbon Panel on Dynamic Congress from 1995-8 before coming to Scoring for the Joint Committee on the University in Pennsylvania as an Taxation of the U.S. Congress. He has assistant professor. He was also the written extensively on Social Security Kaiser Visiting Professor of Economics reform. at Stanford University during the 2000-1

Conferences

Innovation Policy and the Economy

The NBER’s sixth annual con- Adam Jaffe and Josh Lerner, Richard B. Freeman, NBER and ference on Innovation Policy and the “Innovation and its Discontents” Harvard University, “Does Economy took place in Washington Globalization of the on April 19. The conference was Fiona Scott Morton, NBER and Scientific/Engineering Workforce organized by NBER Research , “Consumer Benefit Threaten U.S. Economic Associates Adam B. Jaffe of from Use of the Internet” Leadership?”(NBER Working Paper Brandeis University, Joshua Lerner No. 11457) of Harvard University, and Scott Ernst Berndt, NBER and MIT, Stern of Northwestern University. and Adrian Gottschalk and Richard Gilbert, University of The following papers were discussed: Matthew W. Strobeck, MIT, California, Berkeley, “Market “Opportunities for Improving the Structure and Innovation – What Dennis A. Yao, James J. Anton, Drug Development Process: Results Do We Know?” and Hillary J. Greene, Harvard from a Survey of Industry and the University, “The Policy Implications FDA” of Weak Patent Rights”

Yao, Anton, and Greene explore incentives and entrepreneurial activity. infringement, reduce the amount of a range of work on the economic and Weak patents cause firms to rely more knowledge that becomes publicly dis- policy implications of “weak” patents heavily on secrecy. Thus, the competi- closed, and potentially reduce the — patents that have a significant prob- tive environment is characterized by incentives to innovate. Further, weak ability of being overturned or are rela- private information about the extent of patent rights increase the likelihood of tively easy to circumvent — for inno- the innovator’s know-how. In such an patent litigation over commercially vation and disclosure incentives, environment, weak patents increase valuable patents and raise the specter antitrust policy, and organizational the likelihood of imitation and of anticompetitive settlements. Finally,

18 NBER Reporter Summer 2005 there are some implications for weak ability of products and information. countries. Third, populous low-income property rights in settings involving These gains are particularly useful to countries such as China and India can employee-inventors and employee mis- people with high transactions costs compete with the United States in use of confidential information. In the (busy, rural) and uninformed people. high-tech industries by having many former case, an increase in the strength Of course there are existing and S&E specialists even though those of some legal property rights such as potential attempts by firms to hold on workers are a small proportion of the patents is a force that reduces the to their profits in the face of con- country’s work forces. This threatens employer’s ability to prevent employ- sumers’ lowered search and transaction to undo the “North-South” pattern of ees from leaving with valuable know- costs. Corporate responses include trade in which advanced countries how, in part because stronger property lobbying for legal protection, altering dominate high tech while developing rights increase the value of the product design, restricting the infor- countries specialize in less skilled man- employee’s start-up option. In the lat- mation shared with consumers (obfus- ufacturing. Finally, diminished compar- ter case, an increase in legal penalties cation), and engaging in differential ative advantage in high-tech will create for breach has the expected effect of pricing. a long period of adjustment for U.S. decreasing breach of confidentiality. Berndt, Gottschalk, and workers, of which the offshoring of In the last two decades, the role Strobeck focus on the interactions IT jobs to India, growth of high-tech of patents in the U.S. innovation sys- between the Food and Drug Admini- production in China, and multination- tem has changed profoundly. Two stration (FDA) and the biopharmaceutical al R and D facilities in developing apparently mundane changes in patent companies that perform drug R and D. countries, are harbingers. law and policy have subtly but inex- To assess the current issues and the The effect of market structure on orably transformed the patent system state of communication and interac- innovation incentives has been a con- into a competitive weapon to attack tion between the FDA and industry, troversial subject in economics since rivals, thereby increasing the cost and the authors carried out a survey of Joseph Schumpeter advanced the theo- risk of innovation. Jaffe and Lerner industry leadership in R and D and ry that competitive markets are not review the changes in patent policy regulatory positions, as well as senior necessarily the most effective organiza- that have led to the current situation. leadership at the FDA with responsi- tions to promote innovation. The Then they highlight several principles bility for drug evaluation and over- incentive to innovate is the difference at the heart of what any reform of the sight. Based on 49 industry and 8 FDA in profits before and after innovation patent system ought to be seeking to interviews, the authors find that indus- occurs. The concept is straightforward, accomplish, as well as the limitations try seeks additional structured and yet differences in market structure, the that reform efforts are likely to face. informal interactions with the FDA, characteristics of innovations, and the Scott Morton discusses the especially during Phase II of develop- dynamics of discovery leads to seem- sources of consumer surplus that are ment. Overall, industry places greater ingly endless variations in the theoreti- likely to exist because of the types of value on additional communication than cal relationship between competition Internet sites being used. She also dis- the FDA. Furthermore, industry inter- and expenditures on research and cusses the problems involved in meas- viewees indicate that they are willing to development or the outputs of R and uring all the gains from use of the pay fees during clinical development to D. Gilbert surveys the economic theo- Internet. Web sites that make tradition- ensure that the FDA can hire additional, ry of innovation, focusing on horizon- al sales generate consumer surplus well-qualified staff to assist with proto- tal market structure, the distinction through availability, variety, and con- col reviews and decisionmaking. between product and process innova- venience to the consumer. Price com- Freeman presents four proposi- tions, and the role of exclusive and parison sites allow consumers to quick- tions showing that changes in the glob- non-exclusive right to innovation, and ly and easily gather price quotes from a al job market for science and engineer- attempts to draw conclusions from the variety of sellers, which results in the ing (S&E) workers are eroding U.S. different models. Exclusive rights gen- consumer paying a lower price. dominance in S&E, which diminishes erally lead to greater innovation incen- Information sites provide information the country’s comparative advantage in tives in more competitive markets, which the consumer can use to pick an high-tech goods and services and while non-exclusive rights generally appropriate activity or execute a task threatens the country’s global econom- lead to the opposite conclusion, more efficiently; often these sites save ic leadership. First, the U.S. share of although there are important excep- consumers time in mundane tasks such the world’s S&E graduates is declining tions. Gilbert reviews the large litera- as buying tickets, checking the weather, rapidly as European and Asian univer- ture on empirical studies of innova- or getting driving directions. Likewise, sities, particularly from China, have tion, and argues for greater reliance on matching sites (such as eBay) improve granted more S&E degrees. Second, industry-specific case studies. transactions by hugely increasing the the U.S. job market has worsened for These papers will appear in an quality of the match compared to the young S&E worker and engineers rela- annual volume published by the MIT local garage sale. While it’s clear that tive to many other high-level occupa- Press. Its availability will be announced the Internet increases price competi- tions, which discourages U.S. students in a future issue of the Reporter. They tion, so that consumers pay less for from going on in these fields, but can also be found at “Books in products, it also improves daily life by which still attracts large immigrant Progress” on the NBER’s website. increasing the variety, quality, and avail- flows, particularly from developing NBER Reporter Summer 2005 19 Asset Pricing with Imperfect Trading

An NBER-Universities Research of North Carolina at Chapel Hill Discussant: Kent D. Daniel, NBER Conference on “Asset Pricing with and Northwestern University Imperfect Trading” was held in Robert Novy-Marx, University of Cambridge on May 13 and 14. NBER Chicago, “On the Excess Returns to Antti Petajisto, Yale University, Research Associates Darrell Duffie of Illiquidity” “Why Do Demand Curves for Stanford Business School and Owen Discussant: Robert F. Stambaugh, Stocks Slope Down?” A. Lamont of Yale School of NBER and University of Discussant: Arvind Krishnamurthy, Management organized this program: Pennsylvania Northwestern University Lauren B. Cohen, University of Nicolae Garleanu, University of Sandro C. Andrade and Mark S. Chicago; Karl B. Diether, Ohio Pennsylvania; Lasse Heje Seasholes, University of California, State University; and Christopher J. Pedersen, New York University; Berkeley, and Charles Chang, Malloy, London Business School, and Allen M. Poteshman, Cornell University, “Trading Shocks, “Supply and Demand Shifts in the University of at Urbana- Asset Prices, and the Limits of Shorting Market” Champaign; “Demand-Based Arbitrage” Discussant: Tyler Henry, University Option Pricing” Discussant: Malcolm Baker, NBER of Washington Discussant: Joshua D. Coval, NBER and Harvard University and Harvard University Ronnie Sadka, University of Andrea Frazzini, Yale University, Washington, and Anna Scherbina, Harrison Hong, Princeton and Owen Lamont, “Dumb Harvard University, “Analyst University, and Jialin Yu, Columbia Money: Mutual Fund Flows and the Disagreement, Mispricing, and University, “Gone Fishin’: Cross-Section of Stock Returns” Liquidity” Seasonality in Speculative Trading Discussant: Stefan Nagel, Stanford Discussant: Adam Reed, University and Asset Prices” University

Using proprietary data on stock finds that these stocks tend to be over- Garleanu, Pedersen, and loan fees and stock loan quantities priced, but prices correct downward as Poteshman model the demand-pres- from a large institutional investor, uncertainty about earnings is resolved. sure effect on prices when options Cohen, Diether, and Malloy examine The authors conjecture that one reason cannot be hedged perfectly. The model the link between the shorting market mispricing has persisted is that analyst shows that demand pressure in one and stock prices. Using a unique iden- disagreement coincides with high trad- option contract increases its price by tification strategy, they are able to clas- ing costs. Indeed, they show that in the an amount proportional to the vari- sify shifts in the supply and demand cross-section the less liquid stocks tend ance of the unhedgeable part of the for shorting. They find that shorting to be more severely overpriced. Also, option. Similarly, the demand pressure demand is an economically and statisti- increases in aggregate market liquidity increases the price of any other option cally important predictor of future accelerate convergence of prices to by an amount proportional to the stock returns. The magnitude of this fundamentals. As a result, returns of the covariance of their unhedgeable parts. effect is large: an increase in shorting initially overpriced stocks are negatively Empirically, the authors identify aggre- demand leads to negative abnormal correlated with the time series of inno- gate positions of dealers and end users returns of 2.54 percent in the follow- vations in aggregate market liquidity. using a unique dataset, and show that ing month. Further, they find that the Novy-Marx argues that the high demand-pressure effects help to shorting market is an important mech- expected returns observed on illiquid explain well-known option-pricing anism for revelation of private infor- assets should be expected theoretically puzzles. First, end users are net long mation into prices. Specifically, they but are not actually a premium for illiq- index options, especially out-of-money show that their results are stronger in uidity, per se. Instead, illiquidity — like puts, which helps to explain their environments with less public infor- size — is a proxy for any unobserved apparent expensiveness and the smirk. mation flow, and that net of shorting risk. Therefore, liquidity should have Second, demand patterns help to costs, trading strategies based on their explanatory power in any asset-pricing explain the cross section of prices and identification strategy yield over 3 per- model that is not perfectly specified: the skews of single-stock options. cent per month on average. low measured liquidity will forecast Hong and Yu develop and test a Sadka and Scherbina document high expected returns. The magnitude theory of seasonality in asset prices a close link between mispricing and liq- of the expected premium can be simi- based on the idea that speculative trad- uidity by investigating stocks with high lar to that associated with the omitted ing generates a price premium. The analyst disagreement. Previous research risk factor. authors hypothesize that trading of all

20 NBER Reporter Summer 2005 types, including speculative trades, stock prices are set instead by two sep- shock levels. The model predicts that declines when investors are away on arate classes of investors. Individual trading in one stock can affect prices of vacation but only the prices of specu- investors price the market portfolio other stocks. Empirically, the authors lative assets (those with abnormally based on their collective risk aversion. find support for the model and show high share turnover) will drop at the Individual investors also delegate part that the temporary price impact dou- same time. They test their hypothesis of their wealth to active money man- bles during periods when many stocks using data from the U.S. and Chinese agers who use that wealth to price experience similarly signed shocks. stock markets. As predicted, they find stocks in the cross-section. In equilib- Finally, trading shocks are able to that turnover in the U.S. stock market rium the fee charged by active man- explain approximately half of drops significantly in the summer agers has to equal the before-fee alpha observed co-movement in stock (when investors are gone fishin’) and they earn. This produces wide pricing returns. These results provide insights that the prices of dot-com and liquid, bounds for individual stocks and can into the limits of arbitrage and a mar- high turnover stocks are lowest during account for several empirically ket’s ability to absorb demand shocks. the summer when compared to other observed puzzles, such as the magni- Frazzini and Lamont use mutu- stocks. In China, where investors go tude of the S and P 500 index premium. al fund flows as a measure of individ- on vacation only during the Chinese Andrade, Chang, and Seasholes ual investor sentiment for different New Year (January and February), examine the relationship between trad- stocks. They find that high sentiment turnover and the prices of speculative ing shocks and asset prices. They out- predicts low future returns at long stocks bottom out during the first two line a simple market-clearing model in horizons. Fund flows are dumb months of the year. The authors rule which some investors place orders that money: by reallocating across different out alternative explanations, such as are uncorrelated with asset fundamen- mutual funds, retail investo rs reduce seasonal variations in trading costs. tals. Their empirical results document their wealth in the long run. This Representative agent models, pervasive price pressure at a daily, week- dumb money effect is strongly posi- including the Capital Asset Pricing ly, and monthly frequency. The authors tively related to the value effect. High Model, are not consistent with existing sort stocks into deciles based on trading sentiment also is associated with high empirical evidence for steep demand shocks. Initial distortions in the extreme corporate issuance, interpreted as curves for individual stocks. Petajisto deciles are over 200bp and prices take companies increasing the supply of resolves the issue by proposing that five weeks to mean-revert back to pre- shares in response to investor demand.

*

NBER Reporter Summer 2005 21 Current Account Imbalances

NBER Research Associate Fund; Dirk Muir, Norges Bank; Discussant: Edwin Truman, Institute Richard H. Clarida of Columbia and Paolo Pesenti, Federal Reserve for International Economics University organized an NBER Bank of New York, ”Smooth Conference on “G7 Current Account Landing or Crash? Model-Based , NBER and Imbalances: Sustainability and Scenarios of Global Current University of California, Berkeley, Adjustment”, which was held on June Account Rebalancing” and Kenneth S. Rogoff, NBER 1 and 2. These papers were discussed: Discussant: Lars E.O. Svensson, and Harvard University, “The NBER and Princeton University Unsustainable U.S. Current Account Richard H. Clarida, and Manuela Position Revisited” (NBER Working Goretti and Mark Taylor, Aart Kraay, World Bank, and Paper No. 10869) University of Warwick, “Are There Jaume Ventura, NBER and Discussant: Kristin Forbes, NBER Thresholds of Current Account Universitat Pompeu Fabra, “ The and MIT Adjustment in the G7?” Dot-Com Bubble, the Bush Deficits, Discussant: Robert E. Cumby, and the U.S. Current Account” Muge Adalet, Victoria University, NBER and Georgetown University Discussant: Joseph Gagnon, Federal Wellington, and Barry Reserve Board Eichengreen, NBER and Pierre-Olivier Gourinchas, NBER University of California, Berkeley, and University of California, Michael P. Dooley, NBER and “Current Account Reversals: Always Berkeley, and Helene Rey, NBER University of California, Santa Cruz; a Problem?” and Princeton University, “From David Folkerts-Landau, Deutsche Discussant: Frederic S. Mishkin, World Banker to World Venture Bank; and Peter Garber, NBER NBER and Columbia University Capitalist: The U. S. External and Brown University, “ Direct Adjustment and the Exorbitant Investment, Rising Real Wages, and Catherine L. Mann and Katharina Privilege” the Absorption of Excess Labor in Plueck, Institute for International Discussant: Jose De Gregorio, the Periphery” (NBER Working Economics, “ The U.S. Trade Central Bank of Chile Paper No. 10626) Deficit: A Disaggregated Discussant: Shang-Jin Wei, NBER Perspective” Philip R. Lane, Trinity College and IMF Discussant: Peter Kenen, Princeton Dublin, and Gian Maria Milesi- University Ferretti, International Monetary Menzie D. Chinn, NBER and Fund, “A Global Perspective on University of Wisconsin, and Caroline Freund, World Bank, and External Positions” Jeffrey A. Frankel, NBER and Frank Warnock, University of Discussant: Richard Portes, NBER Harvard University, “Will the Euro Virginia, “ Current Account Deficits and London Business School Eventually Surpass the Dollar as in Industrial Countries: The Bigger Leading international Reserve They Are, The Harder They Fall?” Hamid Faruqee and Douglas Currency?” (NBER Working Paper Discussant: Assaf Razin, NBER and Laxton, International Monetary No. 11510) Tel Aviv University

Clarida, Goretti, and Taylor find adjustment regimes identified by the collapse of the Bretton Woods fixed evidence of threshold behavior in cur- model. exchange rate system. It is mainly rent account adjustment for the G7 Does the center country of the attributable to a “return discount”: countries, such that the dynamics of International Monetary System enjoy within each class of assets, the total adjustment towards equilibrium an “exorbitant privilege” that signifi- return (yields and capital gains) that depend upon whether the current- cantly weakens its external constraint, the United States has to pay to for- account/ net output ratio breaches as has been asserted in some European eigners is smaller than the total return estimated, country-specific current quarters? Using a newly constructed the United States gets on its foreign account surplus or deficit thresholds. dataset, Gourinchas and Rey perform assets. There is also evidence of a Both the speeds of adjustment and the a detailed analysis of the historical evo- “composition effect”: the United size of the thresholds differ signifi- lution of U.S. external assets and liabil- States tends to borrow short and lend cantly across countries. In addition, ities at market value since 1952. They long. As financial globalization acceler- there are shifts found in the means and find strong evidence of a sizeable ated its pace, the United States trans- variances of exchange rate changes, excess return of gross assets over formed itself from a World Banker stock returns, and interest differentials gross liabilities. Interestingly, this into a World Venture Capitalist, invest- that coincide with the current account excess return has increased after the ing greater amounts in high yield assets

