A Service of

Leibniz-Informationszentrum econstor Wirtschaft Leibniz Information Centre Make Your Publications Visible. zbw for Economics

National Bureau of Economic Research (NBER) (Ed.)

Periodical Part NBER Reporter Online, Volume 1997

NBER Reporter Online

Provided in Cooperation with: National Bureau of Economic Research (NBER), Cambridge, Mass.

Suggested Citation: National Bureau of Economic Research (NBER) (Ed.) (1997) : NBER Reporter Online, Volume 1997, NBER Reporter Online, National Bureau of Economic Research (NBER), Cambridge, MA

This Version is available at: http://hdl.handle.net/10419/62109

Standard-Nutzungsbedingungen: Terms of use:

Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Documents in EconStor may be saved and copied for your Zwecken und zum Privatgebrauch gespeichert und kopiert werden. personal and scholarly purposes.

Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle You are not to copy documents for public or commercial Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich purposes, to exhibit the documents publicly, to make them machen, vertreiben oder anderweitig nutzen. publicly available on the internet, or to distribute or otherwise use the documents in public. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, If the documents have been made available under an Open gelten abweichend von diesen Nutzungsbedingungen die in der dort Content Licence (especially Creative Commons Licences), you genannten Lizenz gewährten Nutzungsrechte. may exercise further usage rights as specified in the indicated licence. www.econstor.eu Visit: www.nber.org WINTER 1997/8

In This Issue Program Report: Program Report Labor Economics 1

Research Summaries: Monetary Policy and Inflation Targeting 5 Labor Economics Institutions for Fiscal Stability 8 The Economics of Cities 11 EUchardB.Fveenaan* NBER Profiles 15 Conferences 17 Bureau News 26 When the NBER instituted the Labor Studies Program some 20 years ago, Current Working Papers 43 labor economics was merely a tributary of economics. The main battle­ ground of economic debate was macroeconomics, and most analyses focused on time-series data. Today, many of the big issues in economics are microeconomic labor problems, and their resolution requires analyses of large datasets. The topic that has attracted the greatest attention among NBER labor researchers is the change in the U.S. earnings distribution-the decline in the economic position of low-skilled workers relative to high-skilled work­ ers-an area in which our colleagues in-trade also have worked intensively. Many other labor issues, such as the effect of the minimum wage on employ­ ment and income, the effect of government training programs on worker skills, the relation between family background and the well-being of chil­ dren, the return to schooling and even crime, can be viewed as part of a broad concern for the causes and consequences of inequality. The high level of employment in the United States also has attracted much attention. Many Western Europeans look longingly on the "U.S. model" for its success in creating jobs. By the U.S. model, they do not mean U.S. macro or financial policy, or industrial organization, or trade policy; they mean our labor market. This is a significant change in thinking about the U.S. job mar­ ket. Until the mid-1980s or so, the United States had higher unemployment rates than most European Union countries; the rate of joblessness in West Germany did not rise above the American rate until the 1990s. An Australian economist once remarked that when he was a student, he thought economic models based on the competitive U.S. labor market proved how badly econ­ omists understood labor markets. After all, Australia and Western Europe had markedly better unemployment records than the United States. But this

*The NBER Working Papers and Technical Papers cited in brackets by number throughout this report are listed, and their abstracts and in many cases full text are available, at the NBER World Wide Web Site.

------NBER Reporter Winter 1997/8 1. is no longer the case. We may still misunderstand labor markets, but the facts that need explaining now are quite different. NBER researchers have been ex­ amining foreign labor markets, and The National Bureau of Economic Research is a private, nonprofit research economic systems more broadly, in organization founded in 1920 and devoted to objective quantitative analysis of an effort to understand differences in the American economy. Its officers and board of directors are: outcomes across countries and to President and Chief Executive Officer-Martin Feldstein cast light on the virtues and vices of Director of Finance-Sam Parker the u.s. labor market. They also have analyzed the organization of firms, BOARD OF DlRECTORS asking what leads them to treat work­

Chairman-John H. Biggs ers differently, and how labor rela­ Vice Chairman-Carl EChrist tions and personnel policy, including 'freasurer-GemldA. Polansky compensation policy, contribute to

DlRECTORSAT LARGE firm performance. Finally, mirroring the more central Peter Aldrich Stephen Friedman Peter G. Peterson role of labor issues in economic anal­ Elizabeth E. Bailey George Hatsopoulos Richard N. Rosen John Herron Biggs Karen N. Horn Bert Seidman ysis, the NBER Labor Studies Program Andrew Brimmer Lawrence R. Klein Kathleen P. Utgoff has produced an extraordinarily large Don R. Conlan Leo Melamed Donald S. Wasserman Kathleen B. Cooper Merton H. Miller Marina V. N. Whitman number of research papers since my George C. Eads Michael H. Moskow John O. Wilson last report (307 by my rough count), Martin Feldstein RobertT. Parry which made the preparation of this DlRECTORS BY UNIVERSITY APPOINTMENT article more difficult than previous reports on the Program. Because of George Akerlof, California, Berkeley Joel Mokyr,Northwestern the plethora of papers, I have picked Jagdish w. Bhagwati, Columbia Andrew Postlewaite, Pennsylvania Wtl1iam C. Brainard, Yale Nathan Rosenberg, Stanford only some of the topicS covered in Glen G. Cain, Wisconsin Harold T. Shapiro, Princeton the Program. Franklin FISher, MIT Craig Swan,Minnesota Saul H. Hymans, Michigan David B. Yoffie, Harvard Further mirroring the increased im­ Marjorie B. McElroy, Duke Arnold Zellner, Chicago portance of labor issues, the last two

DlRECTORS BY APPOINTMENT OF OTHER ORGANIZATIONS John Bates Clark Award medalists­ David Card and Kevin M. Murphy­ Marcel Boyer, Canadian Economics Assodation come from our ranks, of which we Mark Drabenstott, American Agricultural Economics Assodation are proud. Wtl1iam C. Dunkelberg, National Association Of Business Economists Gail Fosler, The Conference Board A. Ronald Gallant,American Statistical Association Robert S. Hamada, American Finance Association Inequality and RudolphA. Oswald,American Federation of Labor and Congress of Industrial Organizations Related Issues Gerald A. PolarlSky, American Institute of Certified Public Accountants There is probably not a nook or John J. Siegfried, American Economic Association Josh S. Weston, Committee for Economic Development cranny in the analysis of the rise in Gavin Wright, Economic History Assodation inequality in earnings in the United States that NBER researchers have Contributions to the National Bureau are tax deductible. Inquiries concerning contributions may be addressed to Martin Feldstein, PreSident, NBER, 1050 not explored. They have contributed Massachusetts Avenue, Cambridge, MA 02138-5398. to documenting the facts [5202, 5832, The Reporter is issued for informational purposes and has not been reviewed by 6213, 58231; to examining the effects the Board of Directors of the NBER. It is not copyrighted and can be freely repro­ duced with appropriate attribution of source. Please provide the NBER's Public on immigration and how immigrants Information Department with copies of anything reproduced. have fared in the economy [4972, Preparation of the NBER Reporter is under the supervision of Donna Zerwitz. 4955, 4866, 5454, 5763, 5927, 5388, Requests for subscriptiOns, changes of address, and cancellations should be sent 5837, 61951; considered the role of to Reporter, National Bureau of Economic Research, Inc., 1050 Massachusetts Avenue, Cambridge, MA 02138-5398. Please include the current mailing label. trade [5924, 5940, 5621, 62091, union­ ization, technology [4255, 5534, 5956, 5107, 5941, 5606, 5657, 61661, in-

2. NBER Reporter Winter 1997/8 creases in the supply of WOmen and highlighting the exceptional progress Labor Institutions the labor supply responses of the of American women. Following the completion of the family [5236, 5459]; ana looked at the How much might additional school­ NBER's Comparative Labor Project in effect of neighborhoods and ethnic ing help the workforce prospects of 1995, many NBER researchers have capital on outcomes [6175, 6176]. persons from low income back­ continued to study labor markets in Whereas in some parts of the profes­ grounds? Does class size matter in countries outside the United States sion, the inequality issue is posed student performance after graduation [5237, 50031. Robert Topel and I, solely in terms of the effects of trade and in the job market? Researchers working with Birgitta Swedenborg of versus technology, labor researchers have used data on twins, sought dis­ SNS, directed a major study of the have been looking at diverse institu­ tinct "natural experiments;' and used Swedish welfare state, that paired tional influences as well [4224, 4678, cross-state data and diverse instru­ Swedish and American researchers: 4945, 5093]. A nutshell summary of ments to study the effects of these The Welfare State in Transition: Re­ this research is that there is convinc­ inputs on educational outcomes forming the Swedish Model published ing evidence that most things that we [4874, 5144, 5708, 5450, 6051, 5331, by the University of Chicago Press for expect to matter do in fact matter; 5353, 5274, 5288, 5548]. One impor­ the NBER in 1997. This project high­ that no single factor can explain the tant finding is that the instrumented, lighted both the pluses of the Swed­ pattern of rising inequality; and that or natural experiment, estimates of ish model (conquering poverty) and consensus about the relative magni­ school effects on earnings show that the negatives (in the form of micro­ tudes of different factors has not they remain relatively high. The ef­ economic inefficiencies and the been reached. As in much economic fect of class size and resources on inability to escape high levels of unem­ analysis that seeks to determine the schooling is more controversial; there ployment in the 1990s). John Abowd, "sources of... ", there is a sizeable is no consensus there. Cecilia Rouse working with Francis Kramarz, David residual, leaving an open field for has given a modestly positive assess­ Margolis, and others has exploited judgement calls. An important tool in ment of the effects of the Milwaukee fairly unique data files from France these analyses has been the NBER's parental choice program on student that follow workers across firms CD-ROM on the out-going rotation achievement [5964]. [4917, 5077, T180, 5493, 5551, 6109, group of the Current Population How much might family back­ 6110], allowing the researcher to con­ Survey. ground matter in economic differ­ trol for firm and worker effects and Women have been the "exception ences? Joseph Altonji, working with thus to learn more about how the job to the rule" of rising earnings differ­ various co-authors, has examined market functions. In 1995, Abowd entials; they have improved their both the effects and mechanisms by helped organize an international con­ position in the job market, particu­ which family characteristics influence ference on the use of matched em­ larly at the higher skill levels. A sig­ the young [5072, 5378, 5522]. Derek ployee-employer panel datasets. Other nificant number of researchers have Neal has presented evidence that researchers have examined German explored the improved position of "pre-market factors" are very impor­ institutions in SOme detail [5988, women in the job market and the tant in black-white differences [5124]. 4808, 4825, 5716, 5724, 6167, 5208, particular problems they face. These Several researchers have examined 5829] and have contrasted the United researchers have sought new ways to childbearing issues [5807, 5781] and States, Canada, and France [5487]. In assess issues relating to the determi­ the effect of various programs and winter 1996 the NBER held a confer­ nants and consequences of child­ interventions on children's well-being ence, organized by David Blanch­ bearlng[4911,4224, 5664,5406, 5778, [5805, 59851. Janet Currie currently flower and me, on youth labor mar­ 6047, 6034]; examined the effect of heads the NBER's Program on (the kets. It contrasted the United States school content [5580] and faculty Economics oD Children. and Germany, the United Kingdom, composition [4874] on female prog­ What is the link between crime and Sweden [6031, 6078, 6102, 6105, ress in the job market; and looked at and economic problems? Several re­ 6111,6142,6212]' the changing pattern of organizing searchers have explored issues relat­ Looking at U.S. institutions per se, careers and families [5188], and why ing to crime, examining the distribu­ Patricia Anderson and Bruce Meyer women are underrepresented in the tion of crimes across cities [5430]; the have focused on unemployment in­ field of economics [5299]. Francine relation between crime and wages surance, which has a surprisingly low Blau and Lawrence Kahn have linked [5983]; domestic violence [4939]; and take-up rate in this country [4787, gender pay differentials to the overall employment and crime [4794, 4910, 4960] while Kate Baicker, Claudia wage structure across countries [5664], 5451]. Goldin, and Lawrence Katz [5889]

NBER Reporter Winter 1997/8 3. have studied the development of that Econometric Issues tutions ..without some (albeit very dif­ system over time. Jonathan Gruber, ferent) theoretical basis, perhaps we Huge datasets raise new potential with co-authors, has examined the will see labor researchers contribut­ for statistical testing and open the implications for labor supply of vari­ ing more along the theoretical than door for new strategies for determin­ ous social insurance programs [6041l, they have in the past. while John Bound with co-authors ing behavioral responses to eco­ has focused on disability insurance nomic incentives. A striking pattern LIST OF SELECTED [5159, 5536, 5169l. in much empirical work is the search for appropriate "instruments" from REFERENCES Labor Demand and which to infer behavior. What re­ For the rising pay of higher level Firm Behavior searchers do is seek out factors that workers, see: shift supply or demand incentives Many researchers have examined WP 6213,10/97, Brian]. Hall and Jef­ without directly affecting the relevant labor demand behavior and the in­ frey B. Liebman outcomes. Joshua Angrist, Guido ternal organization of firms. Lawrence "Are CEOs Really Paid Like Bureau­ Imbens, and Don Rubins have greatly Katz has shown that wage subsidies crats?" enhanced our understanding of the can modestly improve the demand advantages and limitations of instru­ For the contribution of institu­ for disadvantaged workers [5679]. mental variable analyses with a set of tions to the distribution of wages, Daniel Hamermesh has examined the see: demand for hours of labor [4394, papers developing the notion of a 59731. Michael Kremer and Eric local average treatment effect, or WP 5093, 4/95, John DiNardo, Nicole Maskin offered a demand-side analy­ LATE [T118, T127, T181, 51921. M. Fortin, and Thomas Lemieux sis linking rising inequality to segre­ Angrist, Krueger, and Imbens have "Labor Market Institutions and the gation by skill [5718]. Two studies examined different ways to use Distribution of Wages, 1973-1992: A have focused on worker characteris­ instrumental variables [T150, T181, Semiparametric Approach" T172l while John Bound and David tics, one on their impact on plant­ For the effect of institutions on Jaeger have pOinted out problems level production [5626l and one on productivity and firm perform­ when instruments are only weakly the effect of affirmative action on ance, see: employee qualifications [5603], correlated with the explanatory vari­ WP 5333, 11/95, Casey Ichniowski, Several researchers have looked at able [5835l. James Heckman and co­ Kathryn Shaw, and Giovanna Pren­ intrafirm issues, ranging from the authors have provided insightful nushi theory of works councils and worker analyses of the problems with social "The Effects of HUman Resource share ownership and worker cooper­ experiments [T166, T184, 5535]. Management Practices on Productivity" atives [4918, 5436, 6118l to other organizational issues [5705, 5802l to What Next? WP 5436, 1/96, Edward P. Lazear and empirical studies of firm performance A visitor to labor studies from the Richard B. Freeman under alternative structures [6120, rest of economics will notice imme­ "Relational Investing: The Worker's 5672l. Doug Kruse and Joseph Blasi diately that the field is strikingly Perspective" empirically-oriented, with researchers have summarized what we know For the debate over the role of reacting to ongoing social problems from many studies of the links school resources on education, and devoted to the "facts, ma'am, just among employee ownership and see: firm performance and employees' the facts." This is a huge strength, but attitudes [5277l. Casey Ichniowski, also in some ways a weakness. In the WP 5708, 8/96, David Card and Alan Katharine Shaw, and Giovanna Pren­ future, I expect more attention to be Krueger nushi have provided evidence that paid to the effects of trade on the "School Resources and Student Out­ packages of human resource prac­ labor market and to the contribution COmes: An Overview of the Literature tices add more to firm productivity of labor markets to the distinct macro­ and New Evidence from North and than individual practices [5333]. economic performance of the United South Carolina" Robert Gibbons and Henry Farber States, which SOme researchers have WP 5288, 10/95, James Heckman, .organized a 1996 NBER-Universities explored [5822, 5538, 5538l and to Anne Layne-Farrar, and Petra Todd Research Conference around the is­ further work On firms and institu­ "The Schooling Quality-Earnings sue of the internal Structure of firms, tional differences among countries. Relationship: Using Economic Theory and this promises to be a growing As it is difficult to analyze trade, to Interpret Functional Forms Con­ topic in future years. macro issues, firms, or country insti- sistent with the Evidence" I 4. NBER Reporter Winter 1997/8 J WP 5548, 4/96, Eric A. Hanushek, "What We Know and Do Not Know Impact of U.S. Unemployment Com­ Steven G. Rivkin, and Lori L. Taylor About the Natural Rate of Unemploy­ pensation" "Aggregation and the Estimated ment" WP 6041, 5/97, Jonathan Gruber and Effects of School Resources" For social insurance issues, see: Aaron Yelowitz "Public Health Insurance and Private For macroeconomic issues, see: WP 5889, 1/97, Katherine Baicker, Savings" WP 5822, 11/96, Olivier Blanchard Claudia Goldin, and Lawrence F. Katz and Lawrence F. Katz "A Distinctive System: Origins and

Research Summaries

Mo~etary Policy and Inflation Targeting Lars E.O. Svensson* research in the last few years has Inflation Targeting largely dealt with understanding in­ as a Remedy Against In the 1990s, several countries flation targeting in relation to other shifted to a new monetary policy monetary policy regimes and investi­ High Inflation regime: an announced quantitative gating how practical monetary policy Inflation targeting can be seen as a inflation target. The reason for this can best be conducted under infla­ potential remedy for persistent high shift was the unsatisfactory perform­ tion targeting. inflation. Other remedies discussed ance under previous regimes. New Practical inflation targeting has sev­ and suggested in the literature in­ Zealand, Canada, Australia, and eral common characteristics: 1) an clude: 1) accepting that the long-run Spain all introduced inflation targets announced quantitative inflation tar­ Phillips curve is vertical and implic­ under persistently high inflation; the get, varying across countries between itly, or explicitly, setting any output United Kingdom, Sweden, and Fin­ 1.5 and 2.5 percent per year, in most or employment target equal to (rather than above) the "natural" level; 2) land did so after having abandoned countries with a tolerance band of fixed exchange rates, which had plus/minus 1 percentage point creating an independent and conser­ vative central bank; and 3) setting up failed to achieve low and stable infla­ around the target; 2) no explicit rule a performance contract (an "inflation tion and had been subject to dra­ on how the central bank shall set its contract") for the central bank gover­ matic speculative attacks. Inflation instrument; 3) a floating exchange nor or governing board. In one of my targeting has received much recent rate (except for Finland and Spain, papers, I examine the relation be­ attention, both among policymakers which are members of the Exchange tween inflation targeting and these and academics. In the United States Rate Mechanism, although the wide remedies. Inflation targeting indeed and in Europe it is debated as a pos­ exchange rate bands there so far sible monetary policy strategy for the can involve elements of all three have not created any conflict between remedies. By announcing a rather low Federal Reserve System and the the inflation target and the exchange future European Central Bank, respec­ inflation target and creating some rate target); and 4) a high degree of degree of commitment to it, inflation tively. Academic research on inflation transparency and accountability. Com­ targeting, both theoretical and empir­ targeting can help to reduce inflation, mentators also often describe infla­ even if an inflationary bias remains, ical, has grown qUickly.! My own tion targeting as a regime without and if inflation more often exceeds an intermediate target for monetary than falls short of the target. This cre­ * Svensson is a professor of international policy (instead, targeting inflation ates a "conservative" central bank in economics at the Institute for Interna­ "directly"). I have argued in some of the sense of having a lower inflation tional Economic Studies, Stockholm my research that this is misleading target rather than, as is common in University, and a Research Associate in and that inflation targeting actually the NBER's Programs on Asset Pricing, the literature since Rogoff's classic International Finance and Macroeco­ implies a particular intermediate tar­ 1985 article, identifying "conser­ nomics, and Monetary Economics. He is get, namely the central bank's infla­ vatism" with a larger weight on a profiled in this issue. tion forecast. given inflation target.