22 NBER Reporter Summer 2005 such as equity and FDI. The authors growth. The adoption of a flexible labor into the industrial sector. The use these findings to cast some light on exchange rate regime in emerging Asia principal constraints are inefficient the sustainability of the current global can help reduce variability in both domestic resource and capital markets, imbalances. regional output and inflation variabili- and resistance to import penetration by Lane and Milesi-Ferretti high- ty. Other simulations consider the labor in industrial countries. The system light the increased dispersion in net effects of fiscal adjustment in the has evolved to overcome these con- external positions in recent years, par- United States, as well as the effects of straints through -led growth and ticularly among industrial countries. growth-enhancing reforms in Europe growth of foreign direct investment. They provide a simple accounting and Japan. Periphery governments’ objectives for framework that disentangles the fac- Over the past decade the United the scale and composition of gross tors driving the accumulation of exter- States has experienced widening cur- trade in goods and financial assets may nal assets and liabilities (such as trade rent account deficits and a steady dete- dominate more conventional concerns imbalances, investment income flows, rioration of its net foreign asset posi- about international capital flows. and capital gains) for major external tion. During the second half of the Might the dollar eventually follow creditors and debtors. They also exam- 1990s, this deterioration was fueled by the precedent of the pound and cede ine the factors driving the foreign asset foreign investment in a booming U.S. its status as leading international portfolio of international investors, stock market. During the first half of reserve currency? To answer the ques- with a special focus on the weight of the 2000s, this deterioration has been tion, Chinn and Frankel econometri- U.S. liabilities in the rest of the world’s fuelled by foreign purchases of rapidly cally estimate determinants of the stock of external assets. Finally, they increasing U.S. government debt. A shares of major currencies in the relate the empirical evidence to the somewhat surprising aspect of the cur- reserve holdings of the world’s central current debate about the roles of port- rent debate is that stock market move- banks. Significant factors include: size folio balance effects and exchange rate ments and fiscal policy choices have of the home country, inflation rate (or adjustment in shaping the external been treated largely as unrelated lagged depreciation trend), exchange adjustment process. events. Stock market movements usu- rate variability, and size of the relevant Faruqee, Laxton, Muir, and ally are interpreted as reflecting exoge- home financial center (as measured by Pesenti re-examine the issues sur- nous changes in perceived or real pro- the turnover in its foreign exchange rounding global rebalancing through ductivity, while budget deficits usually market). The authors do not find that the lens of a dynamic, multi-region are understood as a mainly political net international debt position is an model of the world economy. The decision. Kraay and Ventura chal- important determinant. They do find authors first define a baseline scenario lenge this view and develop two alter- that the relationship between currency of global adjustment, based on the native interpretations. Both are based shares and their determinants is non- premise that the world macroeconom- on the notion that a bubble (the “dot- linear, but changes are felt only with a ic imbalances of the early 2000s can be com” bubble) has been driving the long lag. The advent of the euro inter- attributed to a combination of several stock market, but they differ in their rupts the continuity of the historical related but distinct tendencies: a nega- assumptions about the interactions dataset. So, Chinn and Frankel esti- tive U.S. fiscal shock, a decrease in the between this bubble and fiscal policy mate parameters on pre-1999 data and U.S. private savings rate, an increase in (the “Bush” deficits). The “benevo- then use that data to forecast the EMU the demand for U.S. assets abroad, par- lent” view holds that a change in era. Their equation correctly predicts a ticularly in Asia, a positive shock to investor sentiment led to the collapse (small) narrowing in the gap between productivity growth in emerging Asia, of the dot-com bubble and the Bush the dollar and euro over the period and a negative shock to productivity deficits were a welfare-improving poli- 1999-2004. Whether the euro might growth in Japan and the Euro Area. cy response to this event. The “cyni- rival or surpass the dollar as the world’s The baseline scenario projects a build- cal” view holds instead that the Bush leading international reserve currency up of government debt and a decline deficits led to the collapse of the dot- in the future appears to depend on two in net foreign assets in the United com bubble as the new administration things: 1) do the United Kingdom and States. A gradual depreciation of the tried to appropriate rents from foreign enough other EU members join exchange rate allows the U.S. current investors. The authors discuss the euroland so that it becomes larger than account deficit to move to a sustain- implications of each of these views for the U.S. economy, and 2) does U.S. able level in 10 years time. The authors the future evolution of the U.S. econo- macroeconomic policy eventually then assess a few alternative scenarios my and, in particular, its net foreign undermine confidence in the value of relative to the baseline. A sudden port- asset position. the dollar, in the form of inflation and folio reshuffling in the rest of the Dooley, Folkerts-Landau, and depreciation? According to Chinn and world results in higher real interest Garber set out the political economy Frankel, under two important scenar- rates in the United States and a signifi- behind Asian governments’ participa- ios – that the remaining EU members, cant depreciation of the dollar in tion in a revived Bretton Woods including the United Kingdom, join effective terms. The trade balance System. The overriding problem for EMU by 2020, or that the recent improves significantly but there is a these governments is rapidly integrat- depreciation trend of the dollar per- large contractionary effect on U.S. ing a large pool of underemployed sists into the future – the euro may sur-

NBER Reporter Summer 2005 23 pass the dollar as leading international dence of reversals has been unusually consistent with earlier work showing reserve currency by 2022. great in recent years. The only prior that, in general, current account adjust- Obstfeld and Rogoff show that period that matched the last three ment tends to be associated with slow when one takes into account the glob- decades in terms of frequency was the income growth and a real depreciation. al equilibrium ramifications of an 1920s and 1930s, two decades notori- Overall, the results support claims that unwinding of the U.S. current account ous for the instability of capital flows. the size of the current account deficit deficit, currently running at nearly 6 In contrast, reversals were less com- and the extent to which it is financing percent of GDP, the potential collapse mon in the Bretton Woods and pre- consumption matter for adjustment. of the dollar becomes considerably World War I gold standard eras. The Mann and Pluck prepare new larger (more than 50 percent larger) fact that reversals were relatively few, estimates of the elasticity of U.S. trade than in their previous estimates. It is small, and undisruptive under the gold flows using bilateral trade data for 31 true that global capital market deepen- standard is striking, given the absence countries, different measures of ing appears to have accelerated over of impediments to capital flows and expenditure, and including alternative the past decade and that this deepen- the large size of current account bal- proxies for global supply-cum-variety. ing may have helped the United States ances. Detailed analysis of particular They examine four different categories to allow a record-breaking string of cases points to the degree of market of goods based on the Bureau of deficits. Unfortunately, though, global flexibility and to how reversals were Economic Analysis’s “end-use” classi- capital market deepening turns out to managed as part of the explanation for fication system: autos, industrial sup- be of only modest help in mitigating these patterns. plies and materials (excluding energy), the dollar decline that almost inevitably There are a number of worri- consumer goods, and capital goods. will occur in the wake of global cur- some features of the U.S. current These four categories behave differ- rent account adjustment. As their ear- account deficit. In particular, its size ently from a panel aggregated over the lier analysis showed, and the model of and persistence, the extent to which it full commodity set. The authors find this paper reinforces, adjustments to is financing consumption as opposed that industrial and developing coun- large current account shifts depend to investment, and the reliance on debt tries have different income and relative mainly on the flexibility and global inflows raise concerns about the likeli- price elasticities for these four product integration of goods and factor mar- hood of a sharp adjustment. Freund groups. The relative prices for the kets. Whereas the dollar’s decline may and Warnock examine episodes of industrial countries have plausible be benign, as in the 1980s, Obstfeld current account adjustment in industri- parameter values. Variety is an impor- and Rogoff argue that the current con- al countries to assess the validity of tant variable for capital goods. Whereas juncture more closely parallels the these concerns. They find that: 1) larg- the authors obtain more plausible val- 1970s, when the Bretton Woods sys- er deficits take longer to adjust and are ues for the trade elasticities, consistent tem collapsed. Finally, they use their associated with significantly slower with theoretical priors, their predictions model to dispel some common mis- income growth (relative to trend) dur- for imports are poor compared to the conceptions about what kinds of shifts ing the current account recovery than benchmark model that uses U.S. GDP are needed to help close the U.S. cur- smaller deficits; 2) neither persistent as the measure of expenditure. On the rent account imbalance. For example, deficits nor large net foreign debt posi- other hand, their new estimates do bet- faster growth abroad helps only if it is tions are accommodated by more ter at predicting than does the relatively concentrated in non-tradable extensive exchange rate adjustment or benchmark model that uses GDP as goods; faster productivity growth in for- slower growth; 3) consumption-driven the measure of expenditure. eign tradable goods will actually exacer- current account deficits involve signif- The University of Chicago Press bate the U.S. adjustment problem. icantly larger depreciations than will publish these papers and discus- Adalet and Eichengreen analyze deficits financing investment; and 4) sions in an NBER conference volume. the pre-1970 history of international the composition of the financing of Many of the papers also can be found capital flows and current account the deficit is orthogonal to exchange at “Books in Progress” on the NBER’s reversals, contrasting it with recent rate movements and income growth website. experience. They find that the inci- during adjustment. These findings are

24 NBER Reporter Summer 2005 International Seminar on Macroeconomics

The NBER’s 28th Annual Marianne Baxter and Robert Market-Based Macro Forecast, International Seminar on Macro- King, NBER and Boston Uncertainty, and Risk” economics (ISOM), organized by University, “Fiscal Externalities and Discussants: Christopher Carroll, Jeffrey A. Frankel, NBER and Optimal Taxation in an Economic NBER and Johns Hopkins Harvard University, and Christopher Community” University, and Adam Szeidl, Pissarides, London School of Discussants: Pierpaolo Benigno, University of California, Berkeley Economics, was held on June 17-18 NBER and New York University, at the Magyar Nemzeti Bank (Central and Francesco Giavazzi Zsolt Darvas and Gyorgy Bank of Hungary) in Budapest. Szapary, Magyar Nemzeti Bank, Francesco Giavazzi, NBER and Balazs Vilagi, Magyar Nemzeti and Andrew K. Rose, NBER and Bocconi University, serves as co- Bank, “ Dual Inflation and the Real University of California, Berkeley, chair of ISOM with Frankel. The Exchange Rate in New Open “Fiscal Divergency and Business following papers were discussed at Economy Macroeconomics” Cycle Synchronization: the conference: Discussants: Richard Clarida, NBER Irresponsibility is Idiosyncratic” and Columbia University, and Refet Discussants: Roberto Rigobon, Tito Boeri and Pietro Garibaldi, Gurkaynak, Federal Reserve Board NBER and MIT, and Lucrezia Bocconi University, “Shadow Reichlin, Sorting” Prakash Loungani, IMF, and Discussants: Robert E. Hall, NBER Assaf Razin, NBER and Tel Aviv Linda S. Goldberg, Federal and Stanford University, and University, “ Globalization and Reserve Bank of New York, “Trade Christopher Pissarides Disinflation: The Efficiency Invoicing in the Accession Channel” Countries: Are They Suited to the Yann Algan, Universite Marne la Discussants: Robert King, and Euro?” Vallee, and Pierre Cahuc, Philip Lane, Trinity College Dublin Discussants: , NBER Universite Paris 1, “ The Roots of and University of Wisconsin, and Low European Employment: Family Refet S. Gurkaynak, and Justin Richard Portes, NBER and London Culture?” Wolfers, NBER and University of Business School Discussants: Tito Boeri, and Pennsylvania, “Macroeconomic Alessandra Fogli, New York University Derivatives: An Intitial Analysis of

In an equilibrium model of the model implies that shadow wage gaps individuals. Algan and Cahuc argue labor market with market frictions, should be lower in depressed labor that family labor supply interactions Boeri and Garibaldi investigate the markets and that deregulation of labor and cross-country heterogeneity in border between formal employment, markets is accompanied by a decline in family culture are key to explaining shadow employment, and unemploy- the average skills of the workforce in these stylized facts. The authors pro- ment. From the labor demand side, both legal and shadow sectors. Based vide a simple labor supply model in firms optimally create legal or shadow on microdata on two countries with a which heterogeneity in family prefer- employment through a mechanism sizeable shadow economy — Italy and ences can account for cross-country that is akin to . From the Brazil — Boeri and Garibaldi find variations in the level and the dynamics labor supply side, heterogeneous work- empirical support for these implica- of employment rates of demographic ers sort across the two sectors, with tions of the model. They suggest that groups. Next, they demonstrate — high productivity workers entering the policies aimed at reducing the shadow based on international individual sur- legal sector. This worker sorting is fully economy are likely to increase unem- veys — that family conceptions do dif- consistent with most empirical evi- ployment. Yet, from the job creation fer across countries and are shaped dence on shadow employment. The standpoint, income support for unem- largely by national features. They also model also sheds light on the “shadow ployment may be better than shadow show that cross-country differences in puzzle”: the increasing size of the repression. family culture cause cross-national dif- shadow economy in OECD countries OECD countries experienced ferences in family conception. despite improvements in technologies largely divergent employment rates Studying the correlation between for detecting tax and social security over the last decades. But the bulk of employment rates and family concep- evasion. Shadow employment is corre- the cross-national and cross-temporal tion, the authors then show that the lated with unemployment, and is toler- heterogeneity of unemployment relies stronger preferences for family activi- ated because the repression of shadow on specific demographic groups: ties in European countries may explain activity increases unemployment. The prime-age women, younger, and older both their lower female employment

NBER Reporter Summer 2005 25 rate and the fall in employment rates how globalization induces monetary probabilities. among youth and seniors. authorities, guided by the welfare crite- Using a panel of 21 OECD coun- The Stability and Growth Pact is rion of a representative household, to tries and 40 years of annual data, a continuing source of economic con- put greater emphasis on reducing the Darvas, Rose, and Szapary find that troversy within Europe. The Pact rec- inflation rate than on narrowing out- countries with similar government ognizes that individual member states put gaps. The authors demonstrate budget positions tend to have business experience divergent business cycle that, with capital account liberalization, cycles that fluctuate more closely. conditions that may lead them to run the representative household is able to That is, fiscal convergence (in the form deficits at certain points in time. smooth fluctuations in consumption, of persistently similar ratios of gov- However, the pact is designed to and thus becomes relatively insensitive ernment surplus/deficit to GDP) is encourage member states to adopt fis- to fluctuations in the output gap. With systematically associated with more cal policies that imply zero deficits on trade liberalization the economy tends synchronized business cycles. The average and to limit their deficits to 3 to specialize in the production of rela- authors also find that reduced fiscal percent of GDP at any point in time. tively few varieties of goods. The spe- deficits increase business cycle syn- Baxter and King study the nature of cialization in production resulting chronization. The Maastricht “conver- fiscal externalities within an economic from trade openness increases the dis- gence criteria,” used to determine eligi- community, such as Europe, that lacks tortion associated with fluctuations in bility for EMU, encouraged fiscal con- explicit rules for fiscal policy coordina- the inflation rate. Therefore, when vergence and deficit reduction. Thus tion, assuming that each country trade and financial openness increase, the crieria may have indirectly moved chooses its tax rates optimally given policymakers (guided by efficiency Europe closer to an optimum currency the fiscal stance of other countries. considerations) become more aggres- area, by reducing countries’ abilities to Allowing for real shifts to country pro- sive about inflation and less responsive create idiosyncratic fiscal shocks. These ductivity and public expenditure, the to the output gap. The authors pro- empirical results are economically and authors find that the fiscal deficit can vide evidence on the effect of trade statistically significant and robust. be a poor indicator of fiscal externali- and capital openness on preferences The accession countries to the ties: countries with different labor and towards fluctuations in the output gap euro area increasingly are binding their rates can exert exact- and in inflation, which support their economic activity — external and ly the same external effect but have theory’s predictions. internal — to the euro area countries. very different fiscal deficit behavior. In September 2002, a new market One direction of this phenomenon Trade deficits are, by contrast, much in “Economic Derivatives” was launched concerns the currency invoicing of more informative about the effects that allowing traders to take positions on international trade transactions, by an individual country has on other future values of several macroeconomic which accession countries have members of the community. data releases. Gurkaynak and Wolfers reduced their use of the U.S. dollar in Vilagi studies how the models of provide an initial analysis of the prices invoicing international trade transac- the new open economy macroeco- of these options. They find that expec- tions. Theory predicts that the optimal nomics, which usually focus on the tations derived from these market invoicing choices for accession coun- relationship between the nominal prices are somewhat more accurate tries depend on the composition of exchange rate and the external real than survey-based forecasts, and that goods in exports and imports and on exchange rate, can explain the coexis- they better predict financial market the macroeconomic fluctuations of tence of permanent dual inflation: responses to surprises in data. These trade partners, both bearing on the diverging inflation rates for tradable derivatives also provide market- role of herding and hedging consider- and non-tradable goods and real implied probabilities of the full range ations within exporter profitability. appreciation in emerging market of specific outcomes, allowing the These considerations yield country- economies. He shows that the impact authors to measure uncertainty, assess specific estimates about the optimal of asymmetric sectoral productivity its driving forces, and to compare this degree of Euro-denominated invoic- growth on the real exchange rate measure of uncertainty with the dis- ing of exports. Goldberg finds that depends heavily on the market struc- persion of point-estimates among the exporters of some accession coun- ture, and that the models of new open individual forecasters (a measure of tries, even in their trade transactions economy macroeconomics can be rec- disagreement). The authors also assess with the euro zone and other onciled with the Balassa-Samuelson the accuracy of market-generated European Union countries, might be effect only if pricing to market is probability density forecasts. One con- pricing a larger-than-optimal share of added to models. He also demon- sistent theme is that few of the behav- their transactions in euros rather than strates that the presence of nominal ioral anomalies present in surveys of in dollars, thus taking on excessive risk rigidities and frictions in capital accu- professional forecasts survive in equi- in international markets. mulation help to explain the apprecia- librium, and that these markets are The MIT Press will publish these tion of the external real exchange rate remarkably well calibrated. Finally, the papers in an annual conference volume and the slow adjustment of the relative authors assess the role of risk, finding later this year. They are also available at price of non-tradables to tradables. little evidence that risk aversion drives “Books in Progress” on the NBER’s Razin and Loungani analyze a wedge between market prices and website.

26 NBER Reporter Summer 2005 East Asian Seminar on Economics Focuses on Fiscal Policy and Management

The NBER’s Sixteenth Annual Raj Chetty, NBER and University East Asian Seminar on Economics of California, Berkeley, and Adam Tadashi Fukui, Kyoto Sangyo (EASE), sponsored jointly with Looney, Federal Reserve Board, University, and Yasushi Iwamoto, Hong Kong University of Science “Social Insurance in Developing University of Tokyo, “Policy and Technology (HKU), Philippine Economies” Options for Financing the Future Institute for Development Studies Discussants: Roger Gordon and Medical and Long-Term Care Costs (PIDS), China Center for Economic Mario Lamberte in Japan” Research (CCER), Singapore Discussants: Epictetus Patalinghug Management University (SMU), Chih-Chin Ho, National and Raj Chetty Korea Development Institute (KDI), University, and Brian Erard,B. Tokyo Center for Economic Erard & Associates, “Estate and Young-Sun Koh, KDI, “Reforming Research (TCER), Chung-Hua Gift Taxation in Taiwan: An the Fiscal Management System in Institution for Economic Research Analysis of the Current System and Korea” (CIER), and the Australian Some Proposals for Reform” Discussants: Gilberto Llanto and Productivity Commission, took place Discussants: Takatoshi Ito and Chong Hyun Nam in Manila, Philippines, on June 23-25. Felipe Medalla The organizers were NBER Takero Doi, Keio University; Research Associates Takatoshi Ito, of Wilson Au-Yeung, Jason Toshihiro Ihori, University of Tokyo University, and Andrew K. McDonald, and Amanda Sayegh, Tokyo, and Kiyoshi Mitsui, Rose, of University of California, Australian Treasury, “Australian Gakushuin University, Berkeley. The theme of the meeting Government Balance Sheet “Sustainability, Debt Management, was fiscal policy and management. Management” and Public Debt Policy in Japan” The following papers were discussed: Discussants: Youngsun Koh and Eli Discussants: Eli Remolona and Remolona Michael Alba Roger Gordon, NBER and University of California, San Diego, Yum-Keung Kwan, City University Laurence J. Kotlikoff, NBER and and Wei Li, University of Virginia, of Hong Kong, “The Direct Boston University, and Hans Fehr “Financial, Taxation, and Regulatory Substitution between Government and Sabine Jokisch, University of Structures in Developing Countries” and Private Consumption in East Wurtzberg, “Japan’s Coming Discussants: Francis Lui and Asia” Generational Storm” Ponciano Intal Discussants: Kiyoshi Mitsui and Discussants: Yasushi Iwamoto and Mario Lamberte Dante Canlas Roberto Mariano and Delano Villanueva, SMU, “ Sustainable Xin-Qiao Ping and Jie Bie, CCER, Young Jun Chun, University of External Debt Levels: Estimates for “Fiscal Decentralization and Local Incheon,“Population Aging, Fiscal Selected Asian Countries” Public Good Provision in China” Policies, and National Saving: Discussants: Francis Lui and Discussants: Takero Doi and Cielito Prediction for Korean Economy” Rosario Manasan Habito Discussants: Laurence J. Kotlikoff and Shigeki Kunieda Seok-Kyun Hur, KDI, “Measuring Gilberto Llanto, PIDS, “Dealing the Effectiveness of Fiscal Policy in with Contingent Liabilities: The Korea” Philippines” Discussants: Wei LI and Yum- Discussants: Jason McDonald and Keung Kwan Shigeki Kunieda

Observed economic policies in nontax sources of revenue. Gordon cies given such difficulties. The other, developing countries differ sharply and Li contrast the implications of building on Grossman-Helpman (1994), from those in developed countries and two models designed to explain such presumes that political-economy prob- from those forecast by existing models anomalous policies. One approach, on lems in developing countries are worse, of optimal policies. For example, which they published in 2005, focuses leading to worse policy choices. In developing countries rely little on on the greater difficulties faced in poor this paper, they compare the contrast- broad-based taxes, and make substan- countries in monitoring taxable activi- ing theoretical implications of the two tial use of tariffs and seignorage as ty, and explores the best available poli- models with the data, and find that the