NBER Reporter Winter 1997/8 5. Incidentally, this interpretation of The conventional wisdom is that monetary policy, and sometimes with conservatism solves an empirical price-level targeting would lead to in­ a shorter lag than monetary policy. puzzle about independent central creased inflation variability, as exces­ Given these lags and imperfect banks, inflation, and output variabil­ sive inflation eventually would be control, the central bank necessarily ity. If independent central banks are followed by too little inflation in must adopt a forward-looking per­ more conservative in that they give order to get the price level back in spective, attempting to control infla­ more weight to a specific inflation line. Such variability might then show tion one to two years ahead. Forecasts target, then lower inflation should be up in increased output variability. (projections) of crucial macrovari­ correlated with higher variability of Closer study reveals that this issue abIes become central, and inflation output. A large literature instead has is more complicated. In one of my targeting becomes "inflation-forecast stated that more independent central papers, I show that price-level tar­ targeting": the bank's internal infla­ banks in industrialized countries are geting very well may succeed in tion forecast, conditional on current associated with lower inflation rates, achieving lower variability of both the information and a given path for but not with higher variability of out­ price level and inflation, when the the monetary policy instrument, be­ put. This finding is instead consistent different incentives for monetary pol­ comes the intermediate target. If the with independent central banks sim­ icy under inflation and price-level tar­ inflation forecast is above (below) ply having lower inflation targets. 2 geting, as well as the different expec­ the inflation target, monetary policy tations of future inflation and price should become more restrictive levels, are taken into account. Experi­ (expansionary) . Price-Level Targeting ments in large empirical macro mod­ The effect on the conditional infla­ versus Inflation els also have produced this result. tion forecast is also the main decision Targeting At present, more than half a dozen criterion when new information countries practice explicit inflation arrives. If the new information is Inflation targeting implies "base targeting (and certainly quite a few deemed to shift the inflation forecast drift" of the price level, even if the practice implicit inflation targeting, at a horizon of one to two years, the target is set at zero: if inflation over­ including Germany, the United States, policy instrument should be adjusted shoots its target, then the target for and Switzerland). But there is only to dampen or nullify that shift. If the the next period is related to the new one historical example of price-level new information has no effect on the price level. This base drift means that targeting: the successful but short forecast, there is no need to react to the price level has a unit root; it also experiment in Sweden in the 1930s. it. In practice, inflation-targeting cen­ means that the variance of the future In the next few years, a move to tral banks construct their forecasts price level increases without bound inflation targeting may be sufficiently partly from structural models, partly with the horizon. Therefore, to say challenging for central banks. In from forecasting models, but also that (successful) inflation targeting about another decade, when central from judgements and extraneous infor­ leads to "price stability" is not quite banks hopefully master all the intri­ mation. Thus, inflation targeting uses correct. Nevertheless, the terminol­ cacies of inflation targeting, the time all relevant information.4 ogy has stuck. might be ripe for seriously consider­ Therefore, the implicit instrument Genuine price-level targeting is dif­ ing the pros and cons of the poten­ rule that follows from inflation fore­ ferent: monetary policy then aims at tially more demanding alternative: cast targeting generally will differ keeping the price level constant, or price-level targeting.3 from the well-known Taylor Rule, ac­ around a steady increasing path. cording to which the monetary policy Price-level targeting need not imply instrument would react only to cur­ zero inflation, if a positive inflation rent inflation and output. rate is deemed desirable. The big dif­ Implementing Inflation ference vis-a-vis inflation targeting is Targeting that the variance of the price level How can inflation targeting over­ Strict or Flexible does not increase with the horizon. come the major difficulty that central Inflation Targeting? Thus, the uncertainty about the price banks do not have perfect control Under inflation targeting, what is level in the distant future is less than over inflation? Inflation reacts with the scope for stabilizing macrovari­ under inflation targeting, which "long and variable lags" and with abIes other than inflation: for instance, should facilitate long-term decisions variable magnitude to changes in the output, employment, or the real ex­ about savings and investment, and monetary policy instrument. Inflation change rate? Under "strict" inflation improve resource allocation. also is affected by factors other than targeting, the central bank is only

6. NBER Reporter Winter 1997/8 concerned with achieving the infla­ rather than domestic inflation (in Transparency allows the private tion target; under "flexible" inflation most cases some specific compo­ sector to better assess both the com­ targeting, the central bank is also, to nents are excluded from the index, petence of the central bank and its some extent, concerned with the sta­ for instance mortgage costs). Flexible commitment to the inflation target. If bility of output and/or the real ex­ CPI-inflation targeting appears to be the bank's competence and commit­ change rate. If inflation has deviated better than targeting domestic infla­ ment are deemed adequate, its cred­ from its target, under strict inflation tion when it comes to stabilizing both ibility improves, and it is easier for targeting the bank tries to get infla­ domestic inflation and real exchange the bank to fulfill its target, since pri­ tion back to target as quickly as pos­ rates.? vate sector price- and wage-setting sible. This requires considerable in­ then adapts to the target. At the time, strument movements which also are Monitoring Inflation a lack of transparency may give the likely to move output or real exchange Targeting bank more discretion to pursue any rates. Under flexible inflation target­ As mentioned earlier, inflation-target­ idiosyncratic goals. The incentive for ing, concern about output and real ing regimes may entail a high degree the bank to make monetary policy exchange rate variability would lead of transparency and accountability. more or less transparent thus depends the bank to take inflation back to the Inflation-targeting central banks reg­ in an intricate way on its competence target at a more gradual pace. ularly issue "Inflation Reports;' explain­ and its commitment. Since transpar­ Indeed, I find that concern about out­ ing and motivating their policy to the ency normally seems to be sOcially put and real exchange rate variability general public. In New Zealand, the desirable, conflicts of interest be­ translates into targeting inflation at a Reserve Bank Governor's job is at tween the bank and society cannot longer horizon, say 2.5 years rather risk if inflation is higher than 3 per­ be excluded.9 than 1. 5 years. 5 cent per year or lower than zero. In Concern about output and real ex­ the United Kingdom, the Chancellor Still Too Early to Tell change rate variability is not the only of Exchequer recently announced Explicit inflation targeting appears reason for a longer horizon and a that if inflation deviates more than to have many advantages compared more gradual adjustment of inflation 1 percentage point from the inflation to the available alternatives. Mone­ towards the target. Uncertainty about target, the Bank of England's Gover­ tary policy becomes goal-directed, the lags and magnitudes in the trans­ nor must explain in an open letter incentive-compatible, and transpar­ mission mechanism, that is, model why the divergence has occurred and ent. Yet, flexible inflation targeting uncertainty, as well as concern about what steps the Bank is taking to deal allows some concern about stability interest variability (central banks with it. In the other inflation-target­ of output, employment, and real ex­ seem eager to avoid whip-sawing the ing countries, the central bank's change rates to influence policymak­ interest rate and prefer considerable governor and board certainly suffer ing. Inflation-targeting central banks smoothing) produce the same results.6 considerable embarrassment and crit­ are improving their ability to control Hence, strict inflation targeting is an icism when inflation moves outside inflation. More research adds to the extreme case. Indeed it appears that its designated tolerance interval. understanding of the strong and weak . real-world central banks pursue flex­ An explicit inflation target and an sides of this regime, and to the cen­ ible inflation targeting and to some informative inflation report make it tral bankers' knowledge of how to extent, stabilize output and real ex­ relatively easy to monitor central­ best operate it. Still, these regimes are change rates, or at least smooth in­ bank performance. The quality and very young; the oldest one, in New terest rates. All inflation targeting results of the bank's analysis can be Zealand, is barely 7 years of age. Any economies are very open. In an open scrutinized by external experts and evaluation must be highly prelimi­ .economy, the exchange rate provides observers in order to discover biased nary; we will have to wait for several an additional channel for the trans­ arguments or wishful thinking. Even more years of data, including several mission of monetary policy. There is if the bank chooses to - or is not business cycles, until we can make a also a choice between targeting required to-publish any inflation very reliable evaluation. Meanwhile, domestic inflation (in the GDP defla­ report at all, interested observers can inflation targeting will provide ample tor, for instance) or CPI inflation (the collect inflation forecasts from rep­ opportunities for more research. latter also takes the prices of im­ utable external forecasters and can ported final goods into account). All check whether they are in line with 1 See, for instance, 1. Leiderman and inflation targeting countries have the inflation target at an appropriate I.E.O. Svensson, Inflation Targets, CEPR, opted for targeting CPI inflation horizon.8 London, 1995; A.G. Haldane, Targeting

NBER Reporter Winter 1997/8 7. Inflation, Bank of England, London, and its Consequences for Monetary Pol­ Svensson, "Estimating the Term Structure 1995; Achieving Price Stability, Federal icy, M. King and G. Calvo, eds., London, of Interest Rates for Monetary Policy Reserve Bank of Kansas City, 1996, and MacMillan, 1997. AnalYSiS," Scandinavian Journal of Eco­ B.S. Bernanke and F.S. Mishkin, "Infla­ 3 S. Fischer, "Modern Central Banking," in nomics 98 (1996), pp. 163-83; L.E.a. Non Targeting: A New Framework for The Future of Central Banking, Cam­ Svensson, "Estimating and Interpreting Monetary Policy," Journal of Economic bridge: Cambridge University Press, 1994, Forward Interest Rates: Sweden 1992-94," Perspectives 11 (1997),pp. 97-116. expresses the conventional wisdom on NlJER Working Paper No. 4871, Septem­ ber 1994; and L.E.O. Svensson, "estimat­ 2 The "inflation bias" result was demon­ price-level targeting vs. inflation target­ strated by F. Kydland and E. Prescott, ing. The conference volume Economic ing Forward Interest Rates with the ex­ "Rules Rather Than Discretion: The Behavior and Policy Choice under Price tended Nelson & Siegel Method," Quarterly Inconsistency of Optimal Plans," Journal Stability, Bank ofCanada, 1994, contains Review 3 (1995), pp. 13-26, Sveriges of Political Economy 85(3) (1977), pp. several papers on price-level targeting. Riksbank. 473-90 and R. BarroandD. Gordon, "A /. Fisher, Stable Money: A History of the 5 These results are derived and discussed Positive Theory of Monetary Policy in a Movement, Allen & Unwin, London, in "Inflation Forecast Targeting: Imple­ Natural Rate Model," Journal of Political 1935, and L. jonung, "Kurt Wicksell's menting andMonitoring Inflation Targets," Economy 91 (1983), pp. 589-610. The Norm of Price Stabilisation and Swedish European Economic Review 41, op. cit. result on a "weight-conservative" central Monetary Policy in the 1930s," Journal of G. Rudebusch andL.E.O. Svensson, "Prac­ bank is due to K. Rogoff, "The Optimal Monetary Economics 5 (1979), pp. 459-96, tical Inflation Targeting," mimeo, 1997, Degree of Commitment to a Monetary discuss the Swedish experience in the examine different forms of inflation tar­ Target," Quarterly Journal of Economics 1930s. L.E.O. Svensson, "Price-Level Tar­ geting for the U.S. 100 (1985), pp. 1169-90. Inflation con­ geting vs. Inflation Targeting," NBER Wom­ 6 These results are derived and discussed tracts are examined by G.E. Walsh, "Opti­ ing Paper No. 5719, August 1996, com­ in L.E.O. Svensson, "Inflation Targeting: mal Contracts for Independent Central pares price-level targeting and inflation Some ExtensiOns," NBER Working Paper Bankers," American Economic Review 85 targeting. No. 5962, March 1997. The result for (1995), pp. 150-67, and T. Persson and 4 See L.E.O. Svensson, "Inflation Forecast model uncertainty follows W. Brainard, G. Tabellini, "Designing Institutions for Targeting: Implementing and Monitoring "Uncertainty and the Effectiveness of Monetary Stability," Carnegie-Rochester Inflation Targets," European Economic Policy," American Economic Review 57, Conference Series on Public Policy 39 Review 41 (1997), pp. 1111-46. Under Papers and Proceedings (1967), pp. (1993), pp. 53-84. I.E.O. Svensson, inflation targeting, central banks will use 411-25. "Optimal Inflation Targets, 'Conservative' a variety of indicators to extract useful 7 These and other preliminary results for Central Banks, and Linear Inflation Con­ information. A survey of ways of extract­ an open economy are reported in L.E.O. tracts," American Economic Review 87 ing information from finanCial markets Svensson "Open-Economy Inflation Tar­ (1997),pp. 98-114, relates inflation tar­ is presented in P. S6derlind and L.E.O. geting," mimeo, 1997. gets to these results. Svensson, "New Techniques to Extract 8 These issues are further discussed in The incentives to high inflation that a Market Expectations from Financial "Inflation Forecast Targeting: Inplement­ large nominal public debt creates are Instruments," NBER Working Paper No. ing and Monitoring Inflation Targets," examined, with Sweden as an example, 5877,january 1997,forthcoming inJour­ European Economic Review, op. cit. in M. Persson, T. Persson, and L.E.O. nal of Monetary Economics 40(2) (1997). 9 These issued are clarified and discussed Svensson, "Debt, Cash Flow and Inflation PreViously, I have worked on practical more rigorously in j. Faust and L.E. O. Incentives: A Swedish Example," NBER ways of estimating and interpreting for­ Svensson, "Credibility and Transparency: Working Paper No. 5772, September ward interest rates for monetary policy Monetary Policy with Unobservable Goals," 1996, forthcoming in The Debt Burden purposes, see M. DahlqUist and L.E.O. mimeo, 1997.

Institutions for Fiscal Stability Alberto Alesina* which are unprecedented, except for abandoned fiscal discipline? What the aftermath of major wars. Low explains the very large cross-country The seventies and most of the public savings have been at the root variance in fiscal stance? Why did eighties have been a period of fiscal of Latin America's "lost decade;' the large and persistent deficits appear in profligacy in many countries around eighties. Currently, the goal of achiev­ the mid-seventies and not before? the world. Several OECD countries ing and maintaining fiscal stability is What explains the likelihood of suc­ have accumulated debt/GDP ratios the main macroeconomic issue in cess of fiscal adjustments? Why did many parts of the world. certain countries make swift and very .. Alesina is a Research Associate in the This evolution of fiscal policy successful fiscal consolidations, while NBER's Program on Monetary Economics raises many intriguing questions: others are still struggling? and a professor at Harvard University. Why have many but not all countries The answers to all of these ques- His profile appears later in this issue.

8. NBER Reporter Winter 1997/8 tions cannot rely purely on economic outcomes. We define budget insti­ fiscal rules enforce fiscal discipline factors since economically similar tutions as all of the rules and regula­ without any apparent negative effect countries exhibit very large differ­ tions according to which budgets are on state output volatility. ences in fiscal performance. In a drafted, approved, and implemented. series of recent papers, I have ad­ We focus mainly on three issues: bal­ Procedural Rules dressed these questions by consider­ anced budget rules; procedural and One can identify three phases in ing politico-institutional explanations. voting rules; and transparency. the budget process: the formulation In a paper! coauthored with Ro­ of a budget proposal within the exec­ berto Perotti, I identify two critical Balanced Budget Rules utive; the approval of the budget in institutional variables as important Well-known economic arguments the legislature; the implementation of determinants of the fiscal policy suggest that balanced budget rules the budget within the bureaucracy. In stance: the degree of government are not optimal, because they do not my research I have focused almost fragmentation, and the nature of bud­ allow deficits to fluctuate over the exclusively on the first two points. get institutions. In terms of the for­ cycle and in the event of major and Voting procedures can be classified mer variable, we argue that coalition temporarily high spending needs. on a hierarchical-collegial dimension. governments are more likely to delay However, since for many political Hierarchical procedures are those the adoption of stabilization policies, reasons politicians may have incen­ that, for instance, attribute strong pre­ because of inter-coalition struggles tives to run excessive deficits, bal­ rogatives to the Prime Minister (or leading to legislative deadlocks. Thus, anced budget rules may serve the Treasury Minister) to overrule spend­ after the oil shocks of the seventies, purpose of correcting a politically ing ministers in intragovernmental countries ruled by fragmented coali­ induced distortion in fiscal policy. negotiations. Also, hierarchical insti­ tions reacted more slowly and less However, an unpleasant conse­ tutions limit the latitude of legislative decisively, letting budget deficits quence of balanced budget rules is amendments on the budget. For in­ accumulate. In a second paper,2 we that they generate incentives for cir­ stance, in some cases the legislature show that when coalition governments cumventing them. In so doing, policy­ can change the proposed budget actually attempt to stabilize the bud­ makers engage in tactics of creative without affecting the balance. Even get, they often fail because they do accounting which make the budget more stringent rules require the leg­ not have the political strength to deal less transparent, creating additional islature not to increase either the with structural budget cuts in social obstacles to an effective control on level of spending or the deficit. Thus, spending and government wages. fiscal discipline. in the latter case, the legislature can While this argument about govern­ For all of these reasons, balanced only change the budget allocation ment fragmentation is relatively well budget rules at the national level may between spending programs. Col­ understood, the issue of budget insti­ be counterproductive. Instead, fiscal legial institutions have the opposite tutions is more complex and multi­ discipline can be enhanced by ap­ features: they emphasize "consensus" faceted. In the last few years a vast propriate procedural rules (discussed at every stage of the process, by en­ research program to which I have later) which do not require a numer­ hancing the prerogative of all spend­ contributed has investigated how dif­ ical target on the budget balance. ing ministers in the government, the ferent procedures influence fiscal The same argument, however, may prerogative of the legislature vis-a-vis outcomes, both in the OECD group not apply to subnational levels of the government, and generally by of countries, and in a sample of Latin government, such as American states. upholding the right of the minority in American countries and the American Local and state governments may every stage of the process. states. This research effort leads to need less flexibility because their One can identify a tradeoff be­ the conclusion that budgetary institu­ budget are less cyclically sensitive. tween the two types of institutions. tions do matter as a determinant of Several authors have investigated the Hierarchical institutions are more fiscal outcomes, and therefore differ­ effects of restrictions on budget likely to enforce fiscal restraint, to ent choices about fiscal institutions deficits in American states: this litera­ avoid large and persistent deficits, may lead to a higher or lower pro­ ture concludes that more stringent and to promote swift fiscal adjust­ pensity to run excessive deficits. balanced budget rules lead to more ments when needed. On the other In a third paper,3 Perotti and I iden­ fiscal discipline. My own contribution hand, the same institutions are less tify several theoretical and empirical to the literature includes a paper with respectful of the prerogative of the issues which are central for the dis­ Tamim Bayoumi of the IMF.4 In this minority not in the government, and cussion of how institutions affect fiscal article, we show that more stringent therefore are more likely to generate

NBER Reporter Winter 1997/8 9. budgets tilted in favor of the govern­ beliefs of the taxpayers-voters: 1) tors of all Latin American countries, mental coalition. Collegial institutions Overestimation of the expected and the text of the budget laws, we have the opposite features. growth in the economy, so as to constructed a comprehensive index One related important issue con­ overestimate tax revenues, and to which summarizes several character­ cerns the order of voting. In some underestimate the level of interest istics of fiscal procedures, along the cases the budget procedures imply rates so as to underestimate outlays. "hierarchical transparent" to the "col­ that the legislature first has to ap­ At the end of the fiscal year, the "un­ legial non-transparent" dimensions. prove a balance, often in the context expected" deficits can be attributed We then discussed the relationship of a macroeconomic scenario for the to "bad luck:' 2) Over optimistic fore­ between the index and various com­ coming fiscal year. Then, in later casts of the budget effects of various ponents of it, and the level and evo­ votes the allocation among different policies. 3) Strategic use of what is lution of budget deficits in this region. programs is decided. The alternative kept in and out of the budget, often We also examined cases of changes procedure implies that the balance of with a creative use of the budget of in procedures, namely whether one the budget is the residual of a series various public organizations. 4) Stra­ can detect a difference in the fiscal of votes on specific programs. The tegic use of multi-year budgets, to the position of a country before and after Budget Act of 1974 in the United effect that difficult policies are per­ a reform of its fiscal procedures. Our States implied, among other things, a manently postponed to year two or results confirm that budget proce­ switch from the latter system to the three of a multi-year program and dures do matter. After controlling for former. Intuitively one would think always delayed. several economic determinants of that the system where the balance is Issues of transparency and creative budget deficits, our index of proce­ voted first should promote more fis­ accounting are, in fact, at the fore­ dures was still significantly correlated cal restraint. Indeed, this is what the front of the fiscal debate in Europe. with budget deficits in the expected cross-country empirical evidence The discussion about which coun­ direction. A particularly important seems to suggest. However, the the­ tries can join the European Monetary feature of such procedures is the one oretical underpinnings for this result Union has paid much attention to that requires a vote on the size of the how "real" or "creative" are the fiscal deficit ex ante, in the context of the are not very strong, if one assumes adjustments in many European coun­ approval of the macroeconomic plan rational and forward looking behav­ tries that are reaching the required for the year, before the legislative dis­ ior of legislators. deficit target. Several observers have cussion on the composition and allo­ Transparency of noted how Germany, France, and cation of the budget even begins. especially Italy in recent years have Finally, evidence drawn from the Budget used various ingenious methods to American States, European countries, The budgets of modern economies make their deficits appear as low as and Latin American countries all are very complex, but sometimes possible. points in the same direction: differ­ they are more complex than they In summary, this discussion sug­ ent budget procedures influence fis­ need to be. This complexity, partly gests that "hierarchical-transparent" cal outcomes. Two critical issues then unavoidable, partly artificial, makes it procedures should be associated with follow. What determines institutional possible to hide the real status of more fiscal discipline. Thus, differ­ choice? Why do different countries or public finances, in particular the cur­ ence in procedures can contribute to states choose different fiscal institu­ rent and future burden for the tax­ explaining the cross-country differ­ tions and, therefore, what determines payers of various spending decisions. ences in fiscal policy stance that are institutional change? In other words, Politicians have incentives to hide documented here. the research can be moved one step taxes, emphasize the benefits of Empirical work on this issue shows backward by looking at the determi­ spending programs, and hide govern­ the difficulty of measuring institu­ nants of institutions. The second ment liabilities, equivalent to future tions. Work by von Hagen and his issue is normative: this research can taxes, by using various forms of cre­ associates focused on European coun­ shed light on how to design institu­ ative accounting procedures. The tries and concluded that fiscal insti­ tions which contribute to maintaining more complex is a budget document, tutions do matter in the expected fiscal stability and limit the extent of the easier it is to confuse the public. direction. In my own work with co­ politically induced distortions. The importance of lack of trans­ authors, I have studied Latin Ameri­ parency cannot be overemphasized. can countries from this point of 1 A. Alesina and R. Perotti, "Tbe Political A variety of tricks are used to strate­ view.5 Using the answers from a sur­ Economy of Budget Deficits," NBER gically influence the information/ vey distributed to the budget direc- Working Paper No. 4637, February 1994.

10. NBER Reporter Winter 1997/8------1 2 A. Alesina and R. Perotti, "Fiscal Ex­ Forthcoming in a conference volume No. 5614, June 1996. pansions and Fiscal Adjustments in edited by James M. Poterba and Jurgen 5 A. Alesina, R. Hausmann, R. Hommes, aECD Countries," NEER Working Paper von Hagen and published by University and E. Stein, "Budget Institutions and No. 5214, August 1995. of Chicago Press for NEER. Fiscal Performance in Latin America," 3A. Alesina and R. Perotti, "Budget 4 A. Alesina and T. Bayoumi, "The Costs NBER Working Paper No. 5586, May Deficits and Budget Institutions" NBER and Benefits of Fiscal Rules: EVidence 1996. Working Paper No. 5556, May 1996. from u.s. States," NBER Working Paper

The Economics of Cities Edward 1. Glaeser* people, and ideas. These three dif­ amine why workers are paid sub­ ferent Sources of "agglomeration stantially higher wages in cities. The The fundamental questions of ur­ economies" have eqUivalents in other urban wage premium persists even ban economics are: Why do cities literatures. For example, the role of after we control for a full battery of exist? How does density-or agglom­ reduced transport costs for goods in individual factors (education, race, eration - affect people and firms? the formation of cities is similar to the age), job-related factors (industry and Why do some cities flourish and oth­ idea that the comovement of output occupation dummies), and for differ­ ers decay? Why are social pathologies over the business cycle occurs be­ ential selection into cities. Unless often more extreme in cities? cause a productivity shock to one firm workers in cities were more produc­ These questions address how spill­ increases demand for other firms. tive, firms would leave. Thus, even overs actually operate. If the effects Albert Ades and II measure causes though real wages seem to be con­ of agglomeration and local spillovers of urban agglomeration using cross­ stant over space (as evidence on real lie behind phenomena as important county evidence by looking at the prices suggest that they are), we as economic growth, business cycles, extent to which countries' popula­ believe that there is a productivity and the formation of human capital tions are concentrated in a single city. .premium in cities. (as many researchers now suspect), We find that population is more Surprisingly, and counter to many then urban economics has a special spread out in countries where trans­ theories of agglomeration, the urban role to play in helping us to under­ port costs for physical goods are wage premium does not immediately stand how these spillovers work in lower (measured by development of accrue to workers who come to the their rawest form. My work tries to internal transport networks), and city, and it does not disappear imme­ use the evidence from cities both to when external trade is smaller (which diately (or at all) for workers who understand urban density itself and Paul Krugman and Raul Livas2 argue leave the city. Instead, there appears to shed light on other topics that are is a further implication of transport to be a slow but steady increase in hopefully of interest to the broader cost models of urban agglomeration). the rate of wage growth for workers economics community. While transport costs do matter, our in cities relative to workers outside results suggest that political factors, cities (the urban wage premium is The Causes and Extent for example dictatorship and instabil­ also higher among older workers). ity, are far more important in explain­ One possible interpretation is that the of Agglomeration ing which countries have concen­ urban wage premium works through Economies tration in a single city. For example, faster skill accumulation in cities Cities now exist for three primary dictatorships have 50 percent more which accrues over time, and stays reasons: 1) they reduce transport of the population in their largest with workers when they leave cities. costs for goods; 2) they eliminate the cities than do stable democracies. I4 provide a theoretical analysis of space between people; and 3) cities When political systems are not stable this view and some suggestive evi­ facilitate a faster flow of ideas. More and democratic, politicians respond dence that shows that the individuals precisely, dense agglomeration re­ to the rent-seeking activities of peo­ who choose to live in big cities are duces the transport costs for goods, ple who live in their cities by trans­ drawn disproportionately from groups ferring rents to those cities, and pop­ who would presumably value the ulation flows follow these rents. skill accumulation role of cities (that * Glaeser is a Faculty Research Fellow in the NEER's Program on Economic Fluctu­ To understand why people in the is, young, college educated persons). ations and a Associate Professor at Har­ United States may be more produc­ Hedi Kallal, Jose Scheinkman, vard. He is profiled in this issue. tive in cities, David Mare and 13 ex- Andrei Shleifer, and I5 examine the

------NBER Reporter Winter 1997/8 11. --- connection between local area char­ but these improvements also increase this work and use the Longitudinal acteristics and the growth of particu­ the overall number of relationships, Research Database (which provides lar "city-industries" (for example steel and many of these new relationships almost complete information on in Detroit or retail trade in New York). also will involve face-to-face contact. every manufacturing industry in the We find that employment growth is The theoretical effect of telecommu­ United States) to examine the dy­ faster in city industries that are highly nications on cities is ambiguous. namic components of geographic con­ competitive (where competition is Empirically, we find that across coun­ centration. We find a large amount of the number of firms relative to total tries and across areas within coun­ movement of particular industries employment), not concentrated tries, urbanization and telephone usage across space (even among the most (where concentration is measured by go together, even after controlling for concentrated industries) which casts the share of the city-industry in the costs and income. Telephone contact doubt on the general applicability of total city's employment), and in is higher between regions that are anecdotes used by theorists to sug­ diverse cities (where diversity is mea­ close spatially. Business travel (an­ gest that geographic location often is sured by employment concentration other form of face-to-face contact) has determined by decades-old happen­ of the city in its largest industries). soared as information technology has stance. The longitudinal research We interpret these findings as a test improved. Silicon Valley, which should database also suggests that the geo­ of growth theories; they may imply have the best access to the newest graphic concentration of manufactur­ that diversity and competition, not telecommunications technology, is ing is a delicate balance of the cre­ industrial concentration, inculcate the textbook example of geographic ation of new manufacturing plants, growth, presumably through the gen­ concentration of industry. Still, there the expansion and contraction of eration of new ideas. doesn't seem to be persuasive evi­ pre-existing plants, and plant clo­ Scheinkman, Shleifer, and 16 ex­ dence that the internet will destroy sures. Plant openings occur away tend this work and examine the the city.8 from pre-existing industrial centers connection between city-level char­ Glenn Ellison and 19 attempt to and act to lower geographic concen­ acteristics, population growth, and measure agglomeration economies tration. Plant closures act to reinforce income growth. Population growth by developing an index of geo­ concentration, because they are much and income growth have moved to­ graphic concentration that measures more likely away from pre-existing gether at the city level over the past the degree of concentration of par­ industrial centers (even after control­ 40 years, and all of our results hold ticular industries. This index takes the ling for the fact that these peripheral for either of these variables. Initial overall concentration of manufactur­ areas are likelier to have newer plants). levels of schooling predict later growth. ing employment as given and cor­ We also find that while industries that The connection between schooling rects for the fact that the lumpiness of share input-output relationships and growth has increased since 1970, manufacturing will suggest geo­ locate together, these effects are perhaps because the rise in returns graphic concentration whenever re­ small. The dominant force in deter­ to skill has made the intellectual turns to scale dictate that production mining which manufacturing indus­ spillovers that are available in high take place in a few large plants. We tries locate together is that they hire human capital cities more important. find that the famed geographic clus­ the same kinds of workers. Cities with high unemployment levels ters of particular industries are excep­ that were concentrated in manufac­ tions rather than rules. While the Urban Pathologies turing have seen a substantial decline famous examples are all quite ob­ American cities are not just tech­ in both manufacturing and non-man­ servable in the data, the mass of nological wonderlands. Many urban ufacturing employment. Few attrib­ industries display a statistically sig­ centers face extreme poverty, crime, utes of the public bundle (that is, the nificant level of industrial concentra­ and other social problems. Funda­ level of taxation, or types of spend­ tion that appears (to us at least) to mentally, the social problems of cities ing) seem to influence the growth of be relatively mild economically. The appear to be the result of many of particular cities. fact that industrial clusters appear to the same agglomeration economies Jess Gaspar and 17 ask whether in­ be relatively rare also can be inter­ that fuel urban productivity. The formation technology will make cities preted as suggesting that many of the same intellectual spillovers that make obsolete. Improvements in telecom­ most important agglomeration econ­ Silicon Valley also may lead to the munications will shift time away from omies occur across, rather than within, transmission of ideas (or norms) face-to-face contacts towards elec­ industries. about crime within the inner city. tronic contacts within a relationship, Guy Dumais, Ellison, and 110 extend Bruce Sacerdote and III ask "Why

12. NBERReporter Winter 1997/8------is there more crime in cities?" The city-specific characteristics and look­ a final form of urban pathology: connection between city size and ing at the difference between white riots.1 5 We find a strong connection crime rates is pervasive and long­ and African-American outcomes within between urbanization and rioting standing. We find that only 15 per­ cities. Furthermore, we use a variety across countries, particularly among cent of the connection between city of instruments (including number of countries that are ethnically frag­ size and crime appears to come from governments and topographical bar­ mented. Using a sample of race riots lower probability of arrest in cities, riers) in order to avoid the problem in the 1960s, we find regular evidence and 25 percent appears related to that individuals choose their neigh­ for basic economic hypotheses about higher financial returns from criminal borhoods. All of our results essen­ rioting (higher unemployment in­ activity in cities. Almost 50 percent of tially confirm the striking negative creases rioting; police resources de­ the urban crime premium is related effect of segregation on African­ crease rioting) and little evidence for to the fact that the disadvantaged American outcomes in 1990. We sus­ more sociological hypotheses about (particularly female heads of house­ pect that the mechanism linking relative poverty and deprivation. We hold) end up in cities. racial segregation with poor out­ do find, however, that ethnic diver­ The connection between growing comes for non-whites is not physical sity is deeply linked to conflict, both up with one parent and crime is not access to jobs, but rather the intellec­ in the United States and around the nearly as strong as the connection tual and social isolation of disadvan­ world. The major role of ethnicity between living in areas with large taged communities which then works requires models that move beyond numbers of single parents and the against the acquisition of human cap­ the cost of punishment and the level of crime. While this difference ital among youth. opportunity cost of time into under­ could be explained by omitted area­ Cutler, Jakob Vigdor, and I exam­ standing the far-reaching effects of level variables that are correlated ine the roots of segregation and its ethnic identity. with single-parent families, it may current history.14 We find that segre­ Some of my work has focused on also come from spillovers within gation rose during every decade be­ policy responses to these problems. those areas. Having one parent may tween 1890 and 1960, but that it has One paper published in 1996 sug­ not induce a child to become a crim­ declined quite substantially in the gests that property taxes provide bet­ inal, but growing up in an area twenty years since 1970. Segregation ter incentives for revenue maximizing where all one's peers all have miss­ cannot be explained by many city­ local governments who would ing parents may create peer effects level characteristics except for city choose to maximize property values that lead to crime. Sacerdote, Scheink­ size: bigger cities are generally more in an attempt to maximize total rev­ man, and Il2 create a methodology segregated. We use housing prices enues.16 In a forthcoming paper, I for measuring the extent of social and survey evidence to determine argue that indexing transfer pay­ interaction based on a simple local the extent to which segregation oc­ ments to local price levels could lead interactions model. We find large curs: because of a greater willingness to pernicious results because of amounts of social interaction in crim­ among whites to pay to live in white migration effectsP In a work on my inal behavior, and we believe that neighborhoods (in which case, whites own18 and with Erzo Luttmer,19 I con­ these social interactions explain the should be paying more in more seg­ sider the effects of rent control on wide diversity of crime rates across regated cities) or; because of "collec­ misallocating housing across con­ space. tive action racism," where whites sumers. Overall, though, my primary To further examine the role of collude to force blacks to live in par­ focus has been on the issues in the local spillover effects, David Cutler ticular neighborhoods (in which case, study of urban economies, not on the and I ask whether ghettos are good we would expect African-Americans appropriate government response to or bad.13 We document that young to be paying more to live in more these economies. African-Americans growing up in segregated cities). We find that in segregated areas are more likely to 1940, segregation was driven primar­ Conclusion be unemployed or idle (that is, nei­ ily by collective action racism (sup­ Cities are unique combinations of ther at work nor in school), or to ported by legal devises including the best and worst features of mod­ have a child out of wedlock, and are restrictive covenants), but that by ern society. Agglomeration effects less likely to graduate from high school, 1990, the remaining segregation is appear to create faster learning and than young African-Americans who driven primarily by white tastes for greater productivity, but they also grow up in less segregated areas. white neighborhoods. seem to further crime, riots, segrega­ This is true even after controlling for Denise DiPasquale and I examine tion and ghetto poverty. Further

------NBER Reporter Winter 1997/8 13. l

research on cities can play a role in Section of Cities," Journal of Monetary 13 D. Cutler and E.L. Glaeser, "Are Ghettos helping us to both understand how Economics 36 (1995) pp. 117-43. Good or Bad?" NBER Working Paper No. to improve urban areas and how 7]. Gaspar and E.L. Glaeser, "Information 5163, June 1995, and Quarterly Journal of Economics CXII (1997) pp. 827-72. agglomeration economies actually Technology and the Future of Cities," 14 D. Cutler and E.L. Glaeser, "Tbe Rise operate. NBER Working Paper No. 5562, May 1996, and Journal of Urban Economics, and Decline of the American Ghetto," forthcoming. NBER Working Paper No. 5881, January 1 A.F. Ades and E.L. Glaeser, "Trade and 1997. 8 E.L. Glaeser, "Are Cities Dying," Journal Circuses: Explaining Urban Giants," of Economic Perspectives, Winter 1998, 15 D. DiPasquale and E.L. Glaeser, "Tbe Quarterly Journal of Economics 1995, pp. forthcoming. L.A. Riot and the Economics of Urban 195-228. Unrest," NBER Working Paper No. 5456, 9 G. Ellison and E.L. Glaeser, "Geographic 2 P. Krugman and R. Livas, "Trade Policy February 1996, and Journal of Urban Concentration in U.S. Manufacturing In­ Economics, forthcoming. and the Tbird World Metropolis," NBER dustries: A Dartboard Approach," Jour­ 16 E.L. Glaeser, "Tbe Incentive Effects of Working Paper No. 4238, December 1992. nal of Political Economy 105 (1997) pp. Property Taxes on Local Governments," 3 E.L. Glaeser and D.C. Mare, "Cities an.d 889-927. Skills," Hoover Institution Working Paper Public Choice 89: pp. 93-111. 10 D. Dumais, G. Ellison, and E.L. Glaeser, E-94-11, 1994. 17 E.L. Glaeser, "Should Transfer Payments "Geographic Concentration as a Dy­ be Indexed to Local Price Levels?" NBER 4 E.L. Glaeser, "Learning in Cities," NBER namic Process," NBER Working Paper Working Paper No. 5598, May 1996, and Working Paper No. 6271, November No. 6270, November 1997- Regional Science and Urban Economics, 1997. 11 E.L. Glaeser and B. Sacerdote, "W'hy is forthcoming. 5 H. E.L. Glaeser, Kallal, ]. Scheinkman, Tbere More Crime in Cities?" NBER Work­ 18 E.L. Glaeser, "Tbe Social Costs of Rent and A. Shleijer, "Growth in Cities," Jour­ ing Paper No. 5430, January 1996. Control Revisited," NBER Working Paper nal of Political Economy 100 (1992) pp. 12 E.L. Glaeser, B. Sacerdote, and]. Scheink­ No. 5441, January 1996. 1126-52. man, "Crime and Social Interactions," 19 E.L. Glaeser and E. Luttmer, "Tbe Mis­ 6 E.L. Glaeser,]. Scheinkman, and A. Quarterly Journal of Economics CXI (1996) allocation of Housing Under Rent Con­ Shleijer, "Economic Growth in a Cross- pp.508-48. trol," NBER Working Paper in progress.