NBER Reporter Summer 2005 27 political-economy approach does potential welfare gain from introduc- to reduce risk. Au-Yeung, McDonald, poorly in reconciling many aspects of ing social safety nets in developing and Sayegh provide a framework for the data with the theory. In contrast, economies. Using panel surveys of optimal balance sheet management. the forecasts from their model are con- households in Indonesia and the Their major conclusions are that: 1) sistent with the data currently available. United States, they find that food con- fiscal sustainability depends on both High ratios of external debt to sumption falls by approximately 10 the expected path of future taxation GDP in selected Asian countries have percent when individuals become and the risks around that path; 2) opti- contributed to the initiation, propaga- unemployed in both countries. This mal balance sheet management tion, and severity of the financial and finding is surprising given that the requires knowledge of how risks affect economic crises in recent years, reflect- United States has an extensive social the balance sheet (and therefore ing runaway fiscal deficits and exces- safety net while Indonesia has virtually volatility in tax rates); and 3) the gov- sive foreign borrowing by the private none. Prior studies have interpreted ernment’s financial investment strategy sector. Applying the formal frame- such results as evidence that social should reduce the risk to government work proposed by Villanueva (2003) to insurance is of limited value in devel- finances from macroeconomic shocks a selected group of Asian countries, oping economies. The authors show that permanently affect the budget. Mariano and Villanueva estimate the that this conclusion is incorrect if the Based on this framework, the authors external debt thresholds beyond which consumption path is smooth, because find that a Future Fund portfolio that further debt accumulation will have individuals are highly risk averse. included (among other potential negative effects on growth and will Exploratory tests suggest that investments) domestic nominal securi- become unsustainable. Their frame- Indonesian households are indeed ties and equities of selected countries work extends the standard neoclassical quite risk averse because of subsis- would reduce overall balance sheet risk. growth model that incorporates global tence constraints. Therefore, social Kwan investigates the extent of capital markets. “Sustainability” is safety nets may be valuable in low- direct substitution between govern- measured in terms of the steady-state income economies despite the ment and private consumption in nine ratio of the stock of external debt-to- smoothness of consumption. East Asian countries. His regression GDP, as functions of real world inter- The taxation of property left to uncovers a significantly positive elastic- est rates, risk spreads and their respon- one’s heirs is among the oldest forms ity of substitution between govern- siveness to external debt burdens and of taxation, dating back at least to the ment and private consumption, imply- market perceptions of country risk, time of Ancient Egypt. In Taiwan, as ing that government and private con- marginal propensities to save out of in various other countries, this practice sumption on average are substitutes in national disposable income and for- continues today, with citizens and for- East Asia. A country-by-country eign borrowing, rates of technical eign residents both subject to estate analysis, however, reveals diversity in change, and parameters of the produc- and gift taxation. However, there is the substitutability estimates. The four tion function. The major policy impli- concern that the current estate and gift North East countries — China, Hong cations are that in the long run, fiscal tax structure in Taiwan may be con- Kong, Japan, and Korea – tend to consolidation and the promotion of tributing to resource distortions and share similar and small values of the private saving are critical, and that capital flight as a result of efforts by substitution elasticity. For the five reliance on foreign saving in a global- wealthy citizens and residents to avoid ASEAN countries studied here, the ized financial world has limits, particu- taxation. In this paper, Ho and Erard relationship between private and gov- larly when the risk spreads are posi- summarize the current structure of ernment consumption vary substan- tively correlated with rising external estate and gift taxation in Taiwan, tially, both in the sign and the magni- debt levels. review operational statistics on tax tude of the elasticity of substitution. Using VAR models, Hur esti- payments, and explore the implications Private and government consumption mates a trajectory of GDP induced by of some alternative proposals for in Malaysia and Thailand are strong variations in fiscal expenditure and tax- short-run reform. They also briefly substitutes, but they are complements ation policy. By assigning different discuss the scope for longer-term in Indonesia and Singapore. The combinations of identifying restric- reform involving the integration of the Philippines is in between, with a near- tions on the disturbances and measur- estate and systems. zero elasticity of substitution. ing the corresponding fiscal multipli- Since almost eliminating net debt, Fiscal incentives are closely relat- ers, he compares the robustness of the the Australian government’s attention ed to extra-budgetary revenues. Using estimated values of fiscal multipliers has turned to the financing of broader their definition, Ping and Bei explore with respect to the restrictions. His balance sheet liabilities, such as public the effects of “fiscal incentives” under empirical analysis of Korean data sector superannuation. Australia will decentralization on the responsiveness (from 1979 to 2000) reveals that the be developing a significant financial of public good provision to real local size and the significance of the esti- asset portfolio in the “Future Fund”’ needs. China could have some special mated fiscal multipliers in Korea are to smooth the financing of expenses problems with fiscal decentralization: very small, or that they are decaying through time. This raises the signifi- first, with a huge amount of extra- very fast. cant policy question of how best to budgetary revenue, the size of local Chetty and Looney assess the manage the government balance sheet government could expand, resulting in

28 NBER Reporter Summer 2005 a heavier burden on local citizens and (MHLW). The MHLW forecast over- the crisis, and is set to increase further. peasants. Second, there may be states the future increase in social secu- Government expenditure as a percent- decreasing returns to scale in local rity costs somewhat, although even the age of GDP has stabilized since 2001 at extra-budgetary expenditures. Third, authors’ projections show considerable around 25 percent after rising rapidly in “urbanization” (measured as the ratio growth in costs. Then the authors con- the 1990s, but it may resume its growth of rural population to the total popu- duct some policy simulations that con- and result in worsening fiscal balances lation) is negatively correlated with sider how to finance future health and when these risk factors materialize. local extra-budgetary expenditures on long-term care costs. They project Doi, Ihori, and Mitsui analyze urban maintenance. In China, the these costs from FY2004 to FY2100. sustainability issues with Japan’s fiscal process of industrialization and urban They show that the balanced-budget policy and then discuss the debt man- construction are not consistent. operation of health and long-term care agement policy based on the maturity Llanto draws the attention of insurance will create a large inequity of structure of government bonds. They policymakers and legislators to the fis- burdens among generations. Raising also investigate the coordination cal risk brought by the contingent lia- the premium immediately and pre- between fiscal and monetary authori- bilities arising from explicit or implicit funding for future rising costs will help ties in the direction of fiscal recon- government guarantees. He discusses to equate the burdens of generations. struction. Partial default by the fiscal the current attempt of the Philippine The premium would have to be raised authority is not effective in the long government to address this outstand- by around 90 percent immediately. run, because the gross rate of return ing issue. His paper uses the experi- Koh presents an overview of the on debt adjusts to offset changes in the ence of providing guarantees to infra- development of public finance in size of default. Inflationary taxes structure projects, which have given Korea since the 1970s; analyzes its cur- levied by the central bank will not have rise to large amounts of contingent lia- rent status; explains the institutional such an offsetting effect. Thus, the bilities, to illustrate the fiscal risk faced setup and assesses the recent reform emergency reform by the monetary by the government. Drawing from efforts; and proposes directions for authority may well be better than the existing literature (Lewis and Mody change to maintain financial health and emergency reform by the fiscal author- 1997, Mody and Patro 1996, Irwin and maximize the productivity of public ity as a way to avoid bankruptcy. Given others 1997, Mody 2000), Llanto spending. The Korean government has the present maturity structure of shows how the Philippine government maintained strong fiscal discipline bonds, Japan’s debt management poli- may organize a management frame- since the early 1980s, keeping its budg- cy should be based on the smoothing work for contingent liabilities. He et more or less in balance and its debt rule: as the termination date hovers concludes by pointing out the need for at low levels. There were large deficits within a ten-year period, it is preferable the government to develop credible after the economic crisis of 1997, but not to raise taxes or the inflation rate, regulatory and competition policy a return to surplus in 2000 thanks to or to cut expenditures, but rather to frameworks to minimize the demand the buoyant economy and consolida- reserve fiscal resources, even if the for guarantees in the future. tion efforts. The surplus has continued reliance on government bonds rises To analyze the sustainability of a since then. However, various risk fac- temporarily. It is also necessary not to social security system, it is necessary to tors can adversely affect the govern- cut taxes or to increase expenditures make long-term forecasts of the econ- ment’s finances. The aging population much, but instead to reduce reliance omy and the population for 50 years or and the technological catch-up with on government bonds gradually when longer. However, it is not possible to advanced economies imply much slow- there are relatively few expiring bonds. predict such a long term because of er economic growth in the decades Moreover, it is necessary to restrain the uncertainty over future circumstances. ahead. While revenue growth slows, increasing trend of reliance on govern- Fukui and Iwamoto discuss this and the demand for public expenditure is ment bonds. consider possible improvements for increasing rapidly. The financial sec- Kotlikoff, Fehr, and Jokisch government forecasts and some ideas tor’s restructuring in the wake of eco- develop a dynamic, life-cycle, general for social security reforms. They nomic crisis has left irretrievable debts equilibrium model to study the inter- assume that all present factors except (9 percent of 2004 GDP) in the public dependent demographic, fiscal, and population structure will be sustained sector, and that burden is expected to economic transition path of China, in the future. They project the labor fall mostly on taxpayers. All public Japan, the United States, and the EU. force, economic growth, health care pension schemes have structural prob- Each of these countries/regions is expenditures, and long-term care costs lems because of the imbalance entering a period of rapid and signifi- mechanically, rather than constructing between contributions and benefits. cant aging that will require major fiscal a sophisticated model for forecasting. Some of them (those for civil servants adjustments. But the aging of these Their estimates of health care expendi- and military personnel) are already in societies may be a cloud with a silver tures and the long-term care service serious trouble. Economic cooperation lining, in this case, in the form of cap- costs are roughly consistent with those with North Korea will demand more ital deepening that will raise real wages. estimated by others, and are a bit less and more government support in the In a previous model that excluded than the projection made by the future. Spending on social welfare pro- China, these authors predicted that tax Ministry of Health, Labor and Welfare grams has increased substantially after hikes needed to pay benefits along the

NBER Reporter Summer 2005 29 developed world’s demographic transi- expenditures, and if Chinese technolo- over the next several decades, he uses a tion would lead to a major capital gy and education levels ultimately life-cycle model that incorporates the shortage, reducing real wages per unit catch up with those of the West and generational accounting approach of human capital over time by one Japan, then the model’s long run looks needed to assess the distribution of fifth. A reformulation of their original much brighter. China eventually fiscal burden across generations. He model to incorporate government becomes the world’s saver and, there- finds that rapid population aging and investment suggests that this concern by, the developed world’s savior with long-term budgetary imbalance will was overstated. With the inclusion of respect to its long-run supply of capi- substantially lower the national savings government investment, they find tal and long-run general equilibrium rate in Korea. In addition, estimation much less crowding out over the prospects. Indeed, rather than seeing results for consumption functions with course of the century and only a 4 per- the real wage per unit of human capi- respect to various kinds of wealth sug- cent long-run decline in real wages. tal fall, the West and Japan see it rise by gest that the annuitization of wealth Adding China to the model further one fifth by 2030 and by three fifths by because of the maturing of public pen- alters, indeed dramatically alters, the 2100. Even if the Chinese time prefer- sions and the introduction of reverse model’s predictions. Even though ence rate gradually approaches that of annuity mortgages is likely to further China is aging rapidly, its saving behav- the Americans, real wages in 2030 are decrease the savings rate in the future. ior, growth rate, and fiscal policies are roughly 17 percent higher in 2030 and 4 The University of Chicago Press currently very different from those of percent higher at the end of the century. will publish these papers in an NBER developed countries. If successive Chun evaluates the effects of Conference Volume. Many of them cohorts of Chinese continue to save population aging and fiscal policies on are also available at “Books in like current cohorts, if the Chinese national saving in Korea. To predict Progress” on the NBER’s website. government can restrain growth in the national savings rate for Korea

*

30 NBER Reporter Summer 2005 Bureau News

Bernanke to Chair CEA

Former NBER Research Bernanke currently serves on the Six of the recent Chairs of the Associate Ben S. Bernanke, who had Federal Reserve System’s Board of President’s Council of Economic been a member of the National Governors. In addition, he also serves Advisers were also NBER Research Bureau’s Program of Research on as Professor of Economics and Public Associates prior to their nominations: Monetary Economics, has been con- Affairs at Princeton University, a posi- Martin Feldstein, appointed by Ronald firmed by the Senate to become tion he has held for twenty years. Reagan; Michael Boskin, under Chairman of the Council of Bernanke previously taught at Stanford President George H.W. Bush; Joseph Economic Advisers. Bernanke replaces University, New York University, and Stiglitz, President William J. Clinton; NBER Research Associate N. Gregory MIT. He earned his bachelor’s degree and R. Glenn Hubbard, N. Gregory Mankiw, who will be returning to from Harvard University and his Ph.D. Mankiw, and Harvey Rosen, President Harvard University. from MIT. George W. Bush.

Acemoglu Receives John Bates Clark Medal

NBER Research Associate Daron NBER’s Programs on Labor Studies Previous Clark medalists among Acemoglu, a professor of economics and on Economic Fluctuations and NBER Research Associates were: Zvi at MIT, received the American Growth. He received the medal for his Griliches, Daniel L. McFadden, Martin Economics Association’s John Bates work on how a nation’s political and S. Feldstein, Joseph E. Stiglitz, James J. Clark Medal this year. The award is social institutions play a key role in Heckman, Jerry A. Hausman, Sanford given every two years to the economist guiding its economic destiny. J. Grossman, Paul R. Krugman, under the age of 40 who has made the Acemoglu received his B.A. from Lawrence H. Summers, David Card, most substantial contribution to the the University of York University in Kevin M. Murphy, , and field of economics. 1989 and his Ph.D. from the London Steven Levitt. Acemoglu is a member of the School of Economics in 1992.

NBER Researchers Lead American Economics Association

Three NBER Research Associates ceed Daniel L. McFadden. Sargent is a of California, Berkeley, and a member are the President, nominee for Professor of Economics at New York of the NBER’s Program on Aging. President-Elect, and immediate Past University and a member of the NBER President Martin President, respectively, of the American NBER’s Program on Economic Feldstein, a professor of economics at Economics Association (AEA). Fluctuations and Growth. Harvard University, was President for The AEA’s Executive Committee McFadden became AEA 2004. has chosen Thomas J. Sargent as the President in January 2005. He is a pro- nominee for President-Elect, to suc- fessor of economics at the University

NBER Reporter Summer 2005 31 Behavioral Responses to Taxation

The NBER’s Program on the Economic Effects of Reform” of Michigan, and Matthew D. Public Economics met in Cambridge Shapiro, NBER and University of on April 7 to discuss “Behavioral Susan Yang, Joint Committee on Michigan, “Temporary Investment Responses to Taxation.” Program Taxation, “Dynamic Scoring with Incentives: Theory with Evidence Director James M. Poterba of MIT Government Debt in a Neoclassical from Bonus Depreciation” organized the meeting. These topics Growth Model” Discussant: Austan Goolsbee, were discussed: NBER and University of Chicago Wojciech Kopczuk, NBER and Seth H. Giertz, Congressional Columbia University, “Tax Bases, Jesse Rothstein, NBER and Budget Office, “Recent Literature Tax Rates, and the Elasticity of Princeton University, “Estimation of on Elasticities” Reported Taxable Income” from Variation Discussant: Nada Eissa, NBER and Across the Wage Distribution: The Craig Johnson, U.S. Treasury, “The Georgetown University Earned Income ” Use of the Portfolio Allocation Discussant: Jeffrey Liebman, NBER General Equilibrium Model to Evaluate Christopher L. House, University and Harvard University

Giertz reviews the literature on real and financial decisions and real tax reductions that bring forth adjust- taxable-income elasticity, primarily and financial variables are determined ments in expected future government focusing on empirical studies examin- simultaneously. Businesses and the consumption, capital tax rates, or labor ing the U.S. tax changes of 1981, 1986, government sector issue securities to tax rates. Through the resulting 1990, and 1993 and the bracket creep meet their demand for capital. intertemporal distortions, current tax of the late 1970s and early 1980s. He Meanwhile, households and pension cuts can lower growth. Static scoring provides background on the impor- funds acquire securities in a manner of impacts, which assumes tance of the elasticity of taxable consistent with their risk-return prefer- no feedback from taxes to national income, both for forecasting income ences. The resulting portfolio choices income, does not necessarily overstate tax revenue and for assessing the effi- affect market rates of return, which in the revenue loss when compared with ciency implications of tax rate changes. turn affect the cost of capital to sec- dynamic scoring. He then discusses the major method- tors issuing securities. Total factor sup- Tax reforms usually change both ological issues, emphasizing the differ- plies are fixed, and the model is closed tax rates and tax bases. Using a panel ence in methodologies used and the to international capital flows. The of income tax returns spanning the sensitivity of estimates to an array of model predicts that the combined stat- two major U.S. tax reforms of the factors, including sample selection, the ic individual tax revenue estimate 1980s and a number of smaller under examination, and should be reduced by approximately 10 changes, Kopczuk finds that the elas- econometric techniques (or model percent when accounting for the ticity of income reported on personal specification). Although Giertz recog- behavioral responses included. The income tax returns depends on the nizes advances in the literature in behavioral offsets to the static revenue available deductions. This highlights recent years, he concludes that estimate are considerably larger when the fact that this key behavioral elastic- responses to tax rate changes are far the dividend and capital gains cuts are ity is not an immutable parameter but from fully understood and that there is examined in isolation from the ordi- rather can be controlled to some much to be gained from continued nary rate cuts. For dividend taxes, the extent by policymakers. One implica- research on this topic. revenue offset ranged from 22 to 36 tion is that base broadening reduces Johnson uses a general equilibri- percent; for capital gains taxes, the off- the marginal efficiency cost of taxa- um model with a detailed household set ranged from 44 to 67 percent. tion. The results are very similar for all sector that allows for portfolio choice Neoclassical growth models pre- income categories, indicating that the to examine the economic effects of dict positive growth effects throughout rich are more responsive to tax rates the recent reductions in individual tax the entire transition path after a reduc- because tax rules that apply to them rates on ordinary income, capital gains, tion in capital or labor tax rates when are different (their tax base is narrow- and dividends, and the implications of lump-sum taxes or transfers are used er). The point estimates indicate that the resulting portfolio shift for revenue to balance the government budget. In the Tax Reform Act of 1986 reduced estimating. His model focuses on tax contrast, debt-financed tax cuts can the marginal cost of collecting a dollar distortions in household portfolio generate a complex set of dynamic of tax revenue, with roughly half of decisions and the ensuing effects on interactions between current and this reduction attributable to the base the allocation of physical capital across expected future policy. Yang considers broadening and the other half to the production sectors. Taxes distort both the consequences of bond-financed tax rate reduction. The analysis in this

32 NBER Reporter Summer 2005 paper also offers a reconciliation of from quantity data alone. The bonus DiNardo, Fortin, and Lemieux (1996) disparate estimates obtained by previ- depreciation allowance passed in 2002 to permit fully nonparametric estima- ous studies of the tax responsiveness and increased in 2003 provides a sharp tion of labor supply and wage schedule of income. test of the theory. In the law, certain changes for female workers during this House and Shapiro consider the types of long-lived capital goods qual- period. He finds compelling evidence economy’s reaction to a temporary tax ify for substantial tax subsidies while that the EITC expansion caused sub- subsidy for investment. Because the others do not. The data show that stantial increases in the labor supply of eventual payoff from acquiring a long- investment in capital that qualified for low- and mid-skill single women with lived capital good is unrelated to the the subsidy was substantially higher children. Both the margin of increase date of purchase or installation, there than in capital that did not. The adjust- — large changes in participation and are powerful incentives to delay or ment cost parameters implied by the no changes in hours conditional on accelerate investment to take advan- data are in line with estimates from participation — and its distribution tage of predictable intertemporal vari- earlier studies. Market prices do not across tax brackets indicate that reduc- ations in cost. For such goods, the elas- react to the subsidy, which suggests tions in average tax rates were far more ticity of investment demand is nearly that internal adjustment costs not important than changes in marginal infinite. Consequently, for a temporary reflected in market prices are impor- rates. Estimates of changes in wages tax change, the price of long-lived cap- tant for investment decisions. are much less precise, but generally ital goods fully reflects the tax subsidy Rothstein proposes a new indicate that wages increased slightly regardless of the elasticity of invest- approach to estimating the incidence and insignificantly with labor supply, ment supply. This result is very gener- of federal income taxes, using varia- consistent with perfectly elastic labor al and relies only on an arbitrage argu- tion in exposure to tax changes across demand. Rothstein finds, however, that ment. Thus, contrary to conventional the wage distribution to provide iden- EITC-eligible women’s wages fell rela- wisdom, price data provide no infor- tifying variation in the impacts of tive to those of similarly-skilled but mation on the elasticity of supply. these changes on labor markets for ineligible women. This is not consistent Instead, because the price of invest- workers of different skill levels. with standard incidence models, but he ment goods shifts by exactly the Focusing on the mid-1990s expansion speculates about possible explanations. amount of the subsidy, the elasticity of of the Earned Income Tax Credit, investment supply can be inferred Rothstein extends the approach of