14. NBER Reporter Winter 1997/8------NBER Profile: Alberto Alesina

Alberto Alesina is an NBER vard's Center for European Studies Research Associate in the Monetary since July 1994. He is a consultant to Economics Program and a professor the Italian Treasury Department, and of economics and government at Har­ has held visiting positions at the vard University. He holds an under­ International Monetary Fund and the graduate degree in economics from World Bank. the Universita Bocconi in Milan and Alesina's research has been pub­ a Ph.D. in economics from Harvard. lished in many journals and books. Alesina began his teaching career In addition, he is the coauthor. with at Carnegie-Mellon University in Nouriel Roubini and Gerald Cohen 1986. He joined the Harvard faculty of Political Cycles and the Macro­ as an assistant professor of econom­ economy, which is forthcoming from ics and government in 1988, was the MIT Press, and with Howard named the Paul Sack Associate Pro­ Rosenthal of Partisan Polities, Di­ fessor of Political Economy in 1991, vided Government and the Economy, and became a full professor in 1993. Cambridge University Press, 1995. He teaches macroeconomics and Alesina is single, and lives in political economics. Boston's Back Bay. His hobbies are Alesina was also an NBER Olin rock climbing, skiing and moun- Fellow in 1989-90. In addition, he taineering, and listening to opera. has been a senior associate at Har-

NBER Profile: Carl F. Christ

Carl F. Christ has been a member he returned to Johns Hopkins as of the NBER's Board of Directors a professor of economics. He was since 1975, and its Vice Chairman named the Abram G. Hutzler Profes­ since 1996. He also served as chair­ sor of economics in 1977, a position man of the Universities-National he held until he partially retired in Bureau Committee for Economic Re­ 1989. He also chaired the economics search from 1967-74. department in 1961-6 and again in Christ was born in Chicago in 1969-70. 1923, and received his B.S. in phys­ Christ has half a century of publi­ ics from the University of Chicago. cations in print spanning three areas: After working on the Manhattan econometrics, macroeconomics, and Project from 1943-5 and being an economic policy. He has also taught instructor in physics at Princeton and lectured in many countries. University from 1945-6, he "switched Christ has been elected a Fellow by gears" and began graduate work in the Econometric Society and by the economics. He received his Ph.D. in American Statistical Association. He economics from the University of is a member of Phi Beta Kappa and Chicago in 1950. Sigma Xi. Christ taught political economy at Christ is married and the father of the Johns Hopkins University from three. Collectively, his children have 1950 through 1955, when he became one Ph.D. (in art history) and five chil­ an associate professor of economics dren of their own. In his leisure time, at the University of Chicago. In 1961, he enjoys travel and windsurfing.

------NBER Reporter Winter 1997/8 15. 1

NBER Profile: Edward L. Glaeser

Edward Glaeser, the Paul Sack HalYard's Institute for International Associate Professor of Political Development, where he has worked Economy at HalYard University, has on urban problems in Bolivia, and been a Faculty Research Fellow in urban development in Chile. He the NBER's Program on Economic does research on cities and other Fluctuations and Growth since 1993. topics, ranging from usury laws to He received his A.B. from Princeton privatization to fairy tales. in 1988 and his Ph.D. from the Glaeser currently lives in Chicago University of Chicago in 1992. with his wife, Jenny, and commutes In 1994-5, Glaeser was the Arch into Boston during the week. During W. Shaw National Fellow at Stan­ the winter and spring terms, he will ford University's Hoover Institution. be a Law and Economics Fellow at Glaeser is also a Faculty Associate at the University of Chicago Law School.

NBER Profile: Lars E. O. Svensson

Lars E.O. Svensson is a professor of trade theory to international finance, international economics at the In­ open -economy macroeconomics, stitute for International Economic and monetary policy. Studies, Stockholm University, and a Svensson is a fellow of the Eco­ Research Associate in the NBER's nometric Society, a member of the Programs on Asset Pricing, Interna­ Prize Committee for the Alfred Nobel tional Finance and Macroeconomics, Memorial Prize in Economic Sciences, and Monetary &onomics. He received and a member of Academia Europae. a M.Sc. in physics and applied math­ He has been visiting scholar or visit­ ematics from the Royal Institute of ing professor at universities in Technology, Stockholm, in 1971. France, Israel, Italy, New Zealand, Shifting fields, he received a B.A. in and the United States. He regularly economics and economic history in consults for several international, 1973 and a Ph.D. in economics in U.S., and Swedish agencies and is 1976, both from Stockholm Univer­ active as advisor to Sveriges Riks­ sity. During his graduate studies, he bank (Bank of Sweden). benefited from a year as a Special In his leisure time, he enjoys good Graduate Student at MIT in 1974-5. food and wine with friends, travel­ His primary research interests have ling, reading (John Le Carre and varied over the years, from intertem­ David Lodge are among his favorite poral general equilibrium theory and authors), rock climbing, and spend­ monetary theory via international ing time with his teenage son, Erik.

16. NBER Reporter Winter 1997/8------Conferences

Cutler examines how reductions telephone services to be at least $2.36 family saving and on the degree of in hospital payments by Medicare billion (in addition to the $2.25 bil­ effort and planning that students put affect hospital operations. He looks lion of tax revenue). The efficiency into college. at two episodes of payment reduc­ loss to the economy for every $1 For more than three decades, tions-the late 1980s and the early raised to pay for the Internet access economists have advocated the use 1990s - and finds a large difference discounts thus is an additional $1.05 of the tax system as a means of trans­ in the impact of payment reductions to $1.25 beyond the money raised for ferring income to low-income fami­ in these two time periods. In the the Internet discounts. This cost to lies. Studying the Earned Income Tax 1980s, reduced Medicare payments the economy is extraordinarily high Credit (EITC) offers the opportunity were offset dollar-for-dollar by in­ compared to other taxes used by the to learn how well the tax system creased prices to private insurers. In Federal government to raise revenues. functions in roles traditionally han­ the 1990s, however, payment reduc­ Hoxby investigates the economic dled by the welfare system. There are tions result in lower hospital profits, effects of provisions in the Taxpayer two features of the EITC that distin­ which must ultimately reduce hospi­ Relief Act of 1997 related to higher guish it from other U.S. income trans­ tal costs. Hospitals have responded education. She summarizes the major fer programs. First, only taxpayers to the payment reductions by reduc­ initiatives: Hope Tax Credits; Tax who work are eligible for the EITC. ing the number of beds and nurses, Credits for Lifelong Learning; Edu­ Second, the credit is administered and sometimes by closing entirely, cation IRAs; and tax deductibility of through the tax system rather than but not by reduced acquisition of interest on student loans. She then through the welfare system, and is high-tech equipment. describes the incentives that these usually received as part of a taxpayer's Hausman analyzes the Congres­ provisions generate for attending col­ annual tax refund. Liebman dis­ sionallegislation which established a lege, and discusses whether the peo­ cusses these features of the EITC, and program for all U.S. public schools ple who most need to attend college presents evidence that the EITC has and libraries to receive subsidized are the ones most likely to be in­ increased labor force participation service to the Internet. The cost of duced to attend by the new initia­ among single women with children, the program is currently estimated to tives. Then she synthesizes the existing and has offset a significant share of be $2.25 billion per year, and the leg­ literature on how federal government recent increases in income inequal­ islation directed all users of interstate funds for higher education affect the ity. The Ijmited evidence available telephone service to pay for the pro­ tuition charged by colleges and uni­ suggests that the labor supply impact gram. Using analytical methods from versities, and assesses the likely con­ of the phase out of the credit is min­ public finance, Hausman estimates sequences of the new provisions for imal. Rates of noncompliance are the efficiency cost to the economy of tuition. Finally, she discusses the falling, and are now similar to the the increased taxation of interstate probable effects of the initiatives on overall rate of noncompliance for the

------NBER Reporter Winter 1997/8 17. individual income tax. ficiaries' responses to the earnings cently legislated increase in the exempt The Social Security earnings test test. Her estimates imply substantial amount. reduces a 65-69 year old's benefits deadweight loss from older workers The papers and discussions from by one third and a 62-64 year old's changing their labor supply in order this conference will be published benefits by one half, once earnings to avoid taxation. She predicts a 5.3 by the MIT Press in a conference pass a threshold amount. These are percent boost to aggregate labor sup­ volume. Its availability will be an­ among the highest marginal tax rates ply from eliminating the earnings nounced in a future issue of the in the ecbnomy. Friedberg formu­ test, and at a minimal fiscal cost. In NBER Reporter. lates a model of labor supply to in­ contrast, only a slight decrease in corporate the entire range of bene- labor supply is likely from the re-

Jolls, Sunstein, and Thaler offer scriptive analysis deals with what choices, and what is the appropriate a broad vision of how law and eco­ rules should be adopted to advance domain of paternalism? nomics analysis may be improved by agreed-upon ends; here, they offer The rich law and economics litera­ increased attention to insights about alternatives to standard law and eco­ ture on contract default rules-that actual human behavior. Their analy­ nomics prescriptions based on be­ is, terms that govern relationships sis falls into three categories: positive, havioral insights into the seemingly between contracting parties only if prescriptive, and normative. Positive pernicious or counterproductive ef­ those parties do not explicitly agree analysis of law focuses on how agents fects of the standard proposals. Fi­ to other terms - presumes that the behave in response to legal rules, nally, normative analysis attempts to legal system's choice of such rules and how legal rules are produced; assess more broadly the ends of the will not affect individual negotiators' here, they suggest that a behavioral legal system: for example, should underlying preferences for contract approach improves predictions. Pre- the system always respect. people's terms. The literature suggests that if

18. NBERReporter Winter 1997/8------bargainers perceive default terms as any ex ante understanding about lia­ effects of punitive damages in situa­ part of the status quo, they will pre­ bility that parties may have had. tions of reckless disregard that might fer the substantive content of those When no such mechanisms are avail­ be viewed as outrageous. If the de­ terms more than they would if other able, the courts choose sensibly fendant were making a rational deci­ terms were the legal defaults. Korob­ among second-best solutions. The sion that reflected all of social costs, kin presents a study designed to test legal system thus incorporates and any level of punitive damages would this hypothesis: 151 law students adapts to this limitation on human lower efficiency. If there were inade­ were asked to provide advice to a judgment. quacies of compensatory damages, client in a number of hypothetical Kahneman, Schkade, and Sun­ then costs borne by the defendant contract negotiation scenarios, with stein conducted an experimental would be less than the full social the content of the default terms ma­ study of punitive damage awards in costs. In that case, pUnitive damages nipulated among experimental groups. personal injury cases, using jury-eli­ might improve economic efficiency, The results suggest that the choice of gible respondents. There was sub­ as well as providing retribution. legal default terms affects not only stantial consensus on judgments of A major issue in the current debate what terms contracting parties will the outrageousness of. a defendant's over litigation reform is whether to agree upon but also what terms they actions and of the appropriate sever­ limit the amount of damages that actually prefer. ity of punishment. Judgments of dol­ plaintiffs can receive. In some juris­ Rachlinski notes that past events lar awards made by individuals and dictions in the United States, "dam­ frequently seem inevitable and pre­ synthetic juries were much more age caps" have been placed on pay­ dictable after they unfold, a tendency erratic, though. These results are ments for pain and suffering, punitive that cognitive psychologists have familiar characteristics of judgments damages, and recovery of medical labeled the "hindSight bias;' This bias made on scales of unbounded mag­ expenses. Babcock and Pogarsky affects judgments of liability in the nitude. The degree of harm suffered present experimental evidence on legal system. For example, if adverse by the plaintiff and the size of the the impact of these caps on the judg­ outcomes seem more predictable firm had a pronounced effect on ments and negotiating behavior of (and hence aVOidable) than they awards. subjects assigned to be litigants in a really were, then defendants can be In his first paper, Diamond devel­ pre-trial negotiation. Preliminary held liable for adverse outcomes that ops a typology of different behaviors results from this experiment indicate they could not have foreseen. In that might be viewed as outrageous that when the damage cap is rela­ effect, judgments of liability made by a jury and subjected to punitive tively low, it causes the likelihood of under a negligence standard thus damages. He then derives the level settlement to increase. This is consis­ resemble those made under a stan­ of punitive damages for achieving tent with predictions of the standard dard of strict liability. Courts, how­ economic efficiency in four different legal-economic model of settlement. ever, have historically shown an situations: malice, and three settings However, the legal-econOmic model awareness of the hindsight bias and where a jury might find reckless dis­ offers an incomplete characterization when possible, developed methods regard-a rational response to insuf­ of the settlement process by ignoring of making judgments that avoid ficient compensatory damages; a important psychological mechanisms reliance on it. These adaptations nonrational disregard of risk; and a of information processing and judg­ include barring from the decision­ rational response when compen­ ment formation. making process information acquired satory damages are adequate. In his only later, and carefully enforcing second paper, Diamond explores the

------NBER Reporter Winter 1997/8 19. Feldman and Scharfstein com­ patients tend to be treated by physi­ care organization (Mea) while all pare the quality of care received by cians who perform relatively fewer others are employed by the hospital. cancer patients in fee-for-service and surgeries, and that these patients The Mea attendings have much managed care plans. For the pur­ receive their treatment in lower-vol­ stronger financial incentives to poses of this analysis, they use ume hospitals. The differences across reduce their patients' (almost all of provider volume as a proxy for qual­ the seven managed care plans are whom are insured by the Mea) costs ity: many previous studies have substantial, with one of the plans than do the physicians employed by shown that the patients of high-vol­ actually sending its patients to higher the hospital. Meltzer finds that the ume physicians and hospitals have volume providers than the fee-for­ patients of Mea attendings have sig­ better clinical outcomes than do service plans. nificantly lower costs than do similar other patients. Their dataset contains Meltzer uses data on all admis­ patients of hospital-employed physi­ all Massachusetts hospital discharges sions to a teaching hospital's internal cians. The majority of this cost sav­ in 1995; they compare the experi­ medicine service over a year-and-a­ ing is accomplished through shorter ences of patients with breast cancer, half period to investigate the effect of lengths-of-stay. Physician workload colorectal cancer, and gynecologic the attending physician's financial also has a significant effect on patient cancer, because all three types of incentives on the costs of care. discharge probabilities, and hospitals cancer typically are treated surgically. Within the hospital, some "attend­ may increase their attending physi­ The authors find that managed care ings" are employed by the managed cians' incentives to discharge patients

20. NBERReporter Winter 1997/8------qUickly by reducing their house staff. age Medicare costs are substantially when selling a hospital. For their Philipson uses data from the U.S. higher (by approximately a two-to­ analysis, the authors obtained de­ National Nursing Home Survey to one margin) in Miami. They ascribe tailed information about hospital pur­ determine whether not-for-profit this to the differences in treatments chase prices, the use of the funds nursing homes have significantly of the chronically ill in the two areas. received, and the commitments made higher prices than similar for-profit Next, the authors conduct a cross­ by the buyers to the local communi­ facilities. In a cross-sectional analysis sectional analysis of the 306 U.S. hos­ ties. They calculate appropriate of firms in both 1985 and 1995, pital referral regions, and find that prices under different assumptions Philipson finds no support for the regions with relatively more hospital about future cash flows, and take presence of a not-for-profit premium. beds and physician specialists have into account any community benefits Instead, he finds a 5 percent for­ significantly greater end-of-life Medi­ which are not reflected in the trans­ profit premium in 1985 and no sig­ care expenditures. They also find that action price. Their results suggest that nificant difference in price in 1995. Medicare patients in areas with a communities which deal with a for­ These results suggest that there is greater penetration of for-profit hos­ profit hospital corporation receive a perfect substitution on the demand pitals have significantly higher end-of­ price substantially above the fair side between not-for-profit and for­ life spending. This greater spending price. Interestingly, the authors find profit production. does not appear to lead to improved the reverse when communities deal Gentry and Penrod estimate the outcomes, though, since mortality with not-for-profit or government magnitude of the tax benefits given rates are not significantly related to organizations. Finally, they find that to not-for-profit (NFP) hospitals. NFP intensity of care near the end of life. hospitals which convert to for-profit hospitals are exempt from corporate Frank and Salkever explore the status do, not reduce their provision income taxes (federal and state); causes and effects of the diversifica­ of community benefits. property taxes (state and local); have tion of activities by not-for-profit hos­ Cutler and Horwitz consider why access to tax-exempt bond financing; pitals. The authors cite the recent two large not-for-profit hospitals con­ and can receive charitable donations reduction in the demand for inpatient verted to for-profit status, and what which are tax-deductible for the care as an important reason for the effects these conversions have had. donor. Using data from the Health increase in diversification and joint For their analysis, the authors use Care Financing Administration's 1995 ventures by not-for-profit hospitals. several sources of information, in­ public use file of Medicare Cost Based on the results of three focus cluding interviews with hospital per­ Reports, the authors estimate that the group discussions with executives sonnel, newspaper articles, Medicare income tax exemption is worth $4.6 from 14 (mainly NFP) hospitals in cost reports, and legal documents. billion to not-for-profit hospitals, Boston and Chicago, they find that Their results suggest two primary while the value of their property tax non-teaching, NFP hospitals diversify motivations for conversions to not­ exemption is $1.6 billion. Their re­ their activities and enter into joint for-profit status: to gain access to sults suggest, however, that the net ventures not only to offset reductions cheaper sources of capital; and the benefit of access to tax-exempt bonds in inpatient revenues but also to gain culture of the NFP hospital, since is quite small and does not signifi­ market share, strengthen ties with phy­ hospitals run by businessmen may be cantly reduce the cost of borrowing sicians, and reduce- the uncertainty in much more likely than religious-affil­ for NFP hospitals. The authors do demand. Philanthropy constitutes a iated or physician-run hospitals to point out that, if NFP hospitals en­ very small share (approximately 1 convert. The conversions appear to gage in tax arbitrage by borrowing at percent) of a typical not-for-profit have improved the financial per­ tax-exempt interest rates and invest­ hospital's budget, and these private formance of these hospitals by cut­ ing in financial assets with greater donations fall when a hospital's fi­ ting hospital costs and increasing returns, then the magnitude of this nancial performance improves. Finally, public sector reimbursement. The benefit could be substantial. they find no evidence that diversifi­ authors suggest that this second fac­ Skinner and Wennberg compare cation or a decline in private dona­ tor is attributable to the more skillful Medicare expenditures and physician tions has reduced the provision of exploitation of Medicare loopholes visits in the last six months of life charity care by NFP hospitals. by for-profit hospitals. across U.S. communities to investi­ Analyzing ten hospital conversions McClellan and Staiger compare gate the productive and allocative which occurred in North and South quality at for-profit and not-for-profit efficiency of end-of-life medical care. Carolina since 1981, Sloan, Taylor, hospitals using a dataset which con­ Initially, the authors focus on Miami and Conover investigate whether tains all elderly Medicare beneficia­ and Minneapolis, and find that aver- communities receive a "fair" price ries who are hospitalized from 1984

------NBER Reporter Winter 1997/8 21. through 1994 following their first every ambulance ride in Pennsyl­ Finally, using a national-level dataset heart attack, or hospitalized with vania during 1995 to explore the on the adoption of 911 technologies ischemic heart disease from 1984 direct productivity benefits of a com­ across communities, they find that through 1991. They find a strong munity's adopting a basic or ad­ places with a more conservative po­ negative relationship between hospi­ vanced 911 system. Focusing on cases litical orientation are less likely to tal volume and mortality rates. Also, of cardiac incident, they find that adopt advanced 911 systems, and not-for-profit hospitals have lower both the time to reach an emergency that state legislation governing the mortality than both for-profit and site and the time elapsed from the adoption of 911 has an important government hospitals. Further, differ­ site to the hospital is decreasing with effect on communities' adoption ences in mortality rates between not­ the adoption of advanced 911 serv­ decisions. for-profit and for-profit hospitals ices. However, there is little evidence This article was prepared in large increased between 1985 and 1994. to suggest that mortality rates from part by Mark Duggan of Harvard The authors conclude that, within a cardiac incidents are related to the University, who also attended the specific market, for-profit hospitals adoption of 911 services. The authors conference. It is expected that these actually have higher quality than do also find that a hospital's level of car­ papers and the conference discus­ not-for-profit hospitals. diac technology has an important sions will be published in a confer­ Athey and Stern explore the causes impact on its share of cardiac patients ence volume by the University of and effects of differences across com­ within a market, but there is little evi­ Chicago Press. -Its availability will be munities in emergency services. Ini­ dence that the 911 system influences announced in a future issue of the tially, the authors use a dataset with hospitals' technology investments. NBER Reporter.

Fleming and Remolona note that suggesting that price reactions to persists with high price volatility and the arrival of public information in public information do not require a slightly higher bid-ask spread, sug_· the U.s. Treasury securities market trading. The bid-ask spread widens gesting a sluggish process of price induces striking adjustment patterns dramatically at announcement, and formation as the market reconciles for prices, trading volume, and bid­ narrows shortly thereafter, apparently investors differential private views. ask spreads. The release of a major driven by inventory control rather Gehrig and Jackson examine the macroeconomic announcement occa­ than asymmetric information con­ prices quoted by specialists (or deal­ sions a sharp and immediate price cerns. After the initial sharp price ers) who have monopoly power to change with little trading volume, change, trading volume surges and set prices (bids and asks) for a given

22. NBER Reporter Winter 1997/8------asset, but who face indirect competi­ that are relevant in international authors: assess the quality of the mar­ tion from other specialists who trade macroeconomics. ket through the various stages of in related assets. They compare the Traders in security markets take reform; examine the effect of micro­ equilibrium spreads as the number into account the actions of their peers structure reforms on the emergence (and factor structure) of the assets in making their own trade decisions. of the market; and consider the price that each specialist controls varies. In this paper, Chakrabarti and Roll impact of large block trades. They For some constellations of initial compare a market in which traders also test whether the costs of trading portfolio holdings and asset covari­ "learn" from one another with a mar­ have changed Significantly since the ance, it is SOcially preferable to have ket in which they ignore each other's stock market reforms. Their results competing specialists, while for oth­ actions. The authors measure volatil­ show that the effective spreads and ers it is SOcially preferable to have ity, volume, and the accuracy of mar­ costs of trading have, if anything, actions coordinated (or to have one ket price as a forecast of value in the increased except for the most actively specialist control several assets). In "learning" and the "no learning" cases. traded stock. some situations it is beneficial to While bubbles and cascades do arise Neal and Wheatley examine the have specialist power concentrated some times, on average "learning" performance of two commonly used within industries; in other situations, reduces price and return volatility empirical models for estimating the across industries; and in other situa­ and volume, and improves the accu­ adverse selection component of a tions, not to be concentrated at all. racy of the market price as an indi­ firm's bid-ask spread. They use the Evans studies the high frequency cator of value. The authors examine models to estimate the adverse selec­ beha vior of the interbank foreign the marginal effects of different para­ tion components of a sample of exchange market with a newly cre­ meters on these market characteris­ closed-end funds and a matched ated dataset that provides a compre­ tics, and find that the "learning" sample of common stocks. In con­ hensive picture of activity across the process has a complex and nonlinear trast to the stocks, closed-end funds market. His analysis indicates that impact. report their net asset values weekly, trade activity within the interbank To enhance their understanding of all but eliminating uncertainty about market is distinct from the posting of emerging markets, Ghysels and their current liqUidation values. Thus indicative quotes. Trading and quote­ Cherkaoui study an unusual dataset the authors expect the adverse selec­ making decisions are linked, but the containing all the transaction records tion component to be much smaller links are complicated and poorly of a market over a long span. The for the funds than for the stocks. understood. He also documents the market, which was included in 1996 Estimates of the component, how­ existence of a strong relationship in the International Finance Corpo­ ever, are large and significant for between exchange rate movements ration database, operated under a both samples. This suggests that and a measure of excess dollar de­ dual trading system, consisting of an either adverse selection arises pri­ mand. Empirically, this effect appears upstairs market for large block trades marily from factors other than current important in the determination of and a trading floor exchange. Trans­ liquidation value or that the empirical exchange rates at both high frequen­ actions were recorded separately for models are misspecified. cies and over the longer time spans both segments of the market. The

------NBER Reporter Winter 1997/8 23. I 1.

Arora, Fosfuri, and Gambardella expected investment in LDCs by $100 ogy industries often assumes that study how an independent, upstream million to $ 200 million, with the firrns' abilities to survive depend on capital good sector in a technology increases greater in more mature their own internal Rand D efforts. based industry can act as a mecha­ technologies, and for larger LDCS. Blonigen and Taylor argue that nism for the transmission of growth Lerner and Merges examine the high-technology firrns may choose to across countries. Since the number of determinants of control rights in tech­ specialize either in this internal specialists is determined by the size nology alliances involving biotech­ growth (through R and D) strategy or of the downstream sector, the growth nology firms. They undertake three in an external growth strategy of of the downstream sector in leading case studies and a quantitative analy­ acquiring other firms or finns opera­ countries has beneficial effects for sis of 200 alliances. Consistent with tions. Using a panel of 214 U.S. elec­ the growth of the downstream sector the framework they use, the alloca­ tronics finns over nine years to test in follower countries (less developed tion of control rights to the firm per­ the relationship between Rand D countries, or LDCs). Using a compre­ forming Rand D increases with its intensity and the probability of acqui­ hensive dataset of investments in financial resources. The empirical sition, and controlling for traditional chemical plants in the developing evidence is much more ambiguous determinants of acquisition activity countries during the 1980s, the on the relationship between the stage which include financial constraints, authors find that one additional spe­ of the project at the time the alliance they find a strong and significant cialized supplier in a given process is signed and control rights. negative correlation between Rand D technology would have increased the Economic analysis of high-technol- intensity and the probability of acqui-

24. NBERReporter Winter 1997/8------

...I. sition. This suggests that electronics tivity specifications and measures of work externalities and installed bases firms maybe specializing in one work systems. These findings sup­ contributed to de facto standardiza­ activity or the other. Their results also port the idea that organizational prac­ tion of the VHS format in the U.S. suggest that firms with greater intan­ tices are important determinants of IT home VCR market. His results imply gible assets, higher profitability, and demand and productivity. that: 1) the network advantage of lower debt-to-asset ratios are more Using a high frequency dataset of VHS became increasingly important likely to acquire. advertised prices in the personal during 1981-3, and although there Miravete and Pernias study the computer industry, Deltas and was a push of Betamax in 1983, the existence of strategic complemen­ Zacharias find, after adjusting for all network advantage was the key rea­ tarity between product and process observable product characteristics, son that VHS has outsold Betamax innovation for numerous firms in a that firms which were late in intro­ since 1986; 2) the expected sales ad­ wide range of industries. They show ducing Pentium-based computers vantage was more important than the that general innovative strategies of command a higher price premium as installed base advantage in the adop­ Spanish firms can be considered compared to early entrants. This is tion of the VHS format, however, the mutually complementary and robust true even among firms which have installed base advantage became in­ to the significant existence of firms the same price premium for their creasingly important as time passed; unobservable heterogeneity. Their 486-processor-based computers. Over and 3) the increase in the network empirical evidence is consistent with time, the difference in the Pentium advantage ofVHS, especially through small firms having a comparative price premiums of the late versus the the accumulation of a larger installed advantage in succesfully implement­ early entrants declines to the level of base, was an engine of tipping to­ ing flexible manufacturing methods. the difference in the corresponding ward and de facto standardization of BrynJolfsson and Hitt examine 486 price premiums. One explana­ the VHS format. the influence of organizational design tion for these results might be that Thomas empirically examines the on the demand for information tech­ the late entrants reap short-run rents order in which firms adopt new tech­ nology (IT) and the productivity of from consumers who are loyal enough nologies in the computer disk drive IT investments using data from ap­ to have waited for their entry into the industry. He finds that large firms and proximately 380 U.S. firms. They find market in order to purchase. This incumbents are more likely to adopt greater demand for IT in firms with suggests that firm identity matters earlier than small ones· and new greater decentralization of decision over and above any other product entrants when innovation does not rights (especially the use of self-man­ characteristics in determining con­ rapidly make obsolete existing tech­ aging teams), and greater invest­ sumer willingness to pay for the nology and products. He also finds, ments in human capital, including product. in some cases, that small firms and training and screening by education. Park develops a model of con­ new entrants adopt earlier than large In addition, the output elasticity of IT sumers' choices among incompatible ones and incumbents when innova­ is higher in firms that adopt a more technologies and producers' pricing tion rapidly makes obsolete existing decentralized and human capital­ in the presence of network external­ products. These results are consistent intensive work system. This relation­ ities. He then applies the model to with the economics and strategy lit­ ship is robust to alternative produc- analyzing the extent to which net- eratures on innovation.

------NBER Reporter Winter 1997/8 25.

I L Bureau News

Merton Shares '97 Nobel Robert C. Merton, a Research Asso­ mathematical formula aimed at mea­ Fogel, both NBER Research Associ­ ciate in the NBER's Program on Asset suring the worth of an option. "Thou­ ates and professors of economics at Pricing and professor of economics sands of traders and investors now the University of Chicago, won the at Harvard Business School, and use this formula every day to value prize in 1995 and 1993, respectively. Myron S. Scholes, a professor emeri­ stO'ck options in markets throughout Other researchers who have been tus at Stanford Business School and a the world," the Academy said in its affiliated with the NBER over the former NBER Research Associate, announcement. Options are contracts years and won the Nobel Prize in won the Nobel Prize in Economics that allow, but do not require, an Economics are Simon S. Kuznets, tnis year. investor to buy an asset, like a stock Milton Friedman, Theodore W. The Royal Swedish Academy of or oil, at a fixed price during a spec­ Schultz, George J. Stigler, and Gary S. Sciences awarded the prize to Merton ified period of time. Becker. and Scholes for helping to devise a Robert E. Lucas, Jr. and Robert W.

NBER Researcher is Newest Member of CEA President Clinton recently named Northwestern University, where she Program on Children. NBER Research Associate Rebecca M. teaches economics and directs the Blank holds a doctorate in eco­ Blank to be a member of his Council Northwestern/University of Chicago nomics from MIT, and has written of Economic Advisers. Her nomina­ Joint Center for Poverty Research. extensively about government anti­ tion is subject to Senate confirmation. She is also a member of the NBER's poverty programs. Professor Blank is on leave from Program in Labor Studies and the

26. NBER Reporter Winter 1997/8------Diebold, Ohanian, and Berko­ Adda and Cooper study the ef­ data on job creation and destruction. witz propose a framework for assess­ fects of subsidies on durable goods Cyclical fluctuations in the job de­ ing whether dynamic equilibrium markets. In particular, they consider a struction rate serve to magnify the models and data agree. They evalu­ recent policy in France in which the effects of productivity shocks on out­ ate the significance of deviations governments of Balladur and Juppe put, as well as making the effects between the models and the data, subsidized the replacement of old much more persistent. Interactions and use goodness-of-fit criteria to cars with new ones. They find that between household savings decisions produce estimators that optimize these policies do stimulate the auto­ and separation decisions in employ­ economically-relevant loss functions. mobile sector in the short run but, ment relationships playa key role in They provide a detailed illustrative through the induced changes in the propagating shocks. application to modeling the u.s. cat­ cross-sectional distribution of car The most important event in a tle cycle. ages, they create the basis for subse­ recession is a sharp drop in employ­ In an economy where skilled and quent low activity. Further, while ment. At the same time, firms liqui­ unskilled workers use different these policies increase government date inventories; employment and technologies, the rate of improve­ revenues in the short run, revenues production fall by more than sales. ment of each technology is deter­ in the long run are lower relative to Hall interprets recessions in the fol­ mined by a profit-maximizing Rand D a baseline without intervention. lowing way: episodes of financial sector. When there is a high propor­ A large recent literature shows that stress result in abnormally high dis­ tion of skilled workers in the labor strategic interactions among actors count rates for business decisions. At force, the market for skill-comple­ with conflicting objectives can pro­ the same time, possibly for unrelated mentary technologies is larger, and duce inefficient political decisions. reasons, profitability falls in some more effort will be spent in upgrad­ Romer investigates an alternative important industries. An increase in ing the productivity of skilled work­ explanation of such decisions: if indi­ the discount rate shifts the decision ers. One implication of this theory is viduals' errors in assessing the likely toward shutdown because the firm that when the relative supply of effects of proposed policies are corre­ realizes immediate cash from liqui­ skillc:d workers increases, the skill lated, then democratic decisionmak­ dating inventories and other capital premium decreases in the short run, ing can produce inefficient outcomes upon shutdown, whereas the fore­ but then increases, possibly even even in the absence of distributional gone profit lies well into the future. A above its initial value, because the conflicts or heterogeneous prefer­ decline in profit also shifts the deci­ larger market for skill-complementary ences. Choosing candidates from sion toward shutdown if it lowers technologies has changed the di­ among the best informed members profit in relation to liqUidation value. rection of technical change. This sug­ of the population does not remedy The burst of job destruction and gests that the rapid increase in the the problems created by such errors, inventory decumulation that occurs proportion of college graduates in but subsidizing information and ex­ during the sharp part of the typical the u.s. labor force may have been posing representatives to information contraction results from profit-maxi­ causal in both the decline in the col­ after their election do. Concentration mizing decisions to shut down units lege premium during the 1970s and of power has ambiguous effects. in firms throughout the economy. the large increase in inequality during Finally, the presence of correlated The concentration of recessions dur­ the 1980s. Acemoglu also derives errors tends to create multiple equi­ ing brief episodes is increased as a implications of directed technical libriums in political institutions. result of the heterogeneity of pro­ change for residual wage inequality, Den Haan, Ramey, and Watson ductive units. If an adverse shock hits and shows that calculations of the develop and quantitatively imple­ at a time when there has been a impact of international trade on in­ ment a dynamic general equilibrium buildup of units near the borderline equality that ignore the change in the model with labor market matching of shutdown, the burst of job destruc­ direction of technical progress may and endogenous job destruction. The tion and inventory liquidation will be be misleading. model produces a close match with more severe.

------NBER Reporter Winter 1997/8 27. - Faculty Research Fellows for 1997-8

::;! Yacine Ait-Sahalia Patricia DeVries Ann Hamson Arik Levinson John P. Rust Rosanne Altshuler Francis X. Diebold Judith Hellerstein Steven D. Levitt Xavier Sala-i-Martin Patricia M. Anderson Kathryn M.E. Dominguez Igal Hendel Jeffrey B. Liebman Andrew Samwick Patrick K. Asea Steven N. Durlauf Caroline Minter Hoxby Florencio Lopez-c!e-Silanes tJohn Carl Scholz Susan Athey Janice C. Eberly Hilary W. Hoynes Robin L. Lumsdaine John Shea Orazio Attanasio Aaron Edlin Thomas Hubbard Brigitte Madrian Nachum Sicherman Laurence C. Baker Matthew Eichner Jennifer Hunt Rodolfo Manuelli Hilary Sigman Susanto Basu Nada Eissa Guido Imbens Mark B. McClellan Matthew]. Slaughter Paul Beaudry Eduardo M. Engel Harumi Ito Kathleen M. McGarry Jeffrey Smith Geert Bekaert Jonathan Feinstein Urban Jermann David Meltzer Kathryn E. Spier Dan Ben-David Christopher Foote Christine M. Jolls Rebecca Menes Douglas o. Staiger Roland Benabou Leora Friedberg Charles I. Jones Carolyn Moehling Scott Stem Eli Berman Rachel M. Friedberg Larry E. Jones Fiona Scott Morton Ann Huff Stevens Andrew Bernard Jordi Gali Chinhui Juhn Casey Mulligan Deborah L. Swenson Giuseppe Bertola David Genesove Thomas]. Kane Wallace Mullin Scott Taylor Timothy]. Besley William M. Gentry Wolfgang Keller Derek Neal Alan M. Taylor Gordon Bodnar Alec I. Gershberg Daniel Kessler Thomas J. Nechyba Linda Tesar tLaei Brainard Robert Gertner Sukkoo Kim Aviv Nevo Aaron Tornell Lee G. Branstetter Simon Gilchrist Michael W. Klein Steven Olley Daniel Trefler Moshe Buchinsky Donna B. Gilleskie Peter]. Klenow Rosalie Liccardo Pacula Peter Tufano Jose M. Campa Edward L. Glaeser Michael M. Knetter Martin Pesendorfer Dirnitrios Vayanos Jeffrey R. Campbell Sherry A. Glied Samuel S. Kortum Paolo A. Pesenti Andres Velasco Andrew Caplin William N. Goetzmann Michael Kremer Tomas Philipson Joel Waldfogel Christopher D. Carroll tLinda Goldberg Saul Lach Anne Piehl James R. Walker Kenneth Chay Pinelopi K. Goldberg Francine Lafontaine Jorn-Steffen Pischke Shang-Jin Wei Judith A. Chevalier Paul Gompers David Laibson Canice Prendergast David Weinstein Menzie D. Chinn Austan Goolsbee Robert]. laLonde Thomas J. Prusa Ingrid M. Werner Stephen Coate Jeffrey Grogger Owen Lamont Daniel Raff Matthew White Dora Costa tJonathan Gruber Jenny Lanjouw Sergio Rebelo Holger C. Wolf Kent D. Daniel Brian Hall Thomas Lemieux Paul Rhode Catherine Wolfram Sanjiv Ranjan Das Maria]. Hanratry Josh Lerner Matthew P. Richardson Aaron Yelowitz Donald R. Davis Gordon H. Hanson Margaret Levenstein Nouriel Roubini Alwyn Young Sandra Decker tJames Harrigan Phillip B. Levine Cecilia E. Rouse Luigi Zingales Wouter]. Den Haan