Public Economics

The NBER’s Program on and Interest Rates” University; Richard Romano, Public Economics met in Cambridge University of Florida; and Sinan on April 8. Members discussed these John J. Wallis, NBER and Sarpea, Carnegie-Mellon University, papers: University of Maryland, and Barry “Profiling in Bargaining Over R. Weingast, Stanford University, College Tuitions” Eric M. Engen, American “Equilibrium Federal Impotence: Discussant: Richard Zeckhauser, Enterprise Institute, and R. Glenn Why the States and Not the NBER and Harvard University Hubbard, NBER and Columbia American National Government University, “Federal Government Financed Economic Development Don Fullerton, NBER and Debts and Interest Rates (NBER in the Antebellum Era” University of Texas, and Garth Working Paper No. 10681) Discussant: Brian Knight, NBER Heutel, University of Texas, “The and Brown University General Equilibrium Incidence of William G. Gale and Peter R. Environmental Mandates” Orszag, Brookings Institution, Dennis Epple and Holger Sieg, Discussant: Hilary Sigman, NBER “Budget Deficits, National Savings, NBER and Carnegie Mellon and Rutgers University

Using a standard set of data and a lent to one percent of GDP would specifications to permit comparisons simple analytical framework, Engen increase the real interest rate by about among different approaches taken in and Hubbard reconsider and add to two or three basis points. They present previous research: the effect of an empirical evidence on the effect of their own empirical analysis in two expected, or projected, measure of federal government debt and interest parts. First, they examine a variety of federal government debt on a forward- rates. They begin by analytically deriv- conventional reduced-form specifica- looking measure of the real interest ing the effect of government debt on tions linking interest rates and govern- rate; of an expected, or projected, the real interest rate and find that an ment debt and other variables. In par- measure of federal government debt increase in government debt equiva- ticular, they estimate three types of on a current measure of the real inter-

NBER Reporter Summer 2005 33 est rate; and of a current measure of interest rates by 30 to 60 basis points. and reveals it to the college, then the federal government debt on a current Why did the states and not the college receives additional information measure of the real interest rate. Most federal government promote econom- incorporated in the outside offer. This of the statistically significant estimated ic development during the antebellum information may trigger a counter- effects are consistent with the predic- era? Wallis and Weingast argue that offer that improves financial aid for tion of the simple analytical calcula- the impotence of the federal govern- the applicant. The empirical analysis in tion. Next, the authors provide evi- ment stemmed from political con- this paper suggests that counter-offers dence from vector autoregression straints on Congress. They develop a typically are not correlated with any analysis. In general, these results are theory of legislative choice that illumi- information that is available at the ini- similar to those found in the reduced- nates the problems Congress faces. tial admission stage. However, count- form econometric analysis and are First, majority rule legislatures have er-offers respond strongly to the infor- consistent with the analytical calcula- difficulty voting to fund large infra- mation content of the outside offer tions. Taken together, the bulk of these structure projects with geographically received by a potential student. The empirical results suggest that an concentrated benefits if the expenses empirical evidence is consistent with increase in federal government debt are spread over all through general tax- the notion that signaling and profiling equivalent to one percent of GDP, all ation. Second, benefit, or Lindahl, tax- are important aspects of the college else equal, would be expected to ation affords a potential way out of admission process. increase the long-term real rate of inter- this problem. Third, majority rule leg- Regulations that restrict pollution est by about three basis points, although islatures easily finance large packages by firms also affect decisions about use one specification suggests a larger of small projects, such as rivers and of labor and capital. Thus they affect impact, while some estimates are not harbors or lighthouses. The final piece relative factor prices, total production, statistically significantly different from of the puzzle is institutional: although and output prices. For non-revenue- zero. By presenting a range of results states could use benefit taxation, the raising environmental mandates, what with the same data, the authors illustrate Constitution prevents the national are the general equilibrium effects on the dependence of estimation on spec- government from using it. Hence the wage, the return to capital, and rel- ification and definition differences. states but not the federal government ative output prices? Perhaps surpris- Gale and Orszag provide new could undertake large infrastructure ingly, Fullerton and Heutel cannot estimates of the effects of fiscal policy investments. The authors show how find any existing literature that even on national saving and interest rates. actual spending from 1790 to 1860 fits asks that question, in any model. Their They find that between 50 and 80 per- the model’s predictions. paper starts with the standard two-sec- cent of increases in unified deficits are Prospective college students are tor tax incidence model and modifies consumed, which implies that deficits uncertain about whether they will be one sector to include pollution as a reduce national saving. They also find admitted to any given institution and factor of production that can be a that a sustained increase in the project- thus typically apply to several. Part of complement or substitute for labor or ed unified deficit equal to 1 percent of this uncertainty is attributable to the for capital. Then they look at four GDP raises interest rates by 25 to 35 fact that the applicant does not know types of mandates, and determine con- basis points, and a sustained increase how the college will “read” his or her ditions for each that place more of the of that magnitude in the projected pri- record. Different colleges will receive burden on labor or on capital. Stricter mary deficit (the unified deficit exclud- different signals about the underlying regulation does not always place less ing interest payments) raises interest skills of a student. Epple, Romano, burden on the factor that is a better rates by 44 to 67 basis points. Indeed, Sarpea, and Sieg model the admission substitute for pollution. Also, a restric- despite a rancorous public debate, process of a single college as a bar- tion on the absolute amount of pollu- there appears to be a surprising degree gaining game between the college and tion creates scarcity rents, and thus of convergence in recent estimates of a potential student with sequential raises the dirty sector’s output price by the effects of fiscal policy on interest moves and asymmetric information. more than a relative restriction on pol- rates: results from a variety of econo- Their model predicts that the college lution per unit of output. In some per- metric studies imply that an increase in makes an initial offer based on the verse cases that the authors identify, the unified deficit of 1 percent of GDP, information in the application package. some of those policies might reduce the sustained over ten years, would raise If a student obtains an outside offer dirty sector’s output prices.

34 NBER Reporter Summer 2005 Environmental Economics

The NBER’s Working Group Taylor, NBER and University of Michael B. Ward, University of on Environmental Economics met in Wisconsin, “The Green Solow California, Santa Barbara; and Cambridge on April 8 and 9. NBER Model” (NBER Working Paper No. Timothy K.M. Beatty, University Research Associate Don Fullerton, 10557) of British Columbia, “Are Mercury University of Texas, Austin, organ- Discussant: William Nordhaus, Advisories Effective? Information, ized this program: NBER and Yale University Education, and Fish Consumption” Discussant: Anna Alberini, David Popp, NBER and Syracuse Akie Takeuchi, Maureen University of Maryland University, “Exploring Links Cropper, and Antonio Bento, Between Innovation and Diffusion: University of Maryland, “The Denise L. Mauzerall, Babar Adoption of NOx Control Welfare Effects of Urban Land Use Sultan, Namsoug Kim, and David Technologies at U.S. Coal-Fired Policies on Slum Dwellers: The Case F. Bradf ord, Princeton University, Power Plants” of Mumbai” “Charging NOx Emitters for Health Discussant: Wayne Gray, NBER and Discussant: Patrick Bayer, NBER Damages: An Exploratory Analysis” Clark University and Yale University Discussant: William Pizer, Resources for the Future William A. Brock and M. Scott Jay P. Shimshack, Tufts University;

While many studies have looked mental economics, The Environmental Takeuchi, Cropper, and Bento quan- separately at innovation and adoption Kuznets Curve (EKC), and the core tify these tradeoffs by estimating a of technologies, the two processes are model of modern macroeconomics, model of residential location and com- linked. Advances (and expected the Solow Model, are intimately relat- mute mode choice using data for 5,000 advances) in a single technology ed. Once the authors amend the Solow households in Mumbai, India. The should affect both its adoption rate model to incorporate technological authors compute compensating varia- and the adoption of alternative tech- progress in abatement, the EKC is a tion for in situ improvements in hous- nologies. Moreover, advances made necessary by-product of convergence ing (better roofs and piped water) and abroad may affect adoption differently to a sustainable growth path. The for programs that relocate slum from improvements developed domes- amended model — which they dub the dwellers to better housing. The short- tically. Popp combines plant-level data “Green Solow” — generates an EKC run benefits of relocation programs on U.S. coal-fired electric power plants relationship between the flow of pol- (that is, benefits holding workplace with patent data pertaining to NOx lution emissions and income per capi- location fixed) depend crucially on pollution control techniques to study ta, and the stock of environmental how much they distance workers from these links. He shows that technologi- quality and income per capita. The their jobs. The utility of relocation cal advances, particularly those made resulting EKC may be hump shaped or programs also depends on neighbor- abroad, are important for the adoption strictly declining. The authors explain hood composition, measured in terms of newer post-combustion treatment why current methods for estimating an of religion and mother tongue. technologies, but have little effect on EKC are likely to fail whenever they do Shimshack, Ward, and Beatty the adoption of older combustion not account for cross-country hetero- examine responses to a national FDA modification techniques. Moreover, geneity in either initial conditions or advisory that urged at-risk individuals Popp provides evidence that adaptive deep parameters. Then they develop to limit store-bought fish consumption R and D by U.S. firms is necessary an alternative empirical method, close- because of the dangers of methyl-mer- before foreign innovations are adopted ly related to tests of income conver- cury. The authors find that both edu- in the United States Expectations of gence used in the macro literature, and cation and newspaper readership are future technological advances delay investigate preliminary tests of the important determinants of consump- adoption. Nonetheless, as in other model’s predictions using data from tion response, suggesting that infor- studies of environmental technologies, OECD countries. mation acquisition and assimilation are the effect of other explanatory vari- A key issue in slum upgrading is key factors for risk avoidance. While ables is dominated by the effect of whether residents are made better off the advisory was effective for some environmental regulations, demon- by improving housing in situ, or by groups, the authors do not find a strating that the mere presence of being relocated. The answer to this response among the relatively large environmental technologies is not question depends on the tradeoffs group of at-risk households that met enough to encourage its usage. people are willing to make between neither the education nor the reader- Brock and Taylor demonstrate commuting costs, housing costs, and ship criteria. that a key empirical finding in environ- housing and neighborhood attributes. Mauzerall, Sultan, Kim, and

NBER Reporter Summer 2005 35 Bradford present a proof-of-concept of surface O3, depending on tempera- research shows that a shift of a unit of analysis of the measurement of the ture and local biogenic hydrocarbon NOx emissions from one place or time health damage of ozone (O3) pro- emissions. They also examine the to another could result in large changes duced from nitrogen oxides (NOx = dependence of resulting ozone-related in resulting health effects because of NO + NO2) emitted by individual health damage on the size of the ozone formation and exposure. The large point sources in the eastern exposed population. The investigation authors indicate how the type of mod- United States. Using a regional atmos- is relevant to the increasingly widely eling carried out here might be used to pheric model of the eastern United used “cap and trade” approach to attach externality-correcting prices to States — the Comprehensive Air NOx regulation, which presumes that emissions. Charging emitters fees that Quality Model with extensions shifts of emissions over time and are commensurate with the damage (CAMx) — they quantify the variable space, holding the total fixed over the caused by their NOx emissions would impact that a fixed quantity of NOx course of the summer O3 season, will create an incentive for emitters to re- emitted from individual sources can have a minimal effect on the environ- duce emissions at times and in locations have on the downwind concentration mental outcome. By contrast, this where they cause the largest damage.

Labor Studies

The NBER’s Program on Labor of California, Berkeley, and Jesse Pittsburgh, “Do Women Shy Away Studies met in Cambridge on April Rothstein, NBER and Princeton from Competition?” 15. Program Director Richard University, “Racial Segregation and Freeman and Research Associate the Black-White Test Score Gap” Henry S. Farber, NBER and Lawrence Katz, both of Harvard Princeton University, “Reference University, organized the meeting. Edward P. Lazear and Ulrike Dependent Preferences and Labor These papers were discussed: Malmendier, NBER and Stanford Supply: The Case of New York City University, and Roberto A. Weber, Taxi Drivers” Derek Neal, NBER and University Carnegie Mellon University, of Chicago, “Why Has Black-White “Sorting in Experiments” David Blanchflower, NBER and Skill Convergence Stopped?” Dartmouth College, and Andrew (NBER Working Paper No. 11090) Muriel Niederle, NBER and Oswald, Warwick University, “The Stanford University, and Lise Wage Curve Reloaded” David Card, NBER and University Vesterlund, University of

All data sources indicate that quite significant throughout the twen- desegregation plans, or with the geo- black-white skill gaps diminished over ty-first century. graphic features of a city, confirm this most of the twentieth century. Card and Rothstein study the finding. However, models that include However, black-white skill gaps, as effects of school and neighborhood school segregation, neighborhood seg- measured by test scores among youth segregation on the relative SAT scores regation, and measures of the relative and educational attainment among of black students across different met- exposure of blacks to other character- young adults, have remained constant ropolitan areas, using large microdata istics of their neighbors show much or increased in absolute terms since samples for the 1998-2001 test weaker effects of neighborhood segre- the late 1980s. Neal examines the cohorts. Without controlling for gation. This suggests that race per se is potential importance of discrimination neighborhood segregation, they find not the primary source of these against skilled black workers, changes that school segregation is negatively effects; rather, it is exposure to more in black family structures, changes in associated with black relative test economically successful neighbors. black household incomes, black-white scores, and with relative employment Experiments provide a controlled differences in parenting norms, and and education outcomes measured in environment in which factors can be education policy as factors that may the 2000 Census. In models that isolated and studied more easily than in contribute to the recent stability of include both school and neighborhood the real world, but often they are chal- black-white skill gaps. Without segregation, though, the effect of rela- lenged on their applicability to the real changes in public policy or the econo- tive exposure to black schoolmates is world. One major feature of experi- my that facilitate investment in black uniformly small and statistically ments is that they select subjects ran- children, even best case scenarios sug- insignificant, while neighborhood seg- domly, while markets do not draw indi- gest that approximate black-white skill regation has a strong negative effect. viduals randomly. Markets allow peo- parity is not possible before 2050. Instrumental variables estimates that ple to sort to certain activities and away Equally plausible scenarios imply that are identified by the components of from others based on their prefer- the black-white skill gap will remain school segregation associated with ences, beliefs, and skills. Lazear,

36 NBER Reporter Summer 2005 Malmendier, and Weber design an laboratory experiment, the authors develops an empirical model of daily experiment to demonstrate the impor- examine an environment in which labor supply that incorporates refer- tance of sorting in the context of women and men perform equally well ence-dependent preferences, and social preferences. When individuals under both a noncompetitive piece applies it to data on the daily labor sup- are constrained to play a dictator game, rate and a competitive tournament ply of New York City taxi drivers. His nearly two thirds of the subjects share. scheme. Participants then are asked to estimates suggest that there may be a But when the same subjects are choose the incentive scheme for their reference level of income on a given allowed to avoid the situation altogeth- next performance. The authors find day that affects labor supply. However, er, less than one third tend to share. that twice as many men as women the reference level varies substantially This dramatic reversal of proportions choose the tournament over the piece from day to day for a given driver. This demonstrates the importance of rate. This gender gap in tournament seriously limits the predictive power of accounting for sorting when applying entry cannot be explained by perform- the reference point model and under- experimental results to the real world. ance before or after the entry decision mines the usefulness of the construct The authors also show that institutions has been made. While men are more of the reference income level as a designed to entice pro-social behavior optimistic about their relative perform- determinant of labor supply. actually may induce adverse selection. ance than women, this difference can Blanchflower and Oswald pro- Increasing the total surplus available only explain a small share of the gen- vide evidence for the existence of a for sharing first induces those individ- der gap in tournament entry. Finally, wage curve — a microeconometric uals who are least willing to share to the authors show that gender differ- association between the level of pay sort back into the dictator game. Thus ences exist even when participants and the local unemployment rate — in the impact of social preferences simply decide how to be paid for a past modern U.S. data. Consistent with remains much lower when sorting is performance. They use this decision as recent evidence from more than 40 possible than in a mandatory dictator a control for non-tournament specific other countries, the wage curve in the game, even if sharing is subsidized by gender differences (such as risk aver- United States has a long-run elasticity higher payoffs. sion, feedback aversion, general over- of approximately –0.1. In line with the Competitive high-ranking posi- confidence), and find a large residual paper’s theoretical framework: wages tions still are occupied largely by men, gender effect when participants select are higher in states with more generous and women in academia remain scarce tournament compensation for a future unemployment benefits; the perceived in engineering and sciences. Suggested performance. probability of job-finding is lower in explanations for this fact focus mostly Recent theoretical work has states with higher unemployment; and on discrimination and differences in focused on the importance of refer- employees are less happy in states that abilities or preferences (in terms of ence-dependent preferences. Typically, have higher unemployment. The work hours or field of study). this work assumes a discontinuity in authors conclude that it is reasonable Niederle and Vesterlund explore an marginal utility at some base reference to view the wage curve as an empirical additional factor, namely that women level. Marginal utility below this kink in law of economics. and men may differ in their selection the utility function is higher than mar- into competitive environments. In a ginal utility above the kink. Farber

NBER Reporter Summer 2005 37 Education Program

The NBER’s Program on Multiple Proxies for Quality” Education met in Cambridge on Ofer Malamud, University of April 22. Program Director Caroline Nazmul Chaudhury, Jeffrey Chicago, and Christian Pop- M. Hoxby of Harvard University Hammer, and F. Halsey Rogers, Eleches, Columbia University, “The organized the meeting. These papers World Bank; Michael Kremer, Effect of Vocational Training on were discussed: NBER and Harvard University; and Labor Market Outcomes: Evidence Karthik Muralidharan, Harvard from an Educational Reform in Sarah J. Reber, University of University, “Missing in Action: Romania” California, Los Angeles, Teacher and Health Worker “Desegregation and Educational Absence in Developing Countries” Brian A. Jacob, NBER and Attainment for Blacks: Evidence Harvard University, and Lars from Louisiana” Jordan D. Matsudaira, University Lefgren, Brigham Young of Michigan, “Sinking or University, “What Do Parents Value Dan A. Black, Syracuse University, Swimming? Evaluating the Impact in Education? An Empirical and Jeffrey A. Smith, NBER and of English Immersion versus Investigation of Parents’ Revealed University of Maryland, “Estimating Bilingual Education on Student Preferences for Teachers” the Returns to College Quality with Achievement”

The desegregation of Southern desegregation was more important Bangladesh, Ecuador, India, schools following the Supreme Court’s than the increased exposure to whites. Indonesia, Peru, and Uganda and 1954 Brown decision was one of the A single standard deviation increase in recorded whether teachers and health more important innovations in U.S. initial black enrollment share was asso- workers were present. Averaging education policy in the twentieth cen- ciated with an additional improvement across the countries, about 19 percent tury. Reber assesses the effects of in high school graduation rates of of teachers and 35 percent of health desegregation of Louisiana schools on about 3.3 percentage points. workers were absent. In these coun- its intended beneficiaries, black stu- Existing studies of the effects of tries, the absent education and health dents. Substantial reductions in segre- college quality on earnings typically providers are rarely replaced by substi- gation between 1965 and 1970 were rely on a single proxy variable for col- tutes, unlike in developed countries. accompanied by large increases in per- lege quality. Black and Smith question When providers are absent, clients pupil funding in Louisiana, allowing the wisdom of this approach, given either are not served or are served by districts to “level up” school spending that a single proxy likely measures col- fewer providers. The authors explore a in integrated schools to that previously lege quality with substantial error. wide range of potential determinants experienced only in the white schools. They begin by considering the param- of absence at the individual, facility, Pre-existing black-white spending gaps eter of interest and its relation to the and state/country levels. Absence is were largest in districts with higher parameter estimated in the literature; not concentrated among a small num- shares of initial black enrollment, so this analysis reveals the potential for ber of frequently absent providers, but blacks in those districts experienced substantial bias. Then they consider rather seems to be fairly widespread. larger increases in funding than their three econometric approaches to the Absence rates are generally higher in counterparts in lower black enroll- problem that involve the use of multi- poorer regions; when school infra- ment-share districts. However, blacks ple proxies for college quality: combin- structure is poor; and when a school in high black enrollment-share districts ing the multiple proxies via factor hasn’t been inspected recently. Higher- experienced smaller increases in expo- analysis; using the additional proxies as ranking and more powerful providers, sure to whites (who were higher- instruments; and a GMM estimator such as headmasters and doctors, are income). A single standard deviation derived from a structural measurement absent more often than lower-ranking increase in initial black enrollment error model that generalizes the classi- ones. The relationship between share was associated with an addition- cal measurement error model. Their absence and contractual terms for al increase in per-pupil funding of estimates suggest that the existing liter- teachers seems more complicated than $290 (2003 dollars) but a 14 percentage ature understates the wage effects of often hypothesized. point smaller increase in exposure to college quality. In recent years, the role of bilin- whites. Blacks in high black enroll- Chaudhury, Hammer, Kremer, gual education as the dominant peda- ment-share districts also experienced Muralidharan, and Rogers report gogy for teaching immigrant students larger improvements in educational results from surveys in which enumer- has been challenged by federal and attainment, suggesting that the ators made surprise visits to primary state policy reforms. Critics of bilin- increase in funding associated with schools and health clinics in gual education contend that bilingual