tOn leave for government service

28. NBER Reporter Winter 1997/8------Research Associates for 1997-8

Andrew B. Abel Russell Cooper Alan L. Gustman AndrewW. Lo David S. Salkever John M. Abowd Hope Corman Michael R. Haines Deborah J. Lucas Garth Saloner James D. Adams Robert E. Cumby Bronwyn H. Hall Robert E. Lucas Jr. Thomas J. Sargent Joshua Aizenman Janet Currie Robert E. Hall Lisa M. Lynch David S. Scharfstein Alberto F. Alesina David M. Cutler John C. Haltiwanger Richard K. Lyons Richard Schmalensee Steven G. Allen Michael R. Darby Daniel S. Hamermesh Jeffrey K. MacKie-Mason 'Myron S. Scholes Lee J. Alston Lance E. Davis Lars P. Hansen A. Craig MacKinlay Anna J. Schwartz Joseph G. Altonji Steven J. Davis Eric A. Hanushek Thomas E. MaCurdy G. William Schwert James E. Anderson Angus S. Deaton Jeffrey E. Harris Jacques Mairesse Suzanne Scotchmer Albert Ando J. Bradford De Long Oliver D. Hart N. Gregory Mankiw Nevin S. Scrimshaw Joshua Angrist Jorge Braga de Macedo Campbell R. Harvey Charles F. Manski Douglas Shackelford Richard J. Arnott Peter A. Diamond Jerry A. Hausman Robert A. Margo Jay A. Shanken Orley C. Ashenfelter W. Erwin Diewert Fumio Hayashi James R. Markusen Matthew D. Shapiro Paul Asquith John DiNardo Geoffrey Heal Richard C. Marston Steven Shavell Jeremy Atack JohnJ. Donohue John C. Heaton Bennett T. McCallum Kathryn L. Shaw Andrew Atkeson Michael P. Dooley James J. Heckman Daniel L. McFadden Andrea Shepard Alan J. Auerbach Rudi Dornbusch John F. Helliwell Charles E. McLure Jr. Robert J. Shiller David Backus Allan Drazen Elhanan Helpman James L. Medoff Andrei Shleifer Richard Baldwin Bernard Dumas Patric H. Hendershott Robert C. Merton John B. Shoven Robert E. Baldwin Jonathan Eaton J. Vernon Henderson Gilbert E. Metcalf Christopher A. Sims Laurence M. Ball Sebastian Edwards Rebecca Henderson Bruce D. Meyer Jody L. Sindelar Michael J. Barclay Ronald G. Ehrenberg James R. Hines Jr. Jacob Mincer Burton H. Singer Robert J. Barro Martin S. Eichenbaum Robert J. Hodrick Jeffrey A. Miron Kenneth J. Singleton Robert B. Barsky Barry Eichengreen Bengt R. Holmstrom Frederic S. Mishkin Hans-Werner Sinn Ann P. Bartel Glenn Ellison Douglas Holtz-Eakin Olivia S. Mitchell Jonathan S. Skinner David S. Bates David T. Ellwood Charles Y. Horioka Naci H. Mocan Joel B. Slemrod Marianne Baxter Charles M. Engel V. Joseph Hotz Robert A. Moffitt Frank A. Sloan Lucian A. Bebchuk Stanley L. Engerman R. Glenn Hubbard Michael J. Moore Kenneth L. Sokoloff Ben S. Bernanke Robert F. Engle, III Charles R. Hulten tEdward Montgomery Barbara J. Spencer Ernst R. Berndt William N. Evans Michael D. Hurd John Mullahy Robert W. Staiger B. Douglas Bernheim Henry S. Farber Casey Ichniowski Richard Murnane Robert F. Stambaugh Jeffrey I. Bernstein Daniel R. Feenberg Robert P. Inman Kevin M. Murphy Richard H. Steckel Steven T. Berry Robert C. Feenstra Douglas A. Irwin 'Richard Mussa Jeremy C. Stein Mark Bils Martin S. Feldstein Takatoshi Ito Stewart C. Myers 'Joseph E. Stiglitz Olivier J. Blanchard Raquel Fernandez Adam B. Jaffe M. Ishaq Nadiri James H. Stock David G. Blanchflower Joseph P. Ferrie Michael C. Jensen David Neumark Alan C. Stockman tRebecca Blank Wayne E. Ferson Paul L. Joskow William D. Nordhaus Rene M. Stulz Francine D. Blau 'Stanley Fischer Boyan Jovanovic Maurice Obstfeld tLawrence H. Summers Alan S. Blinder Price V. Fishback Theodore J. Joyce Ariel Pakes Richard C. Sutch Magnus Blomsttom Marjorie Flavin Kenneth L. Judd Christina Paxson Lars E.O. Svensson David E. Bloom Roderick Floud Robert Kaestner Torsten Persson Richard Sylla Tim Bollerslev Robert W. Fogel Edward J. Kane Charles E. Phelps John B. Taylor Michael D. Bordo Richard Frank Shawn E. Kantor Robert S. Pindyck Peter Temin Severin Borenstein tJeffrey A. Frankel Steven N. Kaplan A. Mitchell Polinsky Richard H. Thaler George J. Borjas Richard B. Freeman Louis Kaplow Clayne L. Pope Marie C. Thursby Axel H. B6rsch-Supan Kenneth R. French Anil K Kashyap Robert H. Porter Sheridan Titman Michael J. Boskin 'Jacob A. Frenkel 'Lawrence Katz Richard Portes Robert H. Topel John Bound Benjamin M. Friedman tPatrick Kehoe James M. Poterba Manuel Trajtenberg David F. Bradford Kenneth A. Froot John Kennan Edward C. Prescott James Trussell James A. Brander Victor R. Fuchs Miles S. Kimball Andrew D. Racine Steven F. Venti William H. Branson Don Fullerton 'Mervyn A. King Raghuram Rajan W. Kip Viscusi Timothy F. Bresnahan Melvyn A. Fuss Robert King Valerie A. Ramey Robert W. Vishny Charles C. Brown David Galenson Morris M. Kleiner James E. Rauch JohnJ. Wallis Warren Browner Robert E. Gallman tSanders Korenman Assaf Razin Jiang Wang Willem H. Buiter Alan M. Garber Laurence J. Kotlikoff Peter C. Reiss Mark W. Watson Jeremy I. Bulow Martin Gaynor Kala Krishna J. David Richardson Philippe Weil Ricardo J. Caballero Mark Gertler Alan B. Krueger Mark J. Roberts David N. Weil Charles Calomiris Paul J. Gertler Anne o. Krueger Hugh Rockoff Thomas J. Weiss John Y. Campbell Francesco Giavazzi Paul Krugman Dani Rodrik Kenneth D. West David Card Robert S. Gibbons Douglas L. Kruse Richard Rogerson John Whalley Dennis W. Carlton 'Claudia Goldin Josef Lakonishok Kenneth S. Rogoff Michael D. Whinston Anne Case Fred Goldman Naomi R. Lamoreaux Christina D. Romer Eugene N. White 'Ales sandra Casella Robert J. Gordon Robert Z. Lawrence David H. Romer Jeffrey G. Williamson tStephen G. Cecchetti Roger H. Gordon Edward P. Lazear Paul M. Romer Larry T. Wimmer Frank J. Chaloupka IV Gary B. Gorton John v. Leahy Andrew K. Rose David A. Wise Gary Chamberlain Lawrence H. Goulder Edward E. Leamer Nancy L. Rose Ann Dryden Witte Dov Chernichovsky Wayne Gray Blake D. leBaron Harvey S. Rosen Frank A. Wolak Lawrence Christiano Jerry R. Green John B. Legler Sherwin Rosen Barbara L. Wolfe Richard H. Clarida Shane Greenstein Richard M. Levich Joshua Rosenbloom Mark A. Wolfson Charles T. Clotfelter Zvi Griliches James A. Levinsohn julio J. Rotemberg Michael Woodford John H. Cochrane Reuben Gronau Karen K. Lewis Michael Rothschild Randall Wright lain M. Cockburn Gene M. Grossman Eugene Lewit Richard S. Ruback Victor Zarnowitz Susan M. Collins Herschel I. Grossman Gary D. Libecap Christopher Ruhm Richard J. Zeckhauser George M. Constantinides Michael Grossman Frank R. Lichtenberg Jeffrey D. Sachs Stephen P. Zeldes Philip J. Cook Sanford J. Grossman Robert E. Lipsey Henry Saffer Stanley E. Zin Lynne Zucker 'On leave for non-government service tOn leave for government service

------NBER Reporter Winter 1997/8 29. Wage Inflation and Unemployment On November 4, the NBER held a structs its compensation statistics, and Harvard; members of the NBER's Pro­ meeting on "Wage Inflation and Un­ what the numbers show. gram on Monetary Economics Steven employment" in its Cambridge office In the afternoon discussion of "The G. Cecchetti (currently on leave from which was unusual in that it brought Unemployment-Inflation Relation," the NBER at the Federal Reserve Bank together NBER researchers with gov­ NBER Research Associate Robert ]. of New York), former Program Direc­ ernment officials to discuss this im- . Gordon, a member of the Program on tor Benjamin M. Friedman and cur­ portant topic. In the morning session Economic Fluctuations and Growth, rent Program Director N. Gregory on "Measurement Issues and Market talked about the "Sectoral Origins of Mankiw, both of Harvard; Richard B. Evidence:' NBER Research Associates the Recent Decline in the NAIRU." Freeman, Director of the NBER's Pro­ Lawrence F. Katz and Alan B. Krueger, His remarks were followed by a pre­ gram on Labor Studies; as well as both members of the Labor Studies sentation by Roger Brinner of DRI/ Steven Braun of the President's Coun­ Program and former Chief Econo­ McGraw Hill, an economic consulting cil of Economic Advisers; David Stock­ mists at the U.S. Department of Labor, firm, and by a more general discus­ ton and William Wascher of the Fed­ _presented their thoughts and ques­ sion that included: NBER macro­ eral Reserve Board of Governors; tions on wage statistics. Then John economists Olivier ]. Blanchard of Daniel Sullivan of the Federal Re­ Ruser of the Bureau of Labor Statis­ MIT, Bureau President Martin Feld­ serve Bank of Chicago; and Francis tics explained how that agency con- stein and James H. Stock, both of Harris of the Bureau of Labor Statistics.

Alesina, Baqir, and Easterly pre­ U.S. cities, U.S. metropolitan areas, related to the city's (metro area's/ sent a model that links differences in and U.S. urban counties. They find county's) ethnic fragmentation. Eth­ preferences across ethnic groups in a that productive public goods-edu­ nic fragmentation is related nega­ city to the amount and type of pub­ cation, roads, libraries, sewers, and tively to the share of local spending lic good the city supplies. They test trash pickup-in U.S. cities (metro on welfare. The results are driven the model with three related datasets: areas/urban counties) are inversely mainly by observations in which

30. NBER Reporter Winter 1997/8------majority whites are reacting to vary­ show that a tax system or spending and future low lifetime earners, but ing sizes of minority groups. The program which is not optimal from helps everyone else. It also raises authors conclude that ethnic conflict one point of view still can improve long-run output by over 5 percent. is an important determinant of local taxpayer welfare, because the system The X tax makes everyone better off public finances. creates additional political pressure in the long run and also raises long­ Goolsbee reexamines the respon­ for suppressing the growth of gov­ run output by 7.5 percent. But it ..I siveness of taxable income to changes ernment. Relevant examples include harms initial older generations who in marginal tax rates using detailed the use of inflation taxes, capital taxes, bear its implicit wealth tax. compensation data on several thou­ excise taxes, deficit financing, and The populations in all industrial­ sand corporate executives from 1991 income taxes with many "loopholes." ized countries are aging rapidly and to 1995. The data confirm that the They show that in a broad sample of individual life expectancies are in­ higher marginal rates of 1993 led to a countries for 1972-90, "more effi­ creasing. Yet older workers are leav­ significant decline in taxable income. cient" tax systems - that is, systems ing the labor force at younger and Indeed, this small group of execu­ that rely on broad-based taxes with younger ages. In some countries, the tives can account for as much as 20 fairly flat rate structures-are associ­ labor force participation rates of percent of the aggregate change in ated with larger governments. An 60-to-64-year-old men have fallen wage and salary income for the one analysis of defense spending-espe­ by 75 percent over the past three million richest taxpayers, and one cially wartime spending - and oil decades. This decline in labor force person alone can account for more shocks suggests that the cause and participation magnifies population than 2 percent. The decline, how­ effect is not from spending to tax trends, further increasing the number ever, is almost entirely a short-run structures. of retirees relative to the number of shift in the timing of compensation Auerbach and his coauthors com­ persons who are working. Together rather than a permanent reduction in pare the equity, efficiency, and macro­ these trends have put enormous taxable income. The short-run elas­ economic effects of five alternatives pressure on the financial solvency of ticity of taxable income with respect to the current u.s. federal income social security systems around the to the net-of-tax share exceeds one, tax: a proportional income tax; apro­ world. Ironically, the provisions of but the elasticity after one year is at portional consumption tax; a flat tax; the social security systems them­ most 0.4, and is probably closer to a flat tax with transition relief; and a selves typically contribute to the zero. Disaggregating the data by progressive variant of the flat tax labor force withdrawal. source of income shows that the shift called the "X tax;' Their model pre­ Gruber and Wise summarize the comes from a large increase in the dicts major macroeconomic gains evidence presented in a number of exercising of options in the year be­ (including an 11 percent increase in papers on eleven industrialized coun­ fore the tax change, followed by a long-run output) from replacing the tries, and distill the key conclusions decline in exercising in the year of federal tax system with a propor­ from the collective findings. They the tax change, and is concentrated tional consumption tax. Future mid­ note a strong correspondence be­ among executives at the top of the dle- and upper-income classes gain tween the age at which benefits are income distribution. Executives with­ from this policy, but initial older gen­ available and departure from the out stock options show six times less erations are hurt by the policy's labor force. In addition, the provi­ responsiveness to taxation. Other implicit capital levy. Poor members of sions of social security programs types of compensation, such as salary current and future generations also often imply large financial penalties and bonus or nontaxed income, are lose. The flat tax, which adds a stan­ on labor earnings beyond the social either not responsive to tax rates or dard deduction to the consumption security early retirement age. Further­ not large enough to make a differ­ tax, makes all members of future more, in many countries disability ence. the estimated elasticities indi­ generations better off, but at a cost and unemployment programs effec­ cate that the dead weight loss of of halving the economy's long-run tively provide early retirement bene­ recent tax increases was at most one output gain and harming initial older fits before the official social security quarter of the revenue generated. . generations. Insulating these older early retirement age. Gruber and Becker and Mulligan provide a generations through transition relief Wise conclude that provisions of model for analyzing the effects of the further reduces the long-run gains social security programs indeed have tax system and spending programs from tax reform. Switching to a pro­ contributed to the decline in the on the determination of government portional income tax without deduc­ labor force participation of older per­ spending and taxpayer welfare. They tions and exemptions hurts current sons, substantially reducing the

------NBER Reporter Winter 1997/8 31. potential productive capacity of the als in all previous generations who the transition path. While the model labor force. chose to have children out-of-wed­ is successful in replicating the styl­ Nechyba models the fertility deci­ lock, and the effect of each genera­ ized facts on AFDC and illegitimacy sions of individuals who differ in tion diminishes with time. Nechyba and establishes a link between the their wage rate and their preferences calculates a steady-state level of so­ two through a government induced toward rearing children. More pre­ cial approval in the absence of wel­ change in "values," it also demon­ cisely, each type of person chooses fare programs, and then introduces a strates that welfare reform aimed at her amount of leisure and whether to program similar to AFDC, and calcu­ reducing the incentives for poor have a child out-of-wedlock. Chil­ lates a transition path to the new women to have out-of-wedlock births dren are considered "consumption steady state. He demonstrates that may not be as effective as policy­ goods," yielding utility but costing along the transition path, observed makers who believe in a causal link leisure time. Furthermore, how much cases of illegitimacy are rising both between AFDC and illegitimacy utility individuals receive from hav­ among the poor and the non-poor might suspect. Finally, a spatial ex­ ing a child out-of-wedlock depends despite the fact that AFDC payments tension of the model could explain on the level of "social approval" that are constant or even falling. He then not only the stylized aggregate trends, is associated with having out-of-wed­ performs policy simulations of wel­ but also the empirically important lock children. This social approval is fare reform proposals, from the concentration of out-of-wedlock births a function of the fraction of individu- steady state and at different points on in some communities.

Volatility permeates modern finan­ strategies. In response to this, a volu­ tence of volatility. Meanwhile, when cial theories and decisionmaking minous literature has emerged for judged by standard forecast evalua­ processes. As such, accurate mea­ modeling the temporal dependencies tion criteria, standard volatility mod­ sures and good forecasts of future in financial market volatility at the els seemingly provide poor forecasts. volatility are critical for the imple­ daily and lower frequencies. Most of Andersen and Bollerslev demon­ mentation and evaluation of asset these studies find highly significant strate that, contrary to this conten­ and derivative pricing theories, as in-sample parameter estimates and tion, in empirically realistic situations well as for trading and hedging pronounced intertemporal persis- the models actually produce strik-

32. NBER Reporter Winter 1997/8------ingly accurate interdaily forecasts for going questions on the historical model for stock returns. They find the latent volatility factor that is of mean and standard deviation of the that, far from being excessive, the interest in financial applications. return on equities and bonds and on actual correlations fall below what They also present new methods for the equilibrium demand for these they should be in a unified market. construction of more accurate and securities. The constraint turns out to By way of comparison, they then meaningful ex-post interdaily volatil­ be binding for the young consumers study purely domestic data. They .;, ity measurements extracted from and plays a different role than in close the investigation with an exam­ high-frequency intradaily data. models with infinitely-lived con­ ination of the hypothesis of market Brennan, Chordia, and Subrah­ sumers. The model also highlights segmentation. manyam re-examine the relation­ the function of equity as a hedge The market portfolio (world port­ ship between stock returns, measures against future wage uncertainty, a folio) is in one sense the least impor­ of risk, and a set of other (non-risk) property that has diminishing value tant portfolio to provide to investors; characteristics of securities, including as investors age. there is always a better portfolio for the book-to-market ratio, firm size, Grossman and Zhou develop a social planners to make available to the stock price, the dividend yield, and Simple dynamic general equilibrium them. In a J-agent one-period sto­ lagged returns. Their primary objec­ model to explain: the excessive de­ chastic endowment economy, where tive is to determine whether these pendence of a country's consump­ preferences are quadratic, the market other characteristics have marginal tion on its own income; the home portfolio is never spanned by the explanatory power relative to the equity bias; and the foreign exchange optimal markets a social planner APT benchmark. They use, in turn, risk premium. These "anomalies" are would create. With identical prefer­ the Connor and Korajczyk 1988 (CK) seen to be a direct consequence of ences, the market portfolio is or­ and the Fama and French 1993 (FF) the fact that a country cannot com­ thogonal to all J-1 portfolios which approaches. Fama-MacBeth type pletely equitize the future income achieve a first best solution. These regressions using risk adjusted re­ derivable from its natural and human conclusions rely on the assumption turns provide evidence of return mo­ resources. that the social planner has perfect mentum, size, and book-to-market In an integrated world capital mar­ information about agents utilities. effects, together with a significant ket, the same pricing kernel is applic­ Athanasoulis and Shiller also show and negative relation between re­ able to all securities. If the kernel is that as the contract designer's infor­ turns and trading volume, even after exceSSively volatile, as has been found mation about agents utilities becomes accounting for the CK factors. When in Some studies, it should translate more imperfect, the optimal contracts the analysis is repeated using the into an excessive degree of correla­ will approach contracts that weight FF factors, they find that the size and tion in the returns of different stock individual endowments in proportion book-to-market effects are attenu­ securities. Dumas, Harvey, and to elements of eigenvectors of the ated, and their significance is weak­ Ruiz apply this idea mostly to the variance matrix of endowments. If ened as well, while the other effects stock returns of different countries. there is a substantial market compo­ persist. In addition, after adjusting for They determine first, for a given mea­ nent to endowments then a social risk using either method, Nasdaq stocks sured degree of commonality in planner, for reasons of robustness show significant underperformance. country outputs, what should be the and Simplicity, may conclude that In the context of a stationary, over­ degree of correlation of stock returns. creating a contract to allow trading lapping generations economy in They then match the second mo­ the market portfolio would be a sig­ which consumers are subject to a bor­ ments of the combined model con­ nificant innovation. rowing constraint, Constantinides, taining the statistical multivariate Donaldson, and Mehra address on- model for output and the financial