38 NBER Reporter Summer 2005 programs hinder achievement by reduc- over the most effective method of quence of selection. ing incentives to learn English and by educating immigrant children. In addi- Using a unique dataset showing trapping students in classrooms with tion, the results inform a broader the number of parent requests for low-performing peers. Proponents debate regarding the importance of individual elementary school teachers argue that learning occurs most rapid- educational peer effects, because bilin- and information on teacher attributes, ly when students master concepts in gual programs involve learning in class- including principal reports of teacher their native language first, and there- es surrounded by a group of peers who characteristics that are typically unob- fore that bilingual education programs have markedly lower average achieve- servable (to the econometrician), promote achievement. To distinguish ment scores and are more similar in Jacob and Lefgren examine parents’ between these competing views, ethnic background than those in main- preferences about their children’s edu- Matsudaira exploits quasi-random stream English-immersion classes. cation. The authors present reduced- assignment of students to bilingual Vocational training and general form estimates of the relationship and mainstream (English-immersion) education are the two predominant between teacher characteristics and classes generated by discontinuous forms of secondary schooling around parent requests, and then develop and program eligibility rules in a large the world. Most studies that compare estimate a structural model of parent urban school district. In the District, the effect of vocational and general choice that provides estimates of pref- eligibility for bilingual programs is education on labor market outcomes erence and cost parameters. They find determined by a test of English profi- in the cross-section suffer from selec- that, on average, parents strongly pre- ciency: students scoring below a pre- tion bias because less able students are fer teachers that principals describe as set threshold level are eligible for bilin- more likely to enroll in vocational pro- good at promoting student satisfaction gual classes; students scoring above grams. To avoid this bias, Malamud and place relatively less value on a this threshold are not eligible. Using and Pop-Eleches exploit a 1973 edu- teacher’s ability to raise standardized information on achievement and pro- cational reform in Romania that shift- math or reading achievement. These gram participation from a large admin- ed a large proportion of students from aggregate effects, however, mask strik- istrative dataset, Matsudaira compares vocational training to general educa- ing differences across family demo- students scoring just below and just tion. Using data from the 1992 graphics. Low-income and minority above this threshold for several years Romanian Census, the authors analyze families strongly value student achieve- following the proficiency test. the effect of this policy in the context ment and are essentially indifferent to Compared to students scoring just of a transition economy that experi- the principal’s report of a teacher’s above the threshold, students scoring enced a decline in manufacturing and a ability to promote student satisfaction. just below it are nearly 90 percent reallocation of labor to new jobs. They The results are reversed for higher- more likely to participate in a bilingual find that cohorts affected by the policy income and non-minority families. program and are surrounded by peers were significantly less likely to work in Moreover, the authors find that low- with significantly lower average manual or craft-related occupations income and minority families are sub- achievement scores. Despite these but no less likely to be unemployed or stantially less likely to request a partic- striking differences in classroom envi- out of the labor force than their coun- ular teacher. There is also some evi- ronments, however, he finds negligible terparts who were not affected by the dence that parents have preferences differences in achievement in both policy. Therefore, the authors con- regarding certain teacher characteris- reading and math (measured by tests clude that the cross-sectional differ- tics and even certain aspects of class- given in English) across these two ences in labor market outcomes room organization. For example, par- groups of students. These results between graduates of vocational and ents attempt to avoid first-year teach- speak directly to current policy debates general schools are largely a conse- ers and mixed-grade classrooms.

NBER Reporter Summer 2005 39 Health Economics

The NBER’s Program on Espinosa, University of Maryland, “Medicare Reimbursement, Nurse Health Economics, directed by “Marriage Protection or Marriage Staffing, and Patient Outcomes” Michael Grossman of City Selection?” University of New York, met in Ted Joyce, NBER and Baruch Cambridge on April 22. They dis- Sandra L. Decker, NBER and College, “Changes in Abortions and cussed these papers: International Longevity Center; Births Following Texas Parental Irena Dushi, International Notification Statute: A Regression Bjorn Lindgren, Fredrik Longevity Center; and Partha Deb, Discontinuity Approach” Andersson, and Kristian Bolin, Hunter College, “Medicare at Age Lund University, “Breast Cancer 65: Does It Level the Playing Jeff DeSimone, NBER and Survival, Work, and Earnings in Field?” University of South Florida, and Sweden” Edward Schumacher, Trinity Robert Kaestner, NBER and University, “Compensating Wage William N. Evans, NBER and University of Illinois, and Jose Differentials and AID Risks” University of Maryland, and Javier Guardado, University of Illinois, (NBER Working Paper No. 10861)

In two recent studies, Bradley et nificantly regarding wage rates; and ables bias, in single-equation models, al. (2002, 2005) examined the conse- work fewer hours. the authors should find a stronger quences of breast cancer for women’s Researchers from a variety of marital protection effect for informa- labor market attachment. Using cross- fields have noted a sharp rise in mor- tive deaths than for uninformative section data from the first wave of the tality for widows soon after the death causes of death. In Cox proportional U.S. Health and Retirement Study, of their spouses. Because of assorta- hazard models, they find the death of a linked to longitudinal Social Security tive mating, married couples tend to spouse from an uninformative cause earnings data, they first found that share many of the same characteristics, has a smaller impact on mortality than breast cancer survivors were less likely so this result may reflect an omitted a death from an informative cause. to be employed, but if employed variable bias rather than a causal rela- They also find that for husbands the worked more hours and had higher tionship. In the paper by Espinosa death of a spouse from an uninforma- earnings and wage rates. Their 2005 and Evans is the notion that some tive cause generates a statistically sig- paper used a longitudinal dataset creat- causes of death reveal more informa- nificant increase in mortality. ed from 445 telephone interviews with tion about the surviving spouse than Decker, Dushi, and Deb esti- women diagnosed with breast cancer others. In the extreme, if a cause of mate the effect of gaining Medicare 6, 12, and 18 months after diagnosis. A death was assigned randomly, then coverage at age 65 on the use of health matched control group was created by these types of deaths could be used to services and on health outcomes. drawing a sample from the Current identify the marriage protection effect. Using six waves of data from the Population Survey. Here Bradley and In practice, one cannot specify what Health and Retirement Study (HRS), co-authors found that breast cancer causes of death are randomly assigned, they show that those who have been survivors were less likely to work 6 but instead can identify those that are uninsured use fewer health services months after diagnosis, as in their first not correlated with observed charac- than others before the age of 65, but study, but also that breast cancer sur- teristics. Specifically, Espinosa and more after the age of 65. Although the vivors worked fewer hours, contrary to Evans use data from the National availability of supplementary insurance their previous result. Andersson, Longitudinal Mortality Survey and the reduces the equalizing effect of Bolin, and Lindgren theoretically National Health Interview Survey Medicare on the use of health services, analyze the individual’s response to Multiple Cause of Death supplement a significant equalizing effect still severe illness and its dependence on to create longitudinal datasets of mar- remains. Turning 65 also discontinu- the institutional structure. They also ried couples aged 50 to 70. They ini- ously increases the probability of replicate the estimates performed in tially use this sample to identify those being diagnosed with certain medical the U.S. studies using data from causes of death that are predicted by conditions, such as diabetes. Sweden, where healthcare is financed socio-economic status (income and Kaestner and Guardado study by general taxes and available to all res- education) and those that are not. the effect of plausibly exogenous idents, irrespective of labor market sta- They refer to these two types of deaths changes in Medicare reimbursement tus. They find that breast cancer sur- as informative and uninformative — caused by geographical reclassifica- vivors, compared to the general female causes of death, respectively. If the tion — on hospital staffing (nurses) population: are less likely to be heightened mortality of surviving and patient outcomes. For patient out- employed full-time; do not differ sig- spouses is subject to an omitted vari- comes, they focus on nurse-sensitive

40 NBER Reporter Summer 2005 measures — outcomes that are sensi- ment are not consistently associated law, but that change is sensitive to the tive to changes in the quantity of nurs- with an increase in adverse events, par- statistical specification. ing care — and in-hospital mortality. ticularly associated with nursing care. DeSimone and Schumacher The authors use a quasi-experimental Joyce analyzes changes in abor- examine the effect of HIV/AIDS research design based on a pre- and tions and births among teens in Texas infection risks on the earnings of reg- post-test with comparison group after enforcement of a parental notifi- istered nurses (RNs) and other health approach. The treatment group is hos- cation law in January 2000. The data care workers. The authors combine pitals that are geographically reclassi- are unique because they contain infor- data on metropolitan statistical area fied, which leads to changes in mation on a teen’s exact date of birth, (MSA) AIDS prevalence rates with Medicare reimbursement rates of which enables the author to overcome annual 1987-2001 Current Population about 10 percent on average. The com- an important source of the misclassifi- Survey (CPS) and quadrennial 1988- parison group is either hospitals in the cation bias. Researchers in previous 2000 National Sample Survey of area of reclassification (destination), or studies of parental involvement laws Registered Nurses (SRN) data. hospitals in the originating geographi- have measured a teen’s age at the time Holding constant the wages of control cal area. Using these groups, the of the birth or abortion. However, groups that likely are not exposed to authors compare changes in outcomes three quarters of all 17-year olds who AIDS risks and group-specific MSA surrounding the change in reimburse- conceive as minors, and who thus are fixed effects, a 10 percent increase in ment. Estimates of the association exposed to the law, give birth as 18- the AIDS rate raises RN earnings by between changes in Medicare reim- year olds. This latter group of teens is about 0.8 percent after 1992, when bursement and patient outcomes are often included among the comparison AIDS rates were falling but a more mixed. Changes in reimbursement are group, which results in biased esti- comprehensive categorization of associated with both better and worse mates of the effect of the law on fer- AIDS was used by the CDC. AIDS outcomes. Moreover, the association tility and even abortion. Using age at wage differentials are much larger for between changes in Medicare reim- conception to define exposure, Joyce RNs and non-nursing health practi- bursement and patient outcomes is not finds that abortions in Texas to 17-year tioners than for other nursing and consistent with the association olds fell conservatively 12 percent after health care workers. This suggests that between changes in reimbursement the law. The decline in abortions was the differential represents compensa- and nursing resources. Increases in limited to white Non-Hispanics and tion for job-related exposure to poten- Medicare reimbursement are associat- Hispanics. There was a small increase tially HIV-infected blood. ed with a decrease in nursing in births, between 2 and 4 percent, resources, but increases in reimburse- among the 17-year olds exposed to the

Higher Education

The NBER’s Working Group on Discussant: Ronald G. Ehrenberg, Hideo Akabayashi and Michio Higher Education met in Cambridge NBER and Cornell University Naoi, Keio University, “Why is on April 30. The group’s Director, there No ‘Harvard’ among Japanese Charles T. Clotfelter of Duke Susan Dynarski, NBER and Private Universities?” University, organized this program: Harvard University, “Finishing Discussant: Michael Rothschild, College: The Role of Schooling NBER and Princeton University Mark C. Long, University of Costs in Degree Attainment” Washington, “College Quality and Discussant: Christopher M. Eric P. Bettinger, NBER and Case Early Adult Outcomes” Cornwell, University of Georgia Western Reserve University, and Discussant: Peter Arcidiacono, Duke Bridget T. Long, NBER and University Todd R. Stinebrickner, University Harvard University, “Mass of Western Ontario, and Ralph Instruction or Higher Learning? Bradley R. Curs, University of Stinebrickner, Berea College, “The The Impact of Class Size in Higher Missouri, and Larry D. Singell, Jr. Causal Effect of Studying on Education” and Glen R. Waddell, University of Academic Performance” Discussant: Steven Rivkin, NBER Oregon, “Money for Nothing? The Discussant: Brian Jacob, NBER and and Amherst College Institutional Impact of Changes in Harvard University Federal Financial Aid Policy”

Using panel data from the college quality on a set of early adult with three alternative methods of esti- National Education Longitudinal outcomes. He compares the results mation. In general, he finds that col- Study, Long estimates the effect of using ordinary least squares (OLS) lege quality does have positive signifi-

NBER Reporter Summer 2005 41 cant effects on the outcomes studied, centage points and the share of national university tuition fees and using the OLS estimates. However, Hispanic and nonwhite women subsidies influence the equilibrium dis- there is some evidence of positive attempting or completing any years of tribution of tuition fees and the aca- selection bias in the results. The four college increasing by 6 and 8 percent- demic quality of the private universi- specifications all find evidence of posi- age points, respectively. While the esti- ties. Using cross-section data on tive effects of several college qualities mation strategy cannot separately iden- Japanese universities, the authors find on educational attainment, but mixed tify the effect of aid on entry and per- that private university tuition fees are evidence for the effects on earnings. sistence, Dynarski establishes fairly lower in prefectures where the quality This study points to the importance of tight bounds on the persistence effect, of the national university is high, as examining the educational production concluding that the aid programs theoretically predicted. function as a complex process with increase the college persistence rate by Bettinger and Long attempt to many inputs, each of which can have its between 6 and 11 percent. measure the effects of collegiate class own effects on outcomes of interest. Using an Instrumental Variable size on student outcomes such as per- Using new institution-level data, (IV) approach, Stinebrickner and sistence, subsequent course selection, Curs, Singell, and Waddell assess the Stinebrickner examine the causal and major choice. While class size is a impact of changing levels of federal effect of studying on grade perform- perennial issue in the literature on pri- aid on institution-level Pell revenues. ance. They take advantage of new data mary and secondary schooling, it has They also ask whether changes in the from Berea College in the Berea Panel been largely ignored in research on federal Pell Grant program correlate Study that includes information about higher education. Nonetheless, the differently with the college access of student time-use. They find that study- impact of class size in college is poten- needy students at institutions of differ- ing has a very important causal effect tially very important because it affects ent levels of selectivity. While it is not on student grade performance and that everything from competitive rankings surprising that institutional Pell rev- the IV estimate is much larger than the (20 percent of the U.S. News and World enues are sensitive to the generosity of Ordinary Least Squares estimate. The Report formula focuses on class size) to the Pell Grant program in general, the difference between these two estimates university budgets. Moreover, the authors document significant and arises primarily because individuals range of possible class sizes is substan- interesting asymmetries across institu- react to grade shocks in a particular tial in higher education (as many as 700 tional selectivity, both in terms of semester by changing their study effort per lecture in this data), suggesting magnitude and in terms of which in that semester. These findings sound large differences in students’ class- channel accounts for the measured a cautionary alarm about the use of room experiences and possibly large sensitivity, Pell-award values or Pell fixed effects estimators in cases where effects on students’ outcomes as well. enrollment. behavioral responses to information Few researchers have focused on colle- Half of college students drop out may be present. This paper is perhaps giate class size, though, because data before completing a degree. While the first to identify a specific underly- are rarely available. In addition, there there is strong evidence that financial ing avenue through which peer effects are inherent selection problems aid can increase college attendance, there are transmitted in educational con- because higher ability students appear is little evidence that it increases degree texts; it also provides evidence that to systematically avoid large classes. To completion or even years of complet- video games can have a large causal address this hole in the literature, the ed schooling. However, it is completed effect on educational outcomes. authors use a unique dataset that schooling that is rewarded by the labor The social and academic reputa- includes detailed class size, faculty, and market. Dynarski exploits the intro- tion of private universities in Japan student characteristics; they track two duction of two large state financial aid lags far behind that of the national uni- cohorts of nearly 41,000 students in programs to estimate the impact of aid versities. Akabayashi and Naoi argue four-year, public colleges in Ohio. on completed schooling. She finds that that this arises from the heavy subsidy Using an instrumental variables the aid programs increase the share of provided by, and the low tuition fees approach to control for selection bias, the population that completes a college set by, the central government for the they find that an increase in class size degree by 3 percentage points. The national universities. Using simulations leads to a higher likelihood of drop- effects are strongest among women, based on an assignment model of het- ping out and a reduction in subsequent with white, non-Hispanic women erogeneous students and universities, student interest in a subject. increasing degree receipt by 3.8 per- the authors show that the levels of

42 NBER Reporter Summer 2005 Political Economy

The NBER’s Working Group and Ethan Kaplan, University of NBER and Stanford University on Political Economy met in California, Berkeley, “The Fox News Cambridge on April 30. Group Effect: Media Bias and Voting” Alexander Dyck, University of Director Alberto Alesina of Harvard Discussant: Jesse Shapiro, Harvard Toronto; David Moss, Harvard University organized this program: University University; and Luigi Zingales, NBER and University of Chicago, Rafael Di Tella, Harvard Asim Ijaz Khwaja, Harvard “Media versus Special Interests” University; Sebastian Galiani, University, and Atif Mian, Discussant: Andrei Shleifer, NBER Universidad de San Andres, University of Chicago, “Do Lenders and Harvard University Argentina; and Ernesto Favor Politically Connected Firms? Schargrodsky, Universidad Rent Provision in an Emerging Gene M. Grossman, NBER and Torcuato Di Tella, Argentina, Financial Market” Princeton University, and Elhanan “Property Rights and Beliefs: Discussant: Efraim Ben Melech, Helpman, NBER and Harvard Evidence from the Allocation of Harvard University University, “A Protectionist Bias in Land Titles to Squatters” Majoritarian Politics” (NBER Discussant: Esther Duflo, NBER Guido Tabellini, Bocconi Working Paper No. 11014) and MIT University, “Culture and Institutions: Discussant: Allan Drazen, NBER Economic Development in the and University of Maryland Stefano Dellavigna, NBER and Regions of Europe” University of California, Berkeley, Discussant: Romain Wacziarg,

Possessing property rights may informative as to precisely how prop- Alternatively, voters may display a change the beliefs that people hold. Di erty rights change beliefs, although form of confirmatory bias. Tella, Galiani, and Schargrodsky there is suggestive evidence of a Corruption by the politically con- study this hypothesis using a natural behavioral channel. nected often is blamed for economic experiment from a squatter settlement Does the media affect voting? ills, particularly in less developed in the outskirts of Buenos Aires, DellaVigna and Kaplan address this economies. Using a loan-level dataset ensuring that the allocation of proper- question by looking at the entry of Fox of more than 90,000 firms represent- ty rights is exogenous to the character- News into cable markets. Between ing the universe of corporate lending istics of the squatters. Significant dif- October 1996 and November 2000, in Pakistan between 1996 and 2002, ferences exist in the beliefs that squat- the conservative Fox News Channel Khwaja and Mian study rents to polit- ters with and without property rights was introduced into the cable pro- ically connected firms in banking. declare that they hold. For example, gramming of 20 percent of U.S. towns. Classifying a firm as “political” if its property rights make squatters’ beliefs Using a dataset of voting data for director participates in an election, the closer to those that favor the workings 8,634 towns, the authors ask if authors examine the extent, nature, of a free market, including materialist Republicans gained vote share in and economic costs of provision of and individualist beliefs (such as the towns where Fox News entered the political rents. They find that political belief that money is important for cable market by the year 2000. They firms borrow twice as much and have happiness, or the belief that one can find no significant effect of the intro- 50 percent higher default rates. Such be successful without the support of a duction of Fox News on the vote preferential treatment occurs exclu- large group). These effects appear share in Presidential elections between sively in government banks — private large. The value of a (generated) index 1996 and 2000. In fact, they can rule banks provide no political favors. of pro-market beliefs for squatters out an effect of Fox News larger than Using firm fixed effects and exploiting without property rights is 78 percent 0.5 percentage points. The results are variation across time or lenders, the of that of the general Buenos Aires robust to town-level controls, state and authors show that the observed politi- population. The value for squatters county fixed effects, and alternative cal preference is driven by the political who receive property rights is 98 per- specifications. They also find no signif- status of the firm and not by any cent of that of the general population. icant effect of Fox News on voter unobserved firm characteristic. The In other words, giving property rights turnout. These results imply that Fox political rents thus identified increase to squatters causes a change in their News convinced between 0 and 1.5 with the strength of the firm’s politi- beliefs that makes them indistinguish- percent of its viewers to vote cian and whether he or his party is in able from the general population, in Republican. This evidence is consis- power; they fall with the degree of spite of the large differences in the tent with the view that voters are electoral participation in his con- lives they lead. This experiment is less sophisticated and filter out media bias. stituency. Direct evidence rules out

NBER Reporter Summer 2005 43 alternative explanations, such as social- development through culture. The interests, and why this constraint is ly motivated lending by government indicators of culture that Tabellini uses often not more effective, the authors banks to politicians. The economy- in this paper also are strongly correlat- introduce a simple model of profit- wide costs of the rents identified are ed with economic development and maximizing media. The model sug- estimated to be 0.3 percent to 1.9 per- with available measures of institutions gests that the media are particularly cent of GDP every year. in a cross-country setting. effective in this role when the audience Does culture have a causal effect Dyck, Moss, and Zingales argue is large, when an issue can be more on economic development? The data that profit-seeking media firms can easily converted into entertaining on European regions suggest that it play an important role in reducing the news, and when subscriptions are a does. Culture is measured by indicators influence of powerful economic inter- more important source of revenues of individual values and beliefs, such ests on policymaking. Motivated to than advertising. as trust and respect for others, and reach big audiences by the lure of large Grossman and Helpman devel- confidence in individual self-determi- profits, media firms typically seek to op a novel model of campaigns, elec- nation. To isolate the exogenous varia- transform real events and issues — tions, and policymaking in which the tion in culture, Tabellini relies on two including public policy issues — into ex ante objectives of national party historical variables used as instru- entertaining stories. In so doing, they leaders differ from the ex post objec- ments: the literacy rate at the end of end up informing the public about tives of elected legislators. This gener- the nineteenth century and the political these issues and events, thus overcom- ates a distinction between “policy rhet- institutions in place over the past sev- ing the standard problem of rational oric” and “policy reality” and intro- eral centuries. The political and social ignorance (Downs 1957) that lies at the duces an important role for “party dis- history of Europe provides a rich heart of the economic theory of regu- cipline” in the policymaking process. source of variation in these two vari- lation and Stigler’s hypothesis of regu- The authors identify a protectionist ables at a regional level. The exoge- latory capture. Consistent with this bias in majoritarian politics. When nous component of culture attributa- idea, the authors document that the trade policy is chosen by the majority ble to history is strongly correlated rise of investigative journalism delegation and legislators in the minor- with current regional economic devel- through “muckraking” magazines ity have limited means to influence opment, after controlling for contem- helps to explain the emergence of choices, the parties announce trade poraneous education and for national important progressive-era legislation in policies that favor specific factors, and effects. Moreover, the data do not the early part of the twentieth century. the expected or export subsidy is reject the over-identifying assumption To clarify the circumstances under positive. Positions and expected out- that the two historical variables used as which media can serve as a constraint comes monotonically approach free instruments only influence regional on the political influence of vested trade as party discipline strengthens.