------NBER Reporter Winter 1997/8 33. Hu tests whether teenagers' deci- source of variation in teen drinking, mothers on a path permanently fraught sions to leave or stay in welfare fam- and data from the 1977-92 Monitor- with adversity, although they do sug- j' ilies respond predictably to welfare ing the Future surveys, to demon- gest that a young woman's life changes iIi benefits. He uses a unique dataset strate that teens who faced an MLDA in substantial ways. Although failure I' . that is particularly attractive in two of 18 were substantially more likely to delay childbearing Significantly I 'I'I: i· ways: 1) variation in welfare benefits to drink than teens who faced a reduces the likelihood that women il:;, :r arises from a controlled social exper- higher drinking age. Then using data will obtain a high school diploma, 1;1 iment in California; and 2) the longi- from over 1.3 million respondents teen mothers have a higher probabil- i. 1', tudinal survey provides information who belong to the 1960-9 birth ity than others of obtaining a general on the whereabouts of all children, cohorts in the 1990 Public-Use Micro- equivalency diploma. Teenage child- including those who have left the data Sample, they find that changes bearing does not lead to a reduction household. Hu finds that teenagers in the MLDA had small and statisti- in self-sufficiency or to an increase in with children of their own respond cally insignificant effects on measures dependence on government-spon- to higher benefits by moving out of of educational attainment such as sored public assistance programs, their parents' households. These find- high school completion, college en- they find. While teenage childbearing ings imply that increasing income trance, and college completion. The does reduce earnings and increase through welfare benefits has a per- authors conclude that teen drinking welfare use for a few years after the verse destabilizing effect on poor does not have an independent effect birth of the first child, later in life, as families and alters the share of on educational attainment. her children age and as women who income received by different family Hotz and Sanders have been delayed childbearing begin to have members. engaged in a series of research pro- children· of their own, a teen mother Recent research has suggested that jects that attempt to exploit a "natural becomes more likely to work, will one of the important, life-cycle con- experiment" associated with human earn more income, and is less likely sequences of teen drinking is re- reproduction to measure the causal to participate in government-spon- duced scholastic achievement. Fur- effects and costs of early childbear- sored public assistance programs. In thermore, it has been argued that ing. In these papers, they compare fact, by her early thirties, a teenage state excise taxes on beer and mini- the behavior of teen mothers with mother has worked thousands more mum legal drinking ages (MLDA) can that of women who became preg- hours, and as a consequence has have a positive impact on educa- nant as teenagers but suffered a mis- substantially higher wages, than if tional attainment. Dee and Evans carriage. The results do not support she had delayed childbearing. The , use the increases in the state MLDA the common belief that childbearing authors find that very little of the 'i during the late 1970s and 1980s as a during adolescence places the young approximately $11.3 billion the gov-

34. NBER Reporter Winter 1997/8------,....--- ernment spends on various forms of trol for unobserved factors related to improved medical care may increase public assistance for teen mothers both child outcomes and pregnancy the survival of extremely premature could be saved if all women who intentions. An interesting finding of infants, thus lowering average birth­ bore their first child as a teen were to the analysis is that unwanted preg­ weight of "liveborn" infants. Improved delay that birth by three to four nancy is associated with less healthy medical care also may improve the years. Furthermore, once one takes prenatal and postpartum behaviors, nutritional status of newborns of any .. account of the increased tax rev­ but has little association with infant given gestational age, thus increasing enues that teenage mothers provide health or child cognitive outcomes. average birthweight. Prematurity and through greater work effort, govern­ Women whose pregnancies are weight-for-gestational-age are sepa­ ment would have higher net public unwanted are more likely to delay rable newborn health characteristics, assistance costs if all teen births were prenatal care, smoke during preg­ with different causes and different postponed. nancy, and are less likely to breast health and developmental conse­ One current educational reform feed than women whose pregnancies quences. (The "child health" web­ seeks to reward the "value added" by are wanted. Despite these associa­ page will also include access to a teachers and schools based on the tions, women whose pregnancies are down-Ioadable SAS file which will average change in pupil test Scores unwanted have infants of similar allow researchers to use birthweight, over time. A key assumption of this birth weight, and children that expe­ gestational age, gender and ethnicity policy is that socioeconomic out­ rience no apparent cognitive deficits to classify children's nutritional status comes are a linear function of test as compared to children from wanted at birth.) scores. Using the National Longitudi­ pregnancies. BIau estimates the effect of group nal Survey of Youth, Cawley, Heck­ Norberg gave an overview of the size, staff-child ratio, training, and man, and Vytlacil find a nonlinear features of a large number (40+) of other characteristics of child care on relationship between test scores and publicly available, nationally repre­ the behavioral and mental develop­ (log) wages. They find no consistent sentative databases which have infor­ ment of children using data from the pattern in the curvature of log wage mation pertaining to children's health. National Longitudinal Survey of returns to test Scores. This implies (This information is now available on Youth. In contrast to the data used in that, used alone, the average gain in the NBER web page: from the NBER most previous research on this sub­ test scores is an inadequate measure home page, click on "online data;' ject, the sample of children is large of school performance. and from the online data page, select and nationally representative, the Joyce, Kaestner, and Korenman "links to child health data sets"; this data contain good measures of the investigate the causal link between will take you to a table which pre­ home environment, and there are unintended pregnancy and child sents general features of the available repeated measures of the child devel­ health and development, and paren­ surveys. The table also includes ac­ opment outcomes of interest. The tal behaviors related to child health tive links to websites containing more results show that child care charac­ and development. Their primary information: variable lists, ordering teristics have little association with motivation is that past studies may information, and the like for each child development on average. Asso­ not have controlled adequately for dataset.) She also discussed the use ciations are found for some groups the confounding effects of family and of vital statistics in assessing newborn of children, but they are as likely to social background. To address this health. For example, many studies be of the "wrong sign" as they are to problem, they include an extensive continue to use "low birth weight" as be of the sign predicted in the devel­ set of controls. In addition, they ex­ an indicator of newborn health and opmental psychology literature. ploit information on siblings to con- the quality of medical care. However,

------NBER Reporter Winter 1997/8 35. Katz, Kling, and Liebman exam­ ban and other wealthier communi­ large expansions of the Earned In­ ine the short-run impacts of a change ties, while regular Section 8 assis­ come Tax Credit (EITC) and Medi­ in residential neighborhood on the tance is modestly more effective in caid, changes in the Aid to Families well-being of low-income families. getting a larger share of families out with Dependent Children program, They look at the experiences of 540 of the most distressed communities. and in related training and child care families at the Boston site of Moving Families in all groups report that the programs. Many of the changes were To Opportunity (MTO), a demonstra­ primary reason they want to move intended to encourage low income tion program currently underway in out of public housing is fear of crime. women to work. Meyer and Rosen­ five cities. Families in eligible public The Experimental program appears baum examine the effects of these housing projects in high poverty to have been rather successful in changes on the employment of sin­ Census tracts can apply to MTO and addressing this major concern of the gle mothers. They find that a large are assigned by lottery to one of participants: participants surveyed share of the sharp increase in work three groups. The Experimental group from one to three years after program by single mothers in recent years receives some counseling assistance enrollment indicate much lower can be attributed to the EITC, with and a Section 8 rental subsidy that criminal victimization rates and much smaller shares for welfare benefit can be used to move to a Census higher neighborhood safety than reductions and other changes in wel­ tract that had a poverty rate of less members of the Control group. The fare programs. than 10 percent in 1990. The Section Section 8 program group shows Why do countries have large, per­ 8 Comparison group receives a geo­ somewhat more modest improve­ sistent differences in fertility that are graphically unrestricted rental sub­ ments in safety and reductions in not explained by income levels or by sidy. The Control group continues in criminal victimization. Further, em­ women's labor force participation? public housing and receives no new ployment rates and earnings levels of Applying Becker's intergenerational rental assistance or services. The household heads in all groups have utility function to an analysis of mi­ authors find that one to three years essentially doubled from 1994 to gration as investment, Berman and after participants enter the program, 1997. Also, children in the Experi­ Rzakhanov argue that migration both Experimental and Section 8 mental and Section 8 Comparison costs and differences in altruism Comparison families are fairly suc­ groups attend schools with substan­ induce positive selection. They com­ cessful in using their subsidies to tially higher test scores, but there ap­ pare members of a single birth cohort move out of high poverty neighbor­ pears little difference among groups who migrated in different periods hoods: 48 percent of Experimental on reported absenteeism from school from Eastern Europe to Israel. Mi­ and 58 percent of Section 8 Com­ or hours spent on homework. grants in the early period when parison families move through the In recent years, there have been migration was more costly average program. The Experimental group is enormous changes in welfare and tax 0.9 more children over their lifetime. much more likely to move to subur- policy. In particular, there have been A comparison with women migrating

36. NBERReporter Winter 1997/8------after their fertile years shows that each district, they find that states model analyzes both schooling choices positive selection accounts for about where the school finance system was and post-school on-the-job invest­ two thirds of this large difference in declared unconstitutional in the ment in skills in a framework in fertility rates. 1980s increased the relative funding which different schooling levels serve Card and Payne study the effects of low-income districts. Using micro as an index for different skills. They of school finance reforms on the dis­ samples of SAT scores from this same find that immigration of low skill tribution of school spending across period, they find some evidence that workers contributes little to rising richer and poorer districts, and the the equalization of spending aCross wage inequality. When they extend consequences of spending equaliza­ districts leads to a narrowing of test their model to account for the Baby tion for the distribution of SAT scores score outcomes across family back­ Boom generation, they find that the across children from different family ground groups. same estimates of the supply of backgrounds. Using school district Heckman, Lochner, and Taber human capital that can explain the data from the 1977 and 1992 Cen­ develop and estimate a model of wage history of the last 15 years also suses of Governments to measure the labor earnings, skill formation, and explain the last 35 years of wage correlation between state funding per physical capital accumulation with inequality. pupil and median family income in different levels of human capital. The

Christiano, Eichenbaum, and Changes in government spending In recent years, a number of cen­ Evans develop and implement a strat­ often lead to significant shifts in tral banks have announced numeri­ egy for assessing the plausibility of demand across sectors. Ramey and cal inflation targets as the basis for monetary business cycle models Shapiro estimate the effects of mili­ their monetary strategies. After out­ which focuses on a model's ability to tary buildups on a variety of macro­ lining the reasons why such strate­ reproduce empirical estimates of ac­ economic variables using a new gies might be adopted in the pursuit tual economies' responses to mone­ measure of military shocks. The of price stability, Mishkin and Posen tary policy shocks. Based on their re­ behavior of macroeconomic aggre­ examine the adoption, operational sults with MI, they argue that models gates is consistent with the predic­ design, and experience of inflation in which monetary non-neutralities tions of a multi-sector neoclassical targeting as a framework for mone­ only reflect the effects of unantici­ model. In particular, consumption, tary policy in the first three countries pated movements in money will fail real product wages and manufactur­ to undertake such strategies: New this test. But their limited participa­ ing productivity fall in response to Zealand, Canada, and the United King­ tion model of money does not fail exogenous military buildups in the dom. They also analyze the operation this test. post-World War II United States. of the long-standing German mone-

------NBER Reporter Winter 1997/8 37. t..

tary targeting regime, which incorpo­ transactions become negligible. The total amount of assets. The implica­ rated many of the same features as determination of equilibrium inflation tion is that, instead of studying money later inflation-targeting regimes. The in the cashless limit also provides a demand using time series and look­ key challenge for all these monetary useful approximation of the case of ing at historical interest rate varia­ frameworks has been the appropri­ an economy in which monetary fric­ tions, one can look at a cross-section ate balancing of transparency and tions are present, but small, or in of households and analyze variations flexibility in policymaking. This study which monetary frictions vary over in the amount of assets held. The finds that all of the targeting coun­ time, which may result in substantial authors find that the elasticity of tries examined have maintained low instability of money demand in per­ money demand is very small when rates of inflation and increased the centage terms. Inflation in the cash­ the interest rate is low; the probabil­ transparency of monetary policy­ less limit is a function of the gap ity that a household holds any making without harming the real between the "natural rate" of interest, amount of interest bearing assets is economy through policy rigidity in determined by the supply of goods related positively to the level of fi­ the face of economic developments. A and opportunities for intertemporal nancial assets; and the cost of adopt­ convergence of design choices on substitution, and a time-varying para­ ing financial technologies is related the part of targeting countries with re­ meter of the interest-rate rule in­ positively to age and negatively to gard to operational questions emerges dicating the tightness of monetary the level of education. The finding from this comparative study, suggest­ policy. Inflation can be completely that the elasticity is very small for ing some lines of best practice for stabilized, in principle, by adjusting interest rates below 5 percent sug­ inflation-targeting frameworks. the policy parameter so as to track gests that the welfare costs of infla­ Woodford shows that it is possi­ variation in the natural rate. tion are small. At interest rates of 6 ble to analyze equilibrium inflation Mulligan and SaIa-l-Martin argue percent, the elasticity is close to 0.5, determination without any reference that the relevant decision for the and roughly one half of this elasticity to either money supply or demand, majority of u.s. households is not the can be attributed to the Baumol­ as long as one specifies policy in fraction of assets to be held in inter­ Tobin or intensive margin, and half terms of a "Wicksellian" interest-rate est bearing form, but whether to hold of it can be attributed to the new feedback rule. His central result is a any such assets (we call this "the adopters or extensive margin. The theorem that shows the existence, for decision to adopt" the financial tech­ intensive margin is less important at a simple monetary model, of a well­ nology). They show that the key vari­ lower interest rates and more impor­ behaved "cashless limit" in which the able governing this decision is the tant at higher interest rates. money balances held to facilitate product of the interest rate times the

38. NBER Reporter Winter 1997/8------Household survey data consis- the population implicitly priced annu- behavior, and uses this relationship tently depict large variations in saving ities in a way that was considerably to show that measured levels of and wealth, even among households off a reasonable market price, al- patience in this economy exhibit a with similar socioeconomic charac- though many people were in a rea- substantial degree of variation, both teristics. Within the context of the life sonable range. The sequence of across individuals, and within a sin- cycle hypothesis, families with iden- questions asked affected the answers. gle individual's lifespan. He predicts tical lifetime reSOurces might choose 3) Women tended to have lower that measured levels of patience will \ to accumulate different levels of wealth answers for a given income increase with wealth and age. The ) wealth for a variety of reasons, in- question than men. 4) People in the variation arises because the effective cluding variation in time preference highest income bracket tended to horizons of consumers also vary over rates, risk tolerance, exposure to un- have higher wealth answers relative the life-cycle of a buffer stock con- certainty, relative tastes for work and to lower income people. sumer. For hyperbolic consumers, leisure at advanced ages, income re- Di Tella, MacCulloch, and Os- long-horizons are associated with placement rates, and so forth. These wald use data on the reported well- choices that appear patient, while factors can be divided into a small being levels of approximately one short horizons are associated with number of classes, each with a dis- quarter of a million randomly sam- choices that appear impatient. By tinctive implication concerning the pled Europeans and Americans from contrast, consumers with exponential relation between accumulated wealth the 1970s to the 1990s. After control- discount functions make equilibrium and the shape of the consumption ling for personal characteristics, year choices which are uniformly patient profile. By examining this relation dummies, and country fixed effects, and exhibit no interpersonal or in- empirically, one can test for the pres- they find that the data trace out a trapersonal variation. ence or absence of these particular function that is approximately a lin- "Money illusion" means that peo- explanations for differences in wealth. early additive "misery index." They pIe behave differently when the Using the Panel Study of Income Dy- also calculate the implied dollar value same objective situation is repre- namics and the Consumer Expendi- of a low inflation rate, and examine sented in nominal versus real terms. ture Survey, Bernheiln, Skinner, and the structure of happiness equations To examine the behavioral impact of Weinberg find very little support for across countries and time. money illusion, Fehr and Tyran life cycle models that rely on the above Shiller discusses the Unidad de study the adjustment process of nom- factors to explain wealth variation. Fomento (UF), an indexed unit of inal prices after a fully anticipated The data are, however, consistent with account, which has been used in negative nominal shock in an exper- "rule of thumb" or "mental account- Chile since 1967, and which has been imental setting with strategic com- ing" theories of wealth accumulation. copied in Colombia, Ecuador, Mex- plementarity. They show that seem- "Wealth illusion" refers to the ten- ico, and Uruguay. The UF is a sort of ingly innocuous differences in payoff dency of some people to overvalue a money that has guaranteed real pur- presentation cause large differences stock of wealth relative to the income chasing power, although it is not a in behavior. In particular, if the pay- flow it can finance. Diamond and currency and is not used as a me- off information is presented to sub- Kahneman report on a pilot study dium of exchange. Shiller argues that jects in nominal terms, then price that asked people about the wealth important practical problems in im- stickiness and real effects are much and annuitized real income needed plementing indexation are solved by more pronounced than when payoff for an adequate retirement income creating such units of account. The information is presented in real for a couple described to them. Among use of the units in wagesetting need terms. The driving force behind dif- the preliminary results of the pilot not be inflationary, if the units are ferences in real outcomes is that sub- study are: 1) The coefficient of varia- defined properly. jects expect much more nominal tion tended to be larger for the wealth Studies of animal and human be- inertia in the nominal payoff condi- question than the income question, havior suggest that discount func- tion. Because of strategic comple- reflecting the greater difficulty in tions are approximately hyperbolic. mentarity, these expectations induce , thinking about the wealth needed for Laibson integrates the literature on subjects to adjust rather slowly to the retirement than the income needed hyperbolic preferences with the liter- shock. I for retirement. 2) A sizable fraction of ature on buffer stock consumption

------NBER Reporter Winter 1997/8 39. LaPorta and his coauthors analyze insider trading activities of all com­ proportion of funds from internal why firms pay dividends, the so­ panies listed on the NYSE, Amex, sources increases with age, while the called "dividend puzzle," from the and Nasdaq exchanges during proportion from banks, venture cap­ agency perspective. They outline two 1975-95. Insiders on average pur­ italists, and private investors declines. agency models of dividends: 1) "the chase shares worth 0.6 percent, and There is also evidence that these pat­ outcomes" model, in which divi­ sell shares worth 1.3 percent, of their terns eventually reverse themselves, dends are the result of effective pres­ companies market capitalization each with the proportion of insider finance sure by minority shareholders to year. Large shareholders trading is ultimately declining and the propor­ force corporate insiders to disgorge much less frequent than, but its dol­ tion of outsider finance increasing cash-in this model, stronger minor­ lar volume is comparable to, man­ with age. These findings are consis­ ity shareholder rights should be asso­ agers' trading. In general, there is tent with elements of both reputa­ ciated with higher dividends; and 2) very little market movement in the tion-based and monopoly-lender "the substitutes" model, in which period immediately around insiders theories of firm finance. insiders choose to pay dividends to trading; insiders in the aggregate are Gomes studies the problem of establish a reputation for decent contrarian investors. However, they going public in the presence of moral treatment of minority shareholders, predict market movements better hazard, adverse selection, and multi­ so that firms can raise equity finance than a pure contrarian strategy. In­ ple trading periods. In the multi­ in the future. Under this model, siders' purchases are better predictors period game, managers strategically stronger minority shareholder rights of future long-term stock returns than choose the level of extraction of pri­ reduce the need for establishing a insiders' sales. In addition, insiders' vate benefits and can develop a good reputation, and should be associated activities in smaller firms are more reputation for expropriating low lev­ with lower dividends. The authors informative than their activities in els of private benefits. The costs of compare these models on a cross­ larger firms. going public can be reduced signifi­ section of 4,000 companies from Using data from the Wisconsin En­ cantly because of this reputation around the world, in countries with trepreneurial Climate Study, Fluck, effect, and this can be an important different levels of investor protection, HoltZ-Eakin, and Rosen study the factor in sustaining emerging stock and therefore different strength of sources of finance during the very markets that offer weak protection to minority shareholder rights. Their early stages of firms lives. They focus minority shareholders. Also, allowing findings on payout levels, as well as on the evolution of the mix of finan­ controlling managers to issue non­ other results, support the outcomes cial capital from "insiders" and "out­ voting shares can increase the stock model of dividends. siders" as firms age. They find that at market effiCiency, because the repu­ Lakonishok and Lee document the beginning of firms' life cycles, the tation effect is stronger when man-

40. NBERReporter Winter 1997/8------agers can divest more without losing ing rates of return, or Tobin's Q, for anism used to align the incentives of control. example). Contractual choices and management with ownership, the au­ Firms are governed by a network performance outcomes are deter­ thors examine the determinants of of relationships representing, for mined endogenously by exogenous managerial ownership as a function example, contractual arrangements for (and, in some case, unobserved) of the contracting environment. They financing, capital structure, and man­ changes in the firm's contracting find that managerial ownership is ex­ agerial ownership and compensation. environment. Himme1berg, Hub­ plained by key variables in the con­ Accordingly, it is difficult to focus on bard, and Palia confront this prob­ tracting environment and that, once a single such arrangement and to lem in the context of the firm's they control for observed firm char­ identify the correspondence between compensation contract for managers. acteristics and firm fixed effects, there the contractual choice and firm per­ Because managerial equity stakes are is no effect of changes in firm man­ formance (as measured by account- an important and well-known mech- agerial ownership on performance.