44 NBER Reporter Summer 2005 Health Care

The NBER’s Program on Corporation, and Richard Wang, Differentials and AIDS Risk” Health Care met in Cambridge on Astrazenica, “Advertising and (NBER Working Paper No. 10861) May 6. Research Associate Michael Intellectual Property in the U.S. Chernew of the University of Pharmaceutical Industry” Amalia R. Miller, University of Michigan organized this program: Virginia, “The Impact of John R. Moran, Syracuse Midwifery-Promoting Public Martin Gaynor and William B. University, and Kosali Ilayperuma Policies on Medical Interventions Vogt, NBER and Carnegie Mellon Simon, NBER and Cornell and Health Outcomes” University, and Jian Li, Carnegie University, “Income and the Use of Mellon University, “Is Drug Prescription Drugs by the Elderly: Janet Currie, NBER and University Coverage a Free Lunch? Cross-Price Evidence from the Notch Cohorts” of California, Los Angeles, and Elasticities and the Design of (NBER Working Paper No. 11068) Mark Stabile, NBER and Prescription Drug Benefits” University of Toronto, “Child Jeffrey DeSimone, NBER and Mental Health and Human Capital Tomas Philipson, NBER and University of South Florida, and Accumulation: The Case of University of Chicago; Darius Edward J. Schumacher, Trinity ADHD” Lakdawalla, NBER and RAND University, “Compensating Wage

Recently, many U.S. employers $1 per enrollee per year over two years advertising has on the efficient design have adopted more stringent prescrip- (ignoring discounting). This implies of intellectual property. They estimate tion drug benefits. In addition, the U.S. that more generous prescription drug the demand and supply schedules government will offer prescription benefits may be costly in the short run, underlying these efficiency effects drug insurance to approximately 42 but can generate cost savings in the using patent-expirations in the U.S. million Medicare beneficiaries in 2006. long run. These long-run effects pharmaceutical markets from 1990- Gaynor, Li, and Vogt use data from should be taken into account when 2003 as an instrument. Such expira- Medstat on individual claims and ben- estimating the expected cost of pro- tions display the interesting pattern efit design from 1997-2001 to examine grams like the Medicare prescription that — for a large share of drugs — the dynamic structure of demand for drug benefit. there are output reductions post-expi- health care, focusing specifically on In addition to spending a vast ration because advertising is reduced whether prescription drugs are amount of resources on Research and more than prices are. The authors esti- (dynamic) complements or substitutes Development (R and D), the pharma- mate the quantitative degree to which for inpatient and outpatient care. The ceutical industry is engaged in a sub- there is excessive advertising by U.S. authors study the effect of changing stantial amount of marketing of their drug manufacturers under current consumers’ copayments for prescrip- products, through both detailing to intellectual property law and the tion drugs on the quantity demanded doctors and direct-to-consumer adver- degree to which current intellectual and expenditure on prescription drugs, tising to patients. Given that the share property rights are inefficiently and on inpatient and outpatient care, of sales devoted to marketing is larger designed because of marketing. allowing for effects in the year of the than the share devoted to R and D by Moran and Simon use exoge- copayment change and the following pharmaceutical firms, there has been a nous variation in Social Security pay- year. The results show that a one dollar substantial debate about whether there ments created by the Social Security increase in the consumer copayment is excessive marketing and whether it benefits notch to estimate how for drugs has the following effects: in could be better allocated to productive retirees’ use of prescription medica- the first year after the change, drug R and D. However, previous economic tions responds to changes in their spending, outpatient spending, inpa- analysis of advertising, R and D, and incomes. In contrast to estimates tient spending, and total spending fall intellectual property ignores the obtained using ordinary least squares, by $9.71, $6.46, $3.39, and $20.39, impact one has on the other, making instrumental variables estimates based respectively. In the second year after welfare evaluations of policies that on the notch suggest that lower- the change, drug spending falls by affect both R and D and marketing income retirees exhibit considerable $8.35, while inpatient spending, outpa- infeasible. Lakdawalla and Wang income sensitivity in their use of pre- tient spending, and overall spending seek to remedy this problem by first scription drugs. These estimates are rise by $10.71, $13.03, and $22.85, analyzing the impact that intellectual potentially useful for thinking about respectively. Raising consumer drug property design has on the efficient the health implications of changes in copayment by $1 per prescription actu- degree of advertising by pharmaceuti- transfer payments to the elderly and ally increases overall spending by about cal firms and, vice versa, the impact for evaluating the benefits of the

NBER Reporter Summer 2005 45 recently enacted Medicare prescription tion paid for job-related exposure to work offers a number of innovations: drug benefit. potentially HIV-infected blood. 1) they use large nationally representa- DeSimone and Schumacher Miller measures the impact of tive samples of children from both examine the effect of HIV/AIDS midwifery-promoting public policy on countries. 2) They focus on an ADHD infection risks on the earnings of reg- maternity care in the United States, screener that was administered to all istered nurses (RNs) and other health using national Vital Statistics data on children rather than on diagnosed care workers. They combine data on births spanning 1989-99. State laws cases. 3) They address omitted vari- metropolitan statistical area (MSA) mandating insurance coverage of mid- ables bias by estimating sibling-fixed AIDS prevalence rates with annual wifery services are associated with an effects models. 4) They examine a 1987-2001 Current Population Survey 11- to 17-percentage point rise in mid- range of outcomes and compare the (CPS) and quadrennial 1988-2000 wife-attended births. The laws did not effects of ADHD to the effects of National Sample Survey of Registered increase rates of unassisted vaginal physical health conditions. 5) They ask Nurses (SRN) data. Holding constant deliveries or lead to consistent effects how the effects of ADHD and the wages of control groups that are likely on maternal mortality or APGAR probability of treatment for ADHD not exposed to AIDS risks and group- scores. They did, however, lead to a are mediated by income. Currie and specific MSA fixed effects, the authors statistically significant drop in neonatal Stabile find that ADHD symptoms find that a 10 percent increase in the deaths of about 18/100,000 births. have large negative effects on future AIDS rate raises RN earnings by about Divergence between OLS and natural test scores and schooling attainment. 0.8 percent in post-1992 samples, experiment estimates suggests that The severity of the effects and the per- when AIDS rates were falling but a women are selecting into provider vasiveness of the symptoms suggest more comprehensive categorization of groups based on unobserved prefer- that efforts to find better ways to teach AIDS was used by the CDC. AIDS ences and health. the relatively small number of children wage differentials are much larger for Currie and Stabile examine U.S. diagnosed with ADHD could have a RNs and non-nursing health practi- and Canadian children with symptoms larger payoff in terms of improving tioners than for other nursing and of Attention Deficit Hyperactivity the academic outcomes of large num- health care workers, suggesting that Disorder (ADHD), the most common bers of children with milder symptoms. this differential represents compensa- child mental health problem. This

Market Microstructure

The NBER’s Working Group and Kumar Venkataraman, Panics in a Frictionless Market with on Market Microstructure met in Southern Methodist University, Heterogeneous Expectations” Cambridge on May 6. Organizers “Market Transparency and Discussant: Bruno Biais, Toulouse Bruce Lehmann of NBER and Institutional Trading Costs” University University of California, San Diego, Discussant: David Lesmond, Tulane Duane Seppi of Carnegie Mellon Uni- University Ekkehart Boehmer, Eric Kelley, versity, and Avanidhar Subrahmanyam and Christo Pirinsky, Texas A&M of University of California, Los Mike Aitken, University of New University, “Institutional Investors Angeles, chose these papers to discuss: South Wales; Niall Almeida, and the Informational Efficiency of Deutsche Bank AG; Frederick Prices” Markus K. Brunnermeier, Harris, Wake Forest University; and Discussant: Tarun Chordia, Emory Princeton University, and Lasse Thomas McInish, University of University Heje Pedersen, NBER and New Memphis, “Order Splitting and York University, “Market Liquidity Order Aggressiveness in Electronic Robert Engle, NBER and New and Funding Liquidity” Trading” York University, and Zheng Sun, Discussant: Guillaume Plantin, Discussant: Pankaj Jain, University New York University, “Forecasting Carnegie Mellon University of Memphis Volatility Using Tick by Tick Data” Discussant: Charles Jones, Columbia Hendrik Bessembinder, H. Henry Cao, University of University University of Utah; William North Carolina, and Hui Ou-Yang, Maxwell, University of Arizona; Duke University, “Bubbles and

Brunnermeier and Pedersen’s their availability of funds. Traders pro- by their financiers. In times of crisis, model links a security’s market liquidi- vide market liquidity; their ability to do reductions in market liquidity and ty — that is, the ease of trading it — so depends on their funding, that is, funding liquidity are mutually reinforc- to traders’ funding liquidity — that is, their capital and the margins charged ing, leading to a liquidity spiral. The

46 NBER Reporter Summer 2005 model here provides a natural explana- simultaneously at multiple prices. dimension of market quality. tion for the empirically documented Active institutions submit more Boehmer, Kelley, and Pirinsky study features that market liquidity: 1) can aggressive orders than passive institu- a broad cross-section of NYSE-listed suddenly dry up (that is, is fragile); 2) tions and increase order aggressiveness stocks between 1983 and 2003, using has commonality across securities; 3) is through the day, consistent with measures of the relative informational related to volatility; 4) experiences hypothesized responses to observed efficiency of prices that are construct- “flight to liquidity” events; and 5) co- changes in execution risk and the cost ed from transactions data. They find moves with the market. Finally, the of non-execution. These findings that stocks with a higher fraction of model shows how the Fed can improve highlight the importance of limit order institutional ownership are priced current market liquidity by committing information beyond the best quotes. more efficiently, and this result is to improve funding in a potential When investors have differences robust across a variety of specifica- future crisis. of opinion about the payoffs of a tions. Moreover, they demonstrate that Bessembinder, Maxwell, and stock, Harrison and Kreps (1978) increases in actual institutional trading Venkataraman estimate trade execu- demonstrated, there is a speculative volume are associated with greater effi- tion costs for a sample of institutional bubble in the stock price: that is, the ciency, and this effect appears to be (insurance company) trades in corpo- stock price can exceed the valuation of distinct from the one associated with rate bonds before and after the initia- the most optimistic investor. A crucial cross-sectional differences in institu- tion of public transaction reporting condition supporting this result in tional holdings. for some bonds through the TRACE their model is that investors are not Using tick-by-tick data, Engle system in July 2002. Their results indi- allowed to short sell the stock. Cao and Sun build an econometric model cate a remarkable 50 percent reduction and Ou-Yang demonstrate that specu- for estimating the volatility of unob- in trade execution costs for bonds eli- lative bubbles may arise even without served efficient price change. They gible for TRACE transaction report- their short sales constraint. They also model the joint density of the marked ing; they also suggest the presence of a show that asset panics may arise: that point process of duration and tick-by- “liquidity externality” that results in a is, the stock price may be lower than tick returns within an ACD-GARCH 20 percent reduction in execution the valuations of all individual framework. First they model the dura- costs for bonds not eligible for investors. In particular, even if the tion variable as an ACD process that TRACE reporting. They estimate larg- short sales constraint binds, asset pan- could potentially depend on past er trading cost reductions for less liq- ics can arise. This result suggests that returns. Then they model the return uid and lower-rated bonds, and for Miller’s (1977) insight — that the short variable, conditioning on its current larger trades. These key results are sales constraint causes the stock price duration as well as on past informa- robust to allowances for changes in to be above the average valuation — is tion. The observed return process variables, such as interest rate volatility not robust in a dynamic framework. In admits a state space model, where the and trading activity, which might also the case of a bubble, the model here unobserved efficient price innovation affect execution costs. The authors generalizes the Harrison-Kreps notion and microstructure noises serve as also find decreased market shares for of a resale option: investors believe state variables. After adjusting for bid- large dealers and a smaller cost advan- that they can resell the stock later at a ask spread and a non-linear function of tage for those dealers post-TRACE, higher price. In the case of a panic, this durations, tick by tick returns are dis- suggesting that the corporate bond model develops the notion of a buy- tributed independently of durations, market has become more competitive back option: investors believe that they with volatility that admits a GARCH after TRACE implementation. The can sell the stock now and buy it back process. The authors apply this model point estimates show annual trading later at a lower price. The authors devel- to frequently traded NYSE stock cost reductions of roughly $1.2 billion op intuitive sufficient conditions for transactions data. Contemporaneous per year for the entire corporate bond bubbles and panics and show that under duration appears to have little effect on market, reinforcing the fact that mar- certain conditions, bubbles and panics the conditional volatility per trade, ket design can have first-order effects, can arise even when investors have which means that per second volatility even for relatively sophisticated institu- almost homogeneous expectations. is inversely related to the duration tional customers. The fraction of domestic equity between trades. This is consistent with Aitken, Almeida, Harris, and held by institutional investors has the result of Engle (2000) and Easley McInish examine the orders of active quadrupled during the past four and O’Hara (1987). The authors obtain and passive institutional investors decades, and a prominent share of a new estimate of daily, realized volatil- across five (bid) steps of the limit trading activity is attributable to insti- ity as well as the volatility of efficient order book; they also test recent theo- tutions. Yet, prior research offers price changes. Volatility is forecast ries of order aggressiveness in limit diverging views on how these develop- over calendar time intervals by simula- order placement decisions. The ments affect equity markets. In partic- tion, and the distribution of the num- authors document that institutional ular, we know little about how institu- ber of trades is central in forming investors break their desired quantities tions affect the informational efficien- these forecasts. into several limit orders submitted cy of share prices, one important

NBER Reporter Summer 2005 47 Corporate Governance

NBER researchers whose inter- Lucian Bebchuk, and Yaniv Shareholders and Costs of est is corporate governance met in Grinstein, Cornell University, “Pay Monitoring: Evidence from Cambridge on May 12. NBER and Size” Executive Compensation” Research Associate Lucian Bebchuk Discussant: Arturo Bris, Yale Discussant: Florencio Lopez-de- of Harvard Law School organized University Silanes, NBER and Yale University their program: Martin J. Conyon, John E. Core, Kose John, Lubomir Litov, and David Yermack, New York and Wayne R. Guay, University of Bernard Yeung, New York University, “Golden Handshakes: Pennsylvania, “How High is U.S. University, “Corporate Governance Rewards for CEOs Who Leave” CEO Pay? A Comparison with U.K. and Managerial Risk Taking: Theory Discussant: David S. Scharfstein, CEO Pay” and Evidence” NBER and Harvard University Discussant: Alexander Dyck, Discussant: Rene M. Stulz, NBER University of Toronto and Ohio State University David F. Larcker, Scott A. Richardson, and Irem Tuna, Lily Qui, Brown University, “Which Mihir A. Desai, NBER and University of Pennsylvania, and Institutional Investors Monitor? Harvard University; Alexander Andrew J. Seary, Simon Fraser Evidence from Acquisition Activity” Dyck, University of Toronto; and University, “Back Door Links Discussant: Stuart Gillan, Arizona Luigi Zingales, NBER and Between Directors and Executive State University University of Chicago, “Theft and Compensation” Taxes” Discussant: Henry Hansmann, Yale Andres Almazan, Jay C. Hartzell, Discussant: Enrico Perotti, University and Laura T. Starks, University of Universiteit van Amsterdam Texas, “Active Institutional

Yermack studies separation pay- are required to establish a connection are followed by increases in executive ments made when CEOs leave their between each pair of directors (ignoring pay. The authors find that, controlling firms. In his sample of Fortune 500 the obvious link that occurs when for past stock performance, the com- companies, these packages are wide- directors are on the same board). The pensation of continuing CEOs (both spread and lucrative. Almost 80 per- back door distance is one measure of in absolute terms and relative to their cent of exiting CEOs receive separa- the existence and strength of the com- compensation earlier in their service) is tion pay, and its mean present value munication channel between board positively correlated with the net exceeds $4.5 million. Severance is pos- members that can be used to influence amount of shares issued earlier during itively associated with future pay that decisions by the full board. The the CEO’s service. This effect is eco- CEOs might expect until age 65 and is authors show that CEOs at firms with nomically meaningful: holding other higher when CEOs depart involuntari- a relatively short back door distance things equal (including past stock per- ly. Shareholders react negatively when between inside and outside directors, formance), the average compensation separation agreements are disclosed, or between the CEO and the members of CEOs in the top quartile in terms but only in cases of voluntary CEO of the compensation committee, earn of net share issuances in the preceding turnover. Some evidence suggests that substantially higher levels of total com- three years is 25 percent higher than severance pay acts as a bonding device pensation (after controlling for stan- the average compensation of CEOs in between the board and CEO, while dard economic determinants and other the bottom quartile. Further, stock other evidence accords with theories personal characteristics of the CEO returns are correlated with higher sub- of rent extraction. and the structure for board of direc- sequent CEO pay only to the extent Larcker, Richardson, Seary, tors). This statistical association is con- that they contribute to firm expansion; and Tuna ask whether links between sistent with recent claims that the just the part of stock returns that is inside and outside directors have an monitoring ability of the board is ham- not distributed as dividends is correlat- impact on CEO compensation. Using pered by “cozy” and possibly difficult ed with subsequent CEO pay. The a comprehensive sample of 22,074 to observe relationships between identified link between firm expansion directors for 3,114 firms, the authors directors. under a CEO and subsequent execu- develop a measure of the “back door” Bebchuk and Grinstein study tive pay could provide CEOs with distance between each pair of directors the extent to which decisions to incentives to expand firm size, even on a company’s board. Specifically, expand the size of public companies when such expansion would not maxi- using the entire network of directors — by issuing new equity to finance mize shareholder value. and firms, they compute the minimum acquisitions or investments, or by Conyon, Core, and Guay exam- number of other company boards that avoiding dividends or repurchases — ine U.S. and U.K. CEO pay and incen-

48 NBER Reporter Summer 2005 tives from 1997 to 2003. Pay is the pension fund presence do acquire other incentives with those of the debt hold- total of cash pay, stock and option firms, they perform relatively better in ers. By the same token, higher investor grants, and other pay. The authors the long run. The opposite is the case protection is associated with riskier measure incentives as the change in for mutual fund shareholders — their investment policy and faster firm value of the CEO’s equity holdings for presence encourages acquisitions by growth. The authors test these predic- a 1 percent stock return, which equates firms with potentially higher agency tions in a large Global Vantage panel. $100 in stock to $1 in incentives. costs, and is associated with worse M They find strong empirical confirma- Controlling for firm characteristics, and A performance. tion that corporate risk-taking and U.S. CEOs have higher compensation Although evidence suggests that firm growth rates are positively related and much higher incentives. In 2003, institutional investors play a role in to the quality of investor protection. U.S. CEO pay was 1.6 times U.K. CEO monitoring management, not all insti- Desai, Dyck, and Zingales ana- pay, as compared to about 2.2 times in tutions are equally willing or able to lyze the interaction between corporate 1997. U.S. CEO incentives in 2003 serve this function. Almazan, taxes and corporate governance. They were about 5.2 times U.K. CEO incen- Hartzell, and Starks present a stylized show that the characteristics of a taxa- tives, as compared to 8.4 times in 1997. model that examines the effects of tion system affect the extraction of These narrowing differences over time institutional monitoring on executive private benefits by company insiders. result from substantial increases in compensation. The model predicts A higher tax rate increases the amount U.K. pay and incentives and flat U.S. that institutions’ influence on man- of income that insiders divert and thus pay and incentives. Although U.S. pay agers’ pay-for-performance sensitivity worsens governance outcomes. In is higher than U.K. pay, the much high- and level of compensation is reduced contrast, stronger tax enforcement er incentives imply that U.S. CEOs when institutions have greater implied reduces diversion and, in so doing, can receive less pay per unit of incentives costs of monitoring, but that these raise the stock market value of a com- (about $15 in pay for each $1 in incen- effects are attenuated when the firm- pany in spite of the increase in the tax tives) than U.K. CEOs (about $48 in specific cost of monitoring is high. burden. The authors also show that the pay for each $1 in incentives). This These empirical results are broadly corporate governance system affects ratio is roughly constant over time. consistent with the notion that inde- the level of tax revenues and the sensi- When normed by the larger amount of pendent investment advisors and tivity of tax revenues to tax changes. risky incentives held, U.S. CEO pay investment managers have advantages When the corporate governance sys- does not appear large compared to in monitoring firms’ management. tem is ineffective (that is, when it is U.K CEO pay. John, Litov, and Yeung study easy to divert income), an increase in Qiu shows that the presence of how the investor protection environ- the tax rate can reduce tax revenues. large public pension fund shareholders ment affects corporate managers’ The authors test this prediction in a will reduce firms’ acquisition activity, incentives to take value-enhancing panel of countries. Consistent with the after controlling for ownership endo- risks. In their model, the manager model, they find that rate geneity, firm-level governance provi- chooses higher perk consumption increases have smaller (in fact, nega- sions, and other firm characteristics. when investor protection is low. Since tive) effects on revenues when corpo- Public pension funds particularly perks represent a priority claim held by rate governance is weaker. Finally, this reduce the acquisition frequency by the manager, lower investor protection approach provides a novel justification cash-rich and low-q firms, and the like- leads the manager to implement a sub- for the existence of a separate corpo- lihood of “buying-growth” acquisi- optimally conservative investment pol- rate tax based on profits. tions. When firms with large public icy, effectively aligning her risk-taking

NBER Reporter Summer 2005 49 Corporate Finance

The NBER’s Program on Countries” Corporate Finance met in Bo Becker, University of Illinois, Discussant: Luigi Zingales, NBER Cambridge on May 13. Malcolm “Geographical Segmentation of U.S. and University of Chicago Baker, NBER and Harvard Business Capital Markets” School, and Jeffrey Wurgler, NBER Discussant: Jeremy C. Stein, NBER Simi Kedia, Rutgers University, and and New York University, organized and Harvard University Thomas Phillipon, NBER and this program: New York University, “The Roman Inderst, INSEAD, and Economics of Fraudulent Marianne Bertrand, NBER and Holger M. Mueller, New York Accounting” University of Chicago; Simon University, “Keeping the Board in Discussant: Chris Hennessy, Johnson and Antoinette Schoar, the Dark: CEO Compensation and University of California, Berkeley NBER and MIT; and Krislert Entrenchment” Samphantharak, University of Discussant: Dirk Jenter, NBER and Sergei A. Davydenko and Julian California, San Diego, “Mixing MIT R. Franks, London Business Family with Business: A Study of School, “Do Bankruptcy Codes Thai Business Groups and the Simeon Djankov and Caralee Matter? A Study of Defaults in Families Behind Them” McLiesh, World Bank, and Andrei France, Germany, and the U.K.” Discussant: Randall Morck, NBER Shleifer, NBER and Harvard Discussant: C. Fritz Foley, NBER and University of Alberta University, “Private Credit in 129 and Harvard University

A large fraction of business and ultimate control has been passed in industries that are heavily dependent groups around the world are run by to one of his descendant. One hypoth- on external finance. These base-line families. Bertrand, Johnson, esis that emerges from the analysis is results are robust to variation in the Samphantharak, and Schoar analyze that part of the decay of family-run definition of local market, and to the how the structure of the families groups over time may be attributable inclusion of controls for local wealth, behind these business groups affects to infighting for group resources as as well as to the use of lagged demo- the groups’ organization, governance, control becomes more diluted among graphic variables. Finally, deregulation and performance. To address this different family members. of intrastate branching has significant- question, they construct a unique Is the U.S. capital market integrat- ly cut the effect of local deposit supply, dataset of the family trees and the ed geographically? If not, does seg- suggesting that one benefit of deregu- business groups they run for 70 of the mentation affect economic outcomes? lation is improved geographical capital largest business families in Thailand. And, has deregulation changed the allocation. The authors show that the group head extent of segmentation? Becker The CEO’s control over informa- and his brothers hold the majority of attempts to answer these questions tion in the boardroom limits the family positions within each group. using demographic variation in savings board’s ability to monitor the CEO. However, there is also a positive rela- behavior. Metropolitan areas with a Inderst and Mueller examine how the tionship between family size and large fraction of seniors have higher incentives of CEOs to hide negative involvement of family members in the local volumes of bank deposits. Since information from the board depend business group, especially when the many banks are dependent on deposit on the CEOs’ compensation structure, ultimate control has passed from the financing, this is likely to affect local and how the board can structure CEO founder to one of his descendants. loan supply and economic activity. compensation optimally to mitigate Interestingly, groups that are run by Using the fraction of seniors as an the problem. The optimal compensa- larger families (more male siblings of instrument, Becker shows the effect of tion package consists of options and the group head) tend to have lower deposit supply on local economic out- severance pay. Option pay forces the performance. This negative perform- comes, including the number of firms, CEO “to put his money where his ance effect coincides with a larger the fraction of small firms, and the mouth is” by making it as costly as number of small firms in these group, number of new firms started. He possible for him to falsely assert that more fragmented internal capital mar- examines manufacturing firms in par- the firm’s future prospects are good. kets, and possibly more tunneling ticular because they are the least likely This minimizes the amount of sever- along the pyramidal structure of the to be affected by local demand condi- ance pay needed to compensate the groups. These performance and with- tions, and because data is available fur- CEO for being fired, thus minimizing in-group resource allocation effects are ther back. As predicted, deposit supply his rents. The model suggests how again especially pronounced in groups has a strong positive effect on out- deregulation and technological where the founder is no longer active comes. The supply is more important changes in the 1980s and 1990s may

50 NBER Reporter Summer 2005 have contributed to the dramatic consequences. First, the authors build find that large differences in creditors’ increase in both CEO pay and forced a model in which the costs of earnings rights across countries lead banks to turnover during the same period. management are endogenous and adjust their lending and reorganization Djankov, McLiesh, and Shleifer show that, in equilibrium, bad man- practices to mitigate the expected cred- investigate cross-country determinants agers hire and invest too much, dis- it or unfriendly aspects of bankruptcy of private credit, using new data on torting the allocation of real resources. law. In particular, French banks legal creditor rights and private and The authors then test the predictions respond to a creditor-unfriendly bank- public credit registries in 129 countries. of the model using new historical and ruptcy code by requiring more collater- They find that both creditor protection firm-level data. They show first that al than lenders elsewhere, and by rely- through the legal system and informa- periods of high stock market valua- ing on particular collateral forms that tion sharing institutions are associated tions are systematically followed by minimize the statutory dilution of with higher ratios of private credit to large increases in reported frauds. their claims in bankruptcy. Despite GDP, but that the former is relatively Then they show that, during periods of such adjustments, bank recovery rates more important in the richer countries. suspicious accounting, insiders sell in default differ substantially across the An analysis of legal reforms also their stocks, while misreporting firms three countries, with medians of 92 shows that improvements in creditor hire and invest like the firms whose percent in the United Kingdom, 67 rights and in information sharing pre- income they are trying to match. When percent in Germany, and 56 percent in cede faster credit growth. Also, credi- they are caught, they shed labor and France. Notwithstanding the low level tor rights are extremely stable over capital and improve their true produc- of creditor protection, low recovery time, contrary to the convergence tivity. In the aggregate, this model rates, and high historical bankruptcy hypothesis. Finally, the authors find seems to account for the recent period rates in France, the authors find, pre- that legal origins are an important of jobless and investment-less growth. distress loan spreads there are similar determinant of both creditor rights Davydenko and Franks study to those found in the creditor-friendly and information sharing institutions. how bankruptcy codes and creditors’ United Kingdom. They conclude that, Kedia and Philippon study the rights affect distressed reorganizations despite significant adjustments in lend- consequences of earnings manage- in different countries. Using a sample ing practices, bankruptcy codes still ment for the allocation of resources of 2280 small firms that defaulted on sharply affect default outcomes. among firms, and argue that fraudulent their bank debt in France, Germany, accounting has important economic and the United Kingdom, the authors

Economic Fluctuations and Growth Research Meeting

The NBER’s Program on Troy Davig, College of William Fluctuations” Economic Fluctuations and Growth and Mary, and Eric M. Leeper, Discussant: Steven N. Durlauf, held its summer research meeting in NBER and Indiana University, NBER and University of Wisconsin Cambridge on July 16. NBER “Fluctuating Macro Policies and the Research Associates David Laibson Fiscal Theory” Robert Shimer, NBER and of MIT and Michael Woodford of Discussant: Jordi Gali, NBER and University of Chicago, “Reassessing Columbia University organized the CREI the Ins and Outs of meeting. These papers were dis- Unemployment” cussed: Fatih Guvenen, University of Discussant: Steven J. Davis, NBER Rochester, “Learning Your Earning: and University of Chicago Simi Kedia, Rutgers University, Are Labor Income Shocks Really and Thomas Philippon, NBER Very Persistent?” Guido Lorenzoni, NBER and and New York University, “The Discussant: Jonathan Parker, NBER MIT, “Imperfect Information, Economics of Fraudulent and Princeton University Consumers’ Expectations, and Accounting” Business Cycles” Discussant: Franklin Allen, Xavier Gabaix, NBER and MIT, Discussant: Laura Veldkamp, New University of Pennsylvania “The Granular Origin of Aggregate York University

Kedia and Philippon study the nomic consequences. First they build a torting the allocation of real resources. consequences of earnings manage- model in which the costs of earnings Then they test the predictions of the ment for the allocation of resources management are endogenous and model using new data. They find that among firms. They argue that fraudu- show that, in equilibrium, bad man- periods of high stock market valua- lent accounting has important eco- agers hire and invest too much, dis- tions are followed systematically by

NBER Reporter Summer 2005 51 large increases in reported frauds. and thus initial uncertainty affects er finds a job: the ins and outs of During periods of suspicious account- decisions throughout the life cycle. unemployment. The job finding prob- ing, insiders sell their stocks, while mis- This allows one to estimate prior ability is strongly procyclical; the sepa- reporting firms hire and invest just like uncertainty from consumption behav- ration probability is nearly acyclical, the firms whose income they are trying ior observed later in life. This proce- particularly during the last two to match. When caught, these firms dure implies that 40 percent of the decades. Using the underlying micro- shed labor and capital and improve variation in income growth rates can economic data, Shimer shows that their true productivity. In the aggre- be forecast by individuals at time zero. these results are not attributable to gate, this model seems to account for The resulting model is consistent with compositional changes in the pool of the recent period of jobless and invest- several features of consumption data, searching workers, or to movements of ment-less growth. including: the substantial rise in within- workers in and out of the labor force. Davig and Leeper estimate cohort consumption inequality These results contradict the conven- regime-switching rules for monetary (Deaton and Paxson, 1994); the non- tional wisdom that guided the develop- and over the post-war period concave shape of the age-inequality ment of macroeconomic models of the in the United States and impose that profile; and the fact that consumption labor market during the last 15 years. estimated policy process on a model profiles are steeper for more highly Lorenzoni presents a model of with nominal rigidities. Decision rules educated individuals (Carroll and business cycles driven by shocks to are locally unique and produce a sta- Summers, 1991). These results bring consumers’ expectations about aggre- tionary long-run rational expectations fresh evidence from consumption data gate productivity. Agents are hit by equilibrium; lump sum tax shocks on the nature of income risk. heterogeneous productivity shocks; always affect output and inflation. In The empirical distribution of they observe their own productivity the model, tax non-neutralities arise firms is fat-tailed. Gabaix shows how, and a noisy public signal of aggregate solely through the mechanism articu- in a world with fat-tailed firm size dis- productivity. The shock to this public lated by the fiscal theory of the price tribution, idiosyncratic firm-level fluc- signal, or “news shock”, has the fea- level. The authors quantify that mech- tuations aggregate up to non-trivial tures of an aggregate demand shock: it anism and find it to be important in aggregate fluctuations. He illustrates increases output, employment, and U.S. data, reconciling a popular class of why and how this happens, and con- inflation in the short run and has no monetary models with the evidence tends that aggregate fluctuations in effects in the long run. The dynamics that tax shocks have substantial large part come from idiosyncratic of an economy after an aggregate pro- impacts. Because long-run policy shocks to firms. He shows empirically ductivity shock also are affected by the behavior determines the existence and that idiosyncratic volatility is indeed presence of imperfect information: uniqueness of equilibrium, it is possi- large enough for GDP volatility. The after a productivity shock, output ble to glean more accurate, qualitative idiosyncratic movements of the largest gradually adjusts to its higher long-run inferences from full-sample informa- 100 firms in the United States appear level, and there is a temporary negative tion in a regime-switching environment to explain about 40 percent of varia- effect on inflation and employment. than by conditioning on policy regime. tions in output and the Solow residual. The fraction of short-run fluctuations Guvenen studies the consump- This “granular” hypothesis suggests explained by the news shock increases tion-savings-behavior “heterogeneous new directions for macroeconomic with the level of idiosyncratic noise income profiles” (HIP) model (Lillard research, in particular that macroeco- and is non-monotonic in the precision and Weiss, 1979). In a life-cycle model, nomic questions can be clarified by of the public signal. For relatively high one assumes that individuals enter the looking at the behavior of large firms. levels of idiosyncratic uncertainty, labor market with a prior belief about Shimer uses readily accessible news shocks can generate realistic lev- their individual-specific profiles and data to measure the probability that an els of short-run volatility. learn over time in a Bayesian fashion. employed worker becomes unem- Guvenen finds that learning is slow, ployed and that an unemployed work-

52 NBER Reporter Summer 2005 Bureau Books

The following volumes may be ordered directly from the University of Chicago Press, Order Department, 11030 South Langley Avenue, Chicago, IL 60628-2215; 1-800-621-2736. Academic discounts of 10 percent for individual volumes and 20 percent for standing orders for all NBER books published by the University of Chicago Press are available to universi- ty faculty; orders must be sent on university stationery.

Analyses in the Economics of Aging

Analyses in the Economics of Aging, 401(k) plans. Subsequent chapters devoted to population aging and the edited by David A. Wise, is now avail- present analysis of the growth of plight of widows in India. This vol- able from the University of Chicago Medicare costs; the different aspects of ume should be of interest to any spe- Press. This NBER Conference Report disability; and the evolution of health, cialist or policymaker concerned with costs $85.00. wealth, and living arrangements over ongoing changes in savings and retire- Analyses in the Economics of Aging the lifetime. Keeping with the global ment behaviors. summarizes new research on several tradition of previous volumes, Analyses Wise directs the NBER’s popular and some less-examined top- in the Economics of Aging also includes Program of Research on the ics including retirement savings, the comparative studies on savings behav- Economics of Aging and is the John F. cost and efficiency of medical ior in Italy, the Netherlands, and the Stambaugh Professor of Political resources, and the predictors of health United States; an examination of Economy at Harvard University’s John events. The volume begins with a dis- household savings among different age F. Kennedy School of Government. cussion of the risks and merits of groups in Germany; and a chapter

International Trade in East Asia

International Trade in East Asia, such as China, Japan, Korea, and ent types of protectionism will res- edited by Takatoshi Ito and Andrew K. Taiwan. onate in particular with the financial Rose, is available from the University This volume includes twelve stud- and economic communities who are of Chicago Press for $80.00. This is ies that highlight trading practices trying to keep pace with this dramati- the 14th volume in the NBER series which exist between countries within cally altered landscape. on East Asia Seminars on Economics. as well as outside of the region. The Ito and Rose are both NBER The practice of trading across contributors bring into focus some of Research Associates in the Program on international borders has undergone a the endemic and external barriers to International Finance and Macro- number of changes, many of which international trade and discuss strate- economics. Ito is also a professor at have consequences for the world trad- gies for improving productivity and the University of Tokyo. Rose is the ing community. In International Trade in fostering trade relationships. The stud- B.T. Rocca Professor of Economic East Asia, a group of esteemed con- ies of factors that drive exports, the Analysis and Policy at Berkeley’s Haas tributors summarize the empirical influence of research and develop- School of Business. aspects of international trade that per- ment, the effects of foreign invest- tain specifically to East Asian countries ment, and the ramifications of differ-

NBER Reporter Summer 2005 53 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 sub- scribes 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 avail- able 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 num- ber(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 $2420; for academic libraries and faculty members, $1400. Higher rates apply for foreign orders. The on-line standard rate for a full subscription is $1715 and the on-line academic rate is $700. 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 May 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 dis- cussion 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 11332 Jonathan Gruber Faith-Based Charity and Crowd Out Daniel M. Hungerman During the Great Depression 11333 Charles W. Calomiris Banker Fees and Acquisition Premia for Targets in Donna M. Hitscherich Cash Tender Offers: Challenges to the Popular Wisdom on Banker Conflicts 11334 Roland G. Fryer, Jr. An Empirical Analysis of “Acting White” Paul Torelli 11335 Barry Eichengreen Is China’s FDI Coming at the Expense of Other Countries? Hui Tong 11336 Barry Eichengreen Sterling’s Past, Dollar’s Future: Historical Perspective on Reserve Currency Competition 11337 Pinka Chatteji Adolescent Drinking and High School Dropout Jeff DeSimone 11338 David G. Blanchflower The Wage Curve Reloaded Andrew J. Oswald 11339 Russell Hillberry Trade Responses to Geographic Frictions: David Hummels A Decomposition Using Micro-Data 11340 Geert Bekaert New-Keynesian Macroeconomics and the Term Structure Seonghoon Cho Antonio Moreno

54 NBER Reporter Summer 2005 Paper Author(s) Title 11341 Giancarlo Corsetti The Simple Geometry of Transmission and Paolo Pesenti Stabilization in Closed and Open Economies 11342 Scott Adams The Effects of Living Wage Laws: Evidence from Failed David Neumark and Derailed Living Wage Campaigns 11343 Dalton Conley Gender, Body Mass, and Economic Status Rebecca Glauber 11344 Kevin H. O’Rourke The Worldwide Economic Impact of the Revolutionary and Napoleonic Wars 11345 Peter Englund Adjustment in Property Space Markets: Estimates Ake Gunnelin from the Stockholm Office Market Patric H. Hendershott Bo Soderberg 11346 Benjamin M. Friedman What Remains from the Volcker Experiment? 11347 Thomas J. Kane School Quality, Neighborhoods, and Housing Prices: Douglas O. Staiger The Impacts of School Desegregation Stephanie K. Riegg 11348 Pierre Azoulay The Determinants of Faculty Patenting Behavior: Waverly Ding Demographics or Opportunities Toby Stuart 11349 Philippe Aghion Volatility and Growth: Credit Constraints and George-Marios Angeletos Productivity-Enhancing Investment Abhijit Banerjee Kalina Manova 11350 Antonio Gledson de Carvalho Venture Capital as Human Resource Management Charles W. Calomiris Joao Amaro de Matos 11351 Charles W. Calomiris Monopoly-Creating Bank Consolidation? Thanavut Pornrojnangkool The Merger of Fleet and BankBoston 11352 Michael W. Klein Studying Texts: A Gemara of the Israeli Economy 11353 Kathryn M.E. Dominguez International Borrowing and Macroeconomic Linda L. Tesar Performance in Argentina 11354 William Nordhaus The Sources of the Productivity Rebound and the Manufacturing Employment Puzzle 11355 Jeffrey A. Miron The Opium Wars, Opium Legalization, Chris Feige and Opium Consumption in China 11356 Daron Acemoglu Contracts and the Division of Labor Pol Antras Elhanan Helpman 11357 Joshua D. Coval Asset Fire Sales (and Purchases) in Equity Markets Erik Stafford 11358 Susan T. Stewart A Proposed Method for Monitoring U.S. Population Rebecca M. Woodward Health: Linking Symptoms, Impairments, Chronic David M. Cutler Conditions, and Health Ratings 11359 Benjamin F. Jones Age and Great Invention 11360 Benjamin F. Jones The Burden of Knowledge and the ‘‘Death of the Renaissance Man”: Is Innovation Getting Harder?

NBER Reporter Summer 2005 55 Paper Author(s) Title 11361 Manju Puri Optimism and Economic Choice David Robinson 11362 Jianping Mei Speculative Trading and Stock Prices: Evidence Jose Scheinkman from Chinese A-B Share Premia Wei Xiong 11363 Gary Gorton Bank Credit Cycles Ping He 11364 Gary Gorton Eat or Be Eaten: A Theory of Mergers and Merger Waves Matthias Kahl Richard Rosen 11365 Linda S. Goldberg The International Exposure of U.S. Banks 11366 Joshua Aizenman International Reserves: Precautionary versus Jaewoo Lee Mercantilist Views, Theory and Evidence 11367 Harrison Hong Asset Float and Speculative Bubbles Jose Scheinkman Wei Xiong 11368 Hietor Almeida A Theory of Pyramidal Ownership Daniel Wolfenzon and Family Business Groups 11369 Basant K. Kapur Capital Flows and Exchange Rate Volatility: Singapore’s Experience 11370 Menzie D. Chinn What Matters for Financial Development? Hiro Ito Capital Controls, Institutions, and Interactions 11371 Kishore Gawande Lobbying Competition Over U.S. Trade Policy Pravin Krishna 11372 Kristin J. Forbes The Microeconomic Evidence on Capital Controls: No Free Lunch 11373 Jiro Nemota Productivity, Efficiency, Scale Economies and Mika Goto Technical Change: A New Decomposition Analysis of TFP Applied to the Japanese Prefectures 11374 Laurence Ball Fiscal Remedies for Japan’s Slump 11375 Alessandra Casella A Simple Scheme to Improve the Efficiency of Referenda Andrew Gelman 11376 Jill R. Horwitz Does Corporate Ownership Matter? Service Provision in the Hospital Industry 11377 Jonathan Gruber Religious Market Structure, Religious Participation, and Outcomes: Is Religion Good For You? 11378 Sara Markowitz An Investigation of the Effects of Alcohol Consumption Robert Kaestner and Alcohol Policies on Youth Risky Sexual Behaviors Michael Grossman 11379 Robert S. Huckman Hospital Integration and Vertical Consolidation: An Analysis of Acquisitions in New York State 11380 Yacine Ait-Sahalia Ultra High Frequency Volatility Estimation with Per A. Mykland Dependent Microstructure Noise Lan Zhang 11381 Marcus Noland South Korea’s Experience with International Capital Flows

56 NBER Reporter Summer 2005 Paper Author(s) Title 11382 Kevin Cowan International Borrowing, Capital Controls, and Jose De Gregorio the Exchange Rate: Lessons from Chile 11383 Michael D. Bordo Historical Perspective on Global Imbalances 11384 Murray Leibbrandt Incomes in South Africa Since the Fall of Apartheid James Levinsohn Justin McCrary 11385 Paul Gompers Venture Capital Investment Cycles: The Impact Anna Kovner of Public Markets Josh Lerner David Scharfstein 11386 Raj Chetty A General Formula for the Optimal Level of Social Insurance 11387 Ajay Shah India’s Experience with Capital Flows: Ila Patnaik The Elusive Quest for a Sustainable Current Account Deficit 11388 Diego Comin The Rise in Firm-Level Volatility: Thomas Philippon Causes and Consequences 11389 John Y. Campbell Growth or Glamour? Fundamentals and Systematic Christopher Polk Risk in Stock Returns Tuomo Vuolteenaho 11390 Philip J. Cook Assigning Deviant Youths to Minimize Total Harm Jens Ludwig 11391 Viral V. Acharya Is Cash Negative Debt? A Hedging Perspective on Heitor Almeida Corporate Financial Policies Murillo Campello 11392 Lars E. O. Svensson Optimal Policy Projections Robert J. Tetlow 11393 Marc J. Melitz Market Size, Trade, and Productivity Gianmarco I. P. Ottaviano 11394 Orley Ashenfelter Evaluating the Role of Brown vs. Board of Education in William J. Collins School Equalization, Desgregation, and the Income of Albert Yoon African-Americans 11395 Thomas C. Buchmueller Price and the Health Plan Choices of Retirees 11396 Gene M. Grossman Party Discipline and Pork-Barrel Politics Elhanan Helpman 11397 John Joseph Wallis Equilibrium Impotence: Why the States and Not the Barry R. Weingast American National Government Financed 11398 Edward L. Glaeser Urban Colossus: Why is New York America’s Largest City? 11399 Alberto Alesina Corruption, Inequality, and Fairness George-Marios Angeletos 11400 Lin Peng Investor Attention, Overconfidence, and Category Wei Xiong 11401 Sergio Rebelo Real Business Cycle Models: Past, Present, and Future 11402 Alvin E. Roth Efficient Kidney Exchange: Coincidence of Wants Tayfun Sonmez in a Structured Market M. Utku Unver

NBER Reporter Summer 2005 57 Paper Author(s) Title 11403 Joshua Aizenman FDI and Trade – Two Way Linkages? Ilan Noy 11404 Andrew B. Bernard Importers, Exporters, and Multinationals: A Portrait J. Bradford Jensen of Firms in the U.S. that Trade Goods Peter K. Schott 11405 Kevin M. Murphy The Value of Health and Longevity Robert H. Topel 11406 Andrew G. Biggs Alternative Methods of Price Indexing Social Security: Jeffrey R. Brown Implications for Benefits and System Financing Glenn Springstead 11407 Dennis W. Carlton Tying, Upgrades, and Switching Costs in Michael Waldman Durable-Goods Markets 11408 Emmanuel Farhi Inequality, Social Discounting, and Estate Taxation Ivan Werning 11409 Nittai K. Bergman Employee Sentiment and Stock Option Compensation Dirk Jenter 11410 Richard B. Freeman What Do Unions Do? The 2004 M-Brane Stringtwister Edition 11411 John Cawley The Impact of State Physical Education Requirements Chad D. Meyerhoefer on Youth Physical Activity and Overweight David Newhouse 11412 Gordon H. Hanson Emigration, Labor Supply, and Earnings in Mexico 11413 Geert Bekaert Liquidity and Expected Returns: Lessons From Campbell R. Harvey Emerging Markets Christian Lundblad 11414 Jörn-Steffen Pischke Zero Returns to Compulsory Schooling in Germany: Till von Wachter Evidence and Interpretation 11415 David Popp They Don’t Invent Them Like They Used To: An Examination of Energy Patent Citations Over Time 11416 David G. Blanchflower Happiness and the Human Development Index: The Andrew J. Oswald Paradox of Australia 11417 Stephanie Schmitt-Grohé Optimal Fiscal Policy in a Medium-Scale Martín Uribe Macroeconomic Model: Expanded Version 11418 Barbara J. Spencer International Outsourcing and Incomplete Contracts 11419 Daniel P. Kessler Can Ranking Hospitals on the Basis of Patients’ Travel Distances Improve Quality of Care? 11420 Leemore S. Dafny Do Report Cards Tell Consumers Anything They Don’t David Dranove Already Know? The Case of Medicare HMOs 11421 Judith Chevalier Are Durable Goods Consumers Forward Looking? Austan Goolsbee Evidence from College Textbooks 11422 Marcelle Chauvet Dating Business Cycle Turning Points James D. Hamilton 11423 Brian Duncan Ethnic Identification, Intermarriage, and Unmeasured Stephen J. Trejo Progress by Mexican Americans

58 NBER Reporter Summer 2005 Paper Author(s) Title 11424 Daron Acemoglu Determinants of Vertical Integration: Finance, Simon Johnson Contracts, and Regulation Todd Mitton 11425 Ernst R. Berndt Opportunities for Improving the Drug Development Adrain H. B. Gottschalk Process: Results from a Survey of Industry and the FDA 11426 John R. Graham Investor Competence, Trading Frequency, and Home Bias Campbell R. Harvey Hai Huang 11427 Joseph P. Ferrie Death and the City: Chicago’s Mortality Transition, Werner Troesken 1850-1925 11428 Susan M. Richter Impacts of Policy Reforms on Labor Migration J. Edward Taylor From Rural Mexico to the United States Antonio Yunez-Naude 11429 Sebastian Edwards Establishing Credibility: The Role of Foreign Advisors 11430 Robert S. Pindyck Sunk Costs and Real Options in Antitrust 11431 John A. Rizzo Generic Scrip Share and the Price of Brand-Name Drugs: Richard Zeckhauser The Role of the Consumer 11432 Kurt Unger Regional Economic Development and Mexican Out-Migration 11433 Kelly Noonan Prenatal Drug Use and the Production of Infant Health Nancy E. Reichman Hope Corman Dhaval Dave 11434 Sebastian Edwards Capital Controls, Exchange Rate Volatility, and Roberto Rigobon External Vulnerability 11435 Joanna Lahey Age, Women, and Hiring: An Experimental Study 11436 Judith Shinogle Firms’ Demand for Employment-Based Mental Health David Salkever Benefits 11437 Ronald G. Ehrenberg Crafting A Class: The Trade Off Between Merit Liang Zhang Scholarships and Enrolling Lower-Income Students Jared M. Levin 11438 Ian W.H. Parry The Incidence of Pollution Control Policies Hilary Sigman Margaret Walls Roberton C. Williams III 11439 John Y. Campbell Caught On Tape: Institutional Order Flow and Stock Tarun Ramadorai Returns Tuomo O. Vuolteenaho 11440 Anna Pavlova Wealth of Transfers, Contagion, and Portfolio Constraints Roberto Rigobon 11441 Laurent E. Calvet Multifrequency News and Stock Returns Adlai J. Fisher 11442 Mark Carey The Risks of Financial Institutions René M. Stulz 11443 Lucian Bebchuk The Growth of Executive Pay Yaniv Grinstein

NBER Reporter Summer 2005 59 Paper Author(s) Title 11444 Anna Obizhaeva Optimal Trading Strategy and Supply/Demand Dynamics Jiang Wang 11445 Alan Krueger Race, Income, and College in 25 Years: The Continuing Jesse Rothstein Legacy of Segregation and Discrimination Sarah Turner 11446 Thomas C. Buchmueller Health Insurance Reform and HMO Penetration in the Small Group Market 11447 Jinyoung Kim The Influence of University Research on Industrial Sangjoon John Lee Innovation Gerald Marschke 11448 Andrew B. Bernard A Note on the Empirical Implementation of Lens Raymond Robertson Condition Peter K. Schott 11449 Alan J. Auerbach The 2003 Cuts and the Value of the Firm: Kevin A. Hassett An Event Study 11450 Andrew A. Toole Biomedical Academic Entrepreneurship Through the Dirk Czarnitzki SBIR Program 11451 Paul Beaudry Managerial Skill Acquisition and the Theory of Economic Patrick Francois Development 11452 Raj Chetty The Effects of Taxes on Market Responses to Dividend Joseph Rosenberg Announcements and Payments: What Can We Learn from the 2003 Dividend ? 11453 Daniel S. Hamermesh The Value of Peripatetic Economists: A Sesqui-Difference Evaluation of Bob Gregory 11454 Maria Cancian Labor Supply Effects of the Earned Income Tax Credit: Arik Levinson Evidence from Wisconsin’s Supplemental Benefit for Families with Three Children 11455 David N. Weil Accounting for the Effect of Health on Economic Growth 11456 Pablo Ibarraran Mexican Immigration and Self-Selection: New Evidence Darren Lubotsky from the 2000 Mexican Census 11457 Richard B. Freeman Does Globalization of the Scientific/Engineering Workforce Threaten U.S. Economic Leadership? 11458 Luis Garicano Organization and Inequality in a Knowledge Economy Esteban Rossi-Hansberg 11459 Evgeny Lyandres Investment-Based Underperformance Following Le Sun Seasoned Equity Offerings Lu Zhang 11460 Thomas Hellman The Role of Patents for Bridging the Science-to-Market Gap

11461 Alma Cohen Estimating Risk Preferences from Deductible Choice Liran Einav 11462 Stephen Coate Socially Optimal Districting Brian Knight 11463 Brian A. Jacob Principals as Agents: Subjective Performance Lars Lefgren Measurement in Education

60 NBER Reporter Summer 2005 Paper Author(s) Title 11464 Roland G. Fryer, Jr. Affirmative Action and Its Mythology Glenn C. Loury 11465 Fiona Murray Do Formal Intellectual Property Rights Hinder the Free Scott Stern Flow of Scientific Knowledge? An Empirical Test of the Anti-Commons Hypothesis 11466 Bruce A. Blonigen Spacey Parents: Spatial Autoregressive Patterns in Ronald B. Davies Inbound FDI Helen T. Naughton Glen R. Waddell 11467 James H. Stock Implications of Dynamic Factor Models for VAR Mark W. Watson Analysis 11468 John Y. Campbell Predicting the Equity Premium Out Of Sample: Can Samuel B. Thompson Anything Beat the Historical Average? 11469 Mihir A. Desai Taxation and the Evolution of Aggregate Corporate Ownership Concentration Winnie Fung 11470 Megan MacGarvie Early Academic Science and the Birth of Industrial Jeffrey L. Furman Research Laboratories in the U.S. Pharmaceutical Industry 11471 Richard Baldwin Heterogeneous Firms and Trade: Testable and Untestable Properties of the Melitz Model 11472 Kris James Mitchener Supersanctions and Sovereign Debt Repayment Marc D. Weidenmier 11473 Filippo Altissimo Long-Run Determinants of Inflation Differentials in a Pierpaolo Benigno Monetary Union Diego Rodriguez 11474 Muriel Niederle Do Women Shy Away From Competition? Do Men? Lise Vesterlund 11475 Richard Jensen University Invention, Entrepreneurship, and Start-Ups Celestien Chukumba 11476 Lars Peter Hansen Consumption Strikes Back? Measuring Long-Run Risk John Heaton Nan Li 11477 Sydney C. Ludvigson The Empirical Risk-Return Relation: A Factor Analysis Serena Ng Approach 11478 Michael Greenstone Mandated Disclosure, Stock Returns, and the 1964 Paul Oyer Securities Acts Amendments Annette Vissing-Jorgensen 11479 Chiara Criscuolo Global Engagement and the Innovation Activities of Firms Jonathan E. Haskel Matthew J. Slaughter 11480 Laura X. L. Liu Momentum Profits and Macroeconomic Risk Jerold B. Wagner Lu Zhang 11481 Luojia Hu Layoffs, Lemons, Race, and Gender Christopher Taber 11482 Joshua L. Rosenbloom Reexamining the Distribution of Wealth in 1870 Gregory W. Stutes

NBER Reporter Summer 2005 61 Paper Author(s) Title 11483 Jonathan Gruber Does Falling Smoking Lead To Rising Obesity? Michael Frakes 11484 Timothy Besley Political Competition and Economic Performance: Theory Torsten Persson and Evidence from the United States Daniel Sturm 11485 Wojciech Kopczuk Denial of Death and Economic Behavior 11486 David W. Galenson Filming Images or Filming Reality: The Life Cycles of Joshua Kotin Movie Directors from D.W. Griffith to Federico Fellini 11487 Junsoo Lee Nonrenewable Resource Prices: Deterministic or John A. List Stochastic Trends? Mark Strazicich 11488 Harrison Hong The Only Game in Town: Stock-Price Consequences of Jeffrey D. Kubik Local Bias Jeremy C. Stein 11489 Patric H. Hendershott The Sensitivity of Homeowner Leverage to the Gwilym Pryce Deductibility of Home Mortgage Interest 11490 Mark Aguiar Efficient Fiscal Policy and Amplification Manuel Amador Gita Gopinath 11491 David S. Lee Crime, Punishment, and Myopia Justin McCrary 11492 Guillermo A. Calvo Relative Price Volatility Under Sudden Stops: The Alejandro Izquierdo Relevance of Balance Sheet Effects Rudy Loo-Kung 11493 M. Hashem Pesaran Global Business Cycles and Credit Risk Til Schuermann Björn-Jakob Treutler 11494 Brian A. Jacob What Do Parents Value in Education? An Empirical Lars Lefgren Investigation of Parents’ Revealed Preferences for Teachers 11495 Marco Battaglini Inefficiency in Legislative Policymaking: A Dynamic Stephen Coate Analysis 11496 Paul Beaudry The “News” View of Economic Fluctuations: Evidence Franck Portier from Aggregate Japanese Data and Sectoral U.S. Data 11497 Marie Thursby Are There Real Effects of Licensing on Academic Jerry Thursby Research? A Life-Cycle View Swastika Mukherjee 11498 Thorsten Beck Bank Supervision and Corruption in Lending Asli Demirgüç-Kunt Ross Levine 11499 Luc Laeven Is There a Diversification Discount in Financial Ross Levine Conglomerates? 11500 Thorsten Beck Bank Concentration and Fragility: Impact and Mechanics Asli Demirgüç-Kunt Ross Levine 11501 David W. Galenson Toward Abstraction: Ranking European Painters of the Early Twentieth Century

62 NBER Reporter Summer 2005 Paper Author(s) Title 11502 Ross Levine Law, Endowments, and Property Rights 11503 Diego Comin A Theory of Growth and Volatility at the Aggregate and Sunil Mulani Firm Level 11504 Edward L. Maydew The Changing Role of Auditors in Corporate Tax Planning Douglas A. Shackelford 11505 Jean Helwege Why Do Firms Become Widely Held? An Analysis of the Christo Pirinsky Dynamics of Corporate Ownership René M. Stulz 11506 Jeremy Bulow Matching and Price Competition Jonathan Levin 11507 Patrick Bayer Separate When Equal? Racial Inequality and Residential Hanming Fang Segregation Robert McMillan 11508 Jeffrey A. Frankel Contractionary Currency Crashes in Developing Countries 11509 Leonid Kogan Future Prices in a Production Economy with Investment Dmitry Livdan Constraints Amir Yaron 11510 Menzie Chinn Will the Euro Eventually Surpass the Dollar as Leading Jeffrey Frankel International Reserve Currency? 11511 Edward L. Glaeser Inequality 11512 Francine D. Blau Gender and Assimilation Among Mexican-Americans Lawrence M. Kahn 11513 Raghuram G. Rajan Aid and Growth: What Does the Cross-Country Evidence Arvind Subramanian Really Show? 11514 Ginger Zhe Jin Information and Consumer Choice: The Value of Alan T. Sorensen Publicized Health Plan Ratings 11515 Florian Zettelmeyer How the Internet Lowers Prices: Evidence from Matched Fiona Scott Morton Survey and Auto Transaction Data Jorge Silva-Risso TECHNICAL PAPERS 308 Jesús Fernández-Villaverde A, B, C’s (and D)’s for Understanding VARs Juan Rubio-Ramirez Thomas J. Sargent 309 Christopher D. Carroll The Method of Endogenous Gridpoints for Solving Dynamic Stochastic Optimization 310 Atsushi Inoue A Portmanteau Test for Serially Correlated Errors Gary Solon in Fixed Effects Models 311 Atsushi Inoue Two-Sample Instrumental Variables Estimators Gary Solon 312 Timothy Conley Inference with “Difference in Differences” with a Small Christopher Taber Number of Policy Changes

NBER Reporter Summer 2005 63 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