Greenstone presents new evi­ a county's pollutant-specific high­ 1978 is the annual designation of dence on the impact of federal envi­ regulation designation significantly county air quality attainment status, ronmentallaws on the manufacturing affected the growth of employment, where non-attainment status triggers sector by using pollution categories the rate of net investment, and the specific equipment requirements for established by the EPA after the pas­ growth of shipments of polluting new and existing plants. In the later sage of the Clean Air Act Amend­ plants relative to non-polluting ones. years of regulation, l}onattainment ments of 1970. Specifically, the EPA Moreover, the effects differed across status reduces expected "births" in designated high or low regulation the four regulations and across plants polluting industries by 40-50 per­ status for four pollutants to each decisions to enter, exit, expand or cent, resulting in a shift of polluting county, depending upon whether the contract. activity to cleaner, less populated national air quality standards were Using plant data for 1963 to 1992, attainment areas. Starting in the met there. Controlling for a wide set Becker and Henderson examine 1970s, effects appear first for indus­ of factors, including transitory shocks the unintended effects of air quality tries with bigger plant sizes and then, to polluting industries and plant char­ regulation on decisions of major pol­ within industries, first for corporate acteristics, Greenstone finds that luters. A key regulatory tool since plants relative to the much smaller

------NBER Reporter Winter 1997/8 41. non-affiliate, or single-plant firm sec­ dence of positive regional externalities. effective in mills that have more com­ tor. In all industries, non -affiliates Benkard introduces a new cost plex production processes. face less regulation than the bigger dataset for a commercial aircraft firm Using data from a unique nation­ corporate plants, resulting in a per­ and uses it to analyze the dynamics ally representative sample of busi­ manent shift away from corporate of learning in commercial aircraft nesses, the Educational Quality of the plant production in some industries. production. The dataset is not con­ Workforce National Employers Sur­ Older plants benefit from grand fa­ sistent with the simple learning vey, matched with the Bureau of the thering provisions, greatly enhancing hypothesis, and particularly the pre­ Census Longitudinal Research Data­ survival probabilities. Finally, the diction that a firm's unit cost must base, Black and Lynch examine the negotiation and permitting process decline with its cumulative produc­ impact of workplace practices, infor­ under regulation appears to induce tion. Instead, there is strong support mation technology, and human capi­ much greater up-front investments by for the hypothesis of organizational tal investments on productivity. They new plants, so that, in non-attain­ forgetting, a slightly mOre general use both cross section and panel data ment areas, regulation induces learning model in which unit costs covering 1987-93, and find that what 50-100 percent increases in initial are similarly dependent on a firm's is associated with higher productiv­ plant sizes compared to attainment past production experience, but ity is not so much whether or not an areas. But for plants over 10 years of where that experience depreciates employer adopts a particular work age there are no size differences. overtime. practice, but rather how that work Do firms become more efficient The use of group-based incentive practice is actually implemented after becoming exporters? Do export­ pay and problem-solving teams is within the establishment. In addition, ers generate positive externalities for increasingly common in manu­ their results suggest that those union­ domestically-oriented producers? facturing, yet there is little analysis ized establishments that have adopted Clerides, Lach, and Tybout tackle that assesses the value of these prac­ what have been called new or "trans­ these questions by analyzing the tices. Boning, Ichniowski, and formed" industrial relations practices causal links between exporting and Shaw begin their assessment by that promote joint decisionmaking, productivity using plant-level data. developing a simple model in which coupled with incentive-based compen­ They look for evidence that cost pro­ employees decide on effort levels in sation, have higher productivity than cesses of firms change after they break two possible tasks: producing output other similar non-union plants, while into foreign markets. They find that and working on problem-solving those businesses that are unionized relatively efficient firms become ex­ teams. They show that the gains to but maintain more traditional labor porters; however, in most industries, these innovative incentive practices management relations have lower firms' costs are not affected by previ­ are likely to vary across different productivity. They also find that the ous exporting activities. So the well­ manufacturing environments. Using higher the average educational level documented positive association be­ data from steel minimill production of production workers or the greater tween exporting and efficiency is lines, they find that incentive plans the proportion of non-managerial explained by the self-selection of and problem-solving efforts raise per­ workers who use computers, the the more efficient firms into the ex­ formance on the production line. higher is plant productivity. port market. They also find some evi- However, problem-solving is most

42. NBER Reporter Winter 1997/8 ------Current Working Papers

NBER Working Papers trade and the environment through increased the demand for separate negotiation on policy instruments living quarters, but that demand Trade and Environment: provides significantly inferior devel­ depended upon the rules of the pro­ oping country outcomes. Thus, a gram, in particular whether children Bargaining Outcomes from trade and environment policy-linked were held legally responsible for the Linked Negotiations negotiation may be better than an care of their aged parents. I argue Lisandro Abrego, environment-only negotiation, but that almost half of the decline in the Carlo Perroni, John Whalley, and negotiating compensation to devel­ fraction of older nonmarried women Randall M. Wigle oping countries for environmental living with relatives from 1950 ito NBER Working Paper No. 6216 restraint would be better still. We 1990 can be attributed to rising Social October 1997 provide sensitivity and other analysis Security benefits and expanded eligi­ International Trade and Investment of our results and indicate what fac­ bility and to the fact that Social tors could qualify our main finding, Security benefits were given with no Recent literature has explored strings attached. both physical and policy linkage including the erosion of the most favored nation principle involved between trade and the environment. with environmentally based trade Gradual Incorporation of We explore linkage through leverage actions. Information into Stock in bargaining, whereby developed Prices: Empirical Strategies countries can use trade policy threats A House of Her Own: Sara Fisher Ellison and to achieve improved environmental Old Age Assistance and the Wallace P. Mullin management in a developing coun­ Living Arrangements of NBER Working Paper No. 6218 try, while developing countries can Older Nonmarried Women October 1997 use environmental concessions to Dora L. Costa JEL Nos. G1, G14, e1, LOO achieve trade discipline in developed NBER Working Paper No. 6217 Asset Pricing countries. We use a global numerical October 1997 simulation model to compute bar­ This paper explores environments JEL Nos. J14, N12 gaining outcomes from linked trade in which either the revelation or dif­ Aging, Development of the fusion of information, or its incorpo­ and environment negotiations, com­ American Economy paring developed-developing coun­ ration into stock prices, is gradual. try bargaining only on trade policy I show that the trend towards sin­ We also develop appropriate esti­ with joint bargaining on both trade gle households among older non­ mation techniques. Our paper has and domestic environmental policies. married women, the majority of whom implications both for event study Our results indicate joint gains from were widows, has been ongoing only methodology and for understanding expanding the trade bargaining set to since 1940. I investigate the factors the process by which stock prices include the environment, as opposed that fostered the rise in separate liv­ incorporate information. We highlight to the current developing country ing quarters since the middle of this two environments. reluctance to negotiate in the World century by examining the impact of First, information often is not Trade Organization on this issue. Old Age Assistance on living arrange­ revealed in one announcement but However, compared to bargaining ments in 1940 and 1950. I find that rather through a process of gradual with cash side payments, linking Old Age Assistance substantially public revelation, which may not

------NBER Reporter Winter 1997/8 43. to make participation in the new sys­ tem voluntary may appeal to policy­ makers who need to have the political support of workers. One critical issue in evaluating such a reform and its economic conse­ quences, though, is the extent of unknown differences in households' preferences for consumption. I esti­ mate the distribution of rates of time preference from the wealth data in the 1992 Survey of Consumer Finances using a flexible life-cycle model of consumption under income uncertainty. I then apply this esti­ mated distribution to a variety of reform proposals that incorporate a voluntary choice of how much to contribute to a dedicated retirement account and a rebate of the existing payroll tax that increases with the magnitude of the contribution. My main finding is that an appropriate menu of reform plans can induce the voluntary buyout of 84 percent of existing payroll taxes at an immediate cost to national saving of less than 0.25 percentage point.

The Misallocation of Housing Under Rent Control Edward L Glaeser and Erzo F. P. Luttmer NBER Working Paper No. 6220 be completely observable by a re­ to maximize profits dynamically. We October 1997 searcher. We examine the effect of use the functional form predictions JEL Nos. R20, D45, H80 the evolution of the Clinton health from Kyle in our estimation, and the Public Economics care reform proposal on pharmaceu­ results from a sample of targets of tical stock prices. We estimate by iso­ tender offers are consistent with the When there are binding price con­ tonic regression the expected path of model. trols, there are shortages, and the mar ket -ad jus ted p harmaceu tical allocation of goods across consumers may not be efficient. In general, the prices over September 1992-0ctober Discount Rate costs of price controls in terms of 1993, and find that the major portion Heterogeneity and misallocation are first order, while the of the decline in stock prices Social Security Reform occurred gradually, and did not cor­ classic welfare losses attributable to Andrew A. Samwick respond to identified news events. undersupply are second order. We NBER Working Paper NO. 6219 Second, the trading process itself present an empirical methodology October 1997 may incorporate private information for estimating the degree of misallo­ JEL Nos. H55, E21 into stock prices gradually. That is an cation of housing units caused by Aging and Public Economics implication of the Kyle [19851 model, rent control in New York City. This in which one or a small number of As many countries consider the methodology involves comparing the informed traders use their market privatization of existing pay-as-you­ relative consumption of different power over their private information go social security systems, the option demographic groups within the rent

44. NBERReporter Winter 1997/8------controlled area with the relative lev­ Option Hedging Using Trade and Security, I: els of consumption in a free market Empirical Pricing Kernels Anarchy area. Our best estimate of the costs Joshua V. Rosenberg and James E. Anderson and of rent control in New York attribut­ Robert F. Engle Douglas Marcouiller able to the misallocation of rental NBER Working Paper No. 6222 NBER Working Paper No. 6223 apartments is $200 per apartment October 1997 October 1997 annually. Asset Pricing JEL Nos. D7, Fl International Trade and Investment We develop a method for option hedging which is consistent with Market exchange is subject to an time-varying preferences and proba­ endogenously determined level of Jobless Growth: bilities. The preferences are ex­ predation which impedes speciali­ Appropriability, Factor pressed in the form of an empirical zation and gains from trade. We con­ Stabilization, and pricing kernel (EPK), which mea­ struct a model in which utility­ Unemployment sures the state price per unit proba­ maximizing agents choose between Ricardo J. Caballero and bility. Probabilities are derived from careers in production and careers in Mohamad L Hammour an estimated stochastic volatility predation. Three types of equilibrium NBER Working Paper No. 6221 model of the form GARCH compo­ may emerge: autarky (with no pre­ October 1997 nents with leverage. We estimate dation and no defense); insecure JEL Nos. EO, E2, )6, 03 state prices using the flexible risk­ exchange equilibriums (with preda­ Economic Fluctuations and Growth neutral density method of Rosenberg tion and defense); and secure ex­ A central determinant of the polit­ [1995] and a daily cross-section of change equilibriums (in which ical economy of capital-labor rela­ option premiums. Time-varying pref­ defense completely deters preda­ tions is the appropriability of specific erences over states are linked to a tion). Trading equilibriums, two quasi-rents. This paper is concerned dynamic model of the underlying thirds of them secure, are supported with the general-equilibrium interac­ price of a one-day ahead forecast of only in a narrow range of security tion of appropriability and the char­ derivative price distributions and parameter values. Since changes in acteristics of technology: namely, the minimum variance hedge ratios. the technologies of defense and pre­ embodiment of technology in capi­ Our empirical results suggest that dation have terms-of-trade effects, tal, and substitutability between cap­ risk aversion over S&P500 return some producers may be hurt by ital and labor in the technological states is substantially higher than risk enhanced security. We show cases of menu. Technological embodiment aversion implied by Black-Scholes "irnmiserizing security" in which pro­ means that the supply of capital is state prices and probabilities using ducers in large poor countries are effectively much less elastic in the long-run estimates of S&P500 return harmed by increased security. short than in the long run, and there­ moments. We also find that the daily fore is more exposed to appropri­ level of risk aversion is strongly pos­ ability. Technology choice implies itively autocorrelated, negatively cor­ that an attempt at appropriating cap­ related with S&P500 price changes, Both Sides of Corporate ital will induce a substitution away and positively correlated with the Diversification: The Value from labor in the long run, and this spread between implied and objec­ Impacts of Geographic and constitutes a mechanism for thwart­ tive volatilities. Industrial Diversification ing appropriation. Shifts in European Hedging results reveal that typical Gordon M. Bodnar, Charles Tang, labor relations in the last three hedging techniques for out-of-the­ and Joseph Weintrop decades offer a good laboratory for money S&P500 index options, such NBER Working Paper No. 6224 exploring the empirical relevance of as Black-Scholes or historical mini­ October 1997 those mechanisms. The evolution of mum variance hedging, are inferior )EL Nos. F3, G3 the labor share, the profit rate, the to the EPK hedging method. Thus, International Finance and Macroeconomics capital/output ratio, and unemploy­ time-varying preferences and proba­ ment-which we examine mOre par­ bilities appear to be an important fac­ We examine the effect of geo­ ticularly in the case of France­ tor in the day-to-day pricing of graphic and industrial diversification appears highly supportive. S&P500 options. on firm value for a sample of over

------NBER Reporter Winter 1997/8 45. 20,000 (firm-year) observations of series properties of claims under var­ and wealth, even among households u.s. corporations from 1987-93. We ious degrees of internal production, with similar socioeconomic charac­ find that the average value of a firm and reveal a more pronounced lag teristics. Within the context of the with international operations is 2.2 structure for claims under partial life-cycle hypothesis, families with percent higher than that of compara­ riskbearing than under full self-insur­ identical lifetime resources might ble domestic single-activity firms, ance or market insurance. These pre­ choose to accumulate different levels while the average value of a firm dictions are generated by a funda­ of wealth for a variety of reasons, with activities in multiple industrial mental diseconomy of scale that including variation in time preference segments is 5.4 percent lower than a offsets the standard scale economy rates, risk tolerance, exposure to un­ portfolio of comparable focused do­ associated with risk-pooling. The certainty, relative tastes for work and mestic firms in similar activities. More tradeoff facing a group in its make­ leisure at advanced ages, income re­ importantly, we demonstrate that fail­ or-buy decision is that self-insurance placement rates, and so forth. These ure to control simultaneously for rewards self-protection but forgoes factors can be divided into a small both dimensions of diversification the pooling of risk with members number of classes, each with a dis­ results in overestimation of the neg­ outside the group. tinctive implication for the relation­ ative value impact of industrial diver­ ship between accumulated wealth sification, but has little impact on The Surprising Symmetry and the shape of the consumption estimates of the positive value impact of Gross Job Flows profile. By empirically examining this of geographic diversification. Christopher L. Foote relation, one can test for the presence NBER Working Paper No. 6226 or absence of particular explanations Consumption Versus October 1997 for differences in wealth. Using the Production of Insurance JEL No. E32 Panel Study of Income Dynamics and Tomas Philipson and Labor Studies and Monetary Economics the Consumer Expenditure Survey, George Zanjani we find very little support for life­ A large literature attempts to ex­ NBER Working Paper No. 6225 cycle models that rely on these fac­ plain the asymmetric behavior of job October 1997 tors to explain variation in wealth. destruction and job creation over the JEL No. 11 However, the data are consistent with business cycle. This paper contends Health Care "rule of thumb" or "mental account­ that much of this asymmetry is spuri­ ing" theories of wealth accumulation. Many forms of insurance are pro­ ous. Analyzing gross flows in relation duced by groups themselves rather to the net flow virtually eliminates than purchased in the market. For Evaluating Density cyclical asymmetry in annual data Forecasts of Inflation: example, coverage for workers' com­ and substantially reduces it in quar­ pensation provided by employers The Survey of terly data. To the extent that gross Professional Forecasters often is produced by the employer, flows are symmetric, there is a fun­ in the sense that the employer bears Francis X. Diebold, Anthony S. damental identification problem in Tay, and Kenneth F. wallis some or all of the financial risk asso­ moving between gross flows and the NBER Working Paper No. 6228 ciated with the insurance. We gener­ net flow that is reminiscent of the October 1997 alize the theory of insurance to ana­ earlier empirical literature on sectoral JEL Nos. E3, C1 lyze what factors determine whether shifts. Economic Fluctuations and Growth groups produce insurance internally by self-insuring or instead consume Since 1968, the Survey of Profes­ it by purchasing coverage in the mar­ What Accounts for the sional Forecasters has asked respon­ ket. The theory makes cross-sectional Variation in Retirement dents to provide a complete proba­ predictions on which firms will Wealth Among u.S. bility distribution of expected future choose to produce insurance, as well Households? inflation. We evaluate the adequacy B. Douglas Bernheim, Jonathan as on how prices and loss experience of those density forecasts using the Skinner, and Steven Weinberg will vary with the production deci­ framework of Diebold, Gunther, and NBER Working Paper No. 6227 sion. The theory also predicts which Tay [1997]. The analysis reveals sev­ October 1997 lines of insurance are likely to be eral interesting features, and several JEL Nos. E2, D1 associated with internal production deficiencies of the density forecasts Aging and Public Economics and which lines will have coverage in relation to realized inflation. The provided entirely by the market. Household survey data consis­ probability of a large negative infla­ Furthermore, we analyze the time- tently depict large variations in saving tion shock generally is overestimated,

46. NBER Reporter Winter 1997/8------and in more recent years the proba­ crises constituted the earliest per­ Trend and cycles are interrelated bility of a large shock of either sign is ceived reasons for economic instabil­ and have common causes. Periods overestimated. Inflation surprises are ity. Classical literature on the subject, and countries with high growth had serially correlated, although agents developed in the second half of the low instability. eventually adapt. Expectations of low nineteenth and first half of the twen­ Certain variables have long been inflation are associated with reduced tieth century favored for the most critically important in business cycles, uncertainty. The results suggest sev­ part the concept of self-sustaining or as shown by historical studies within eral promising directions for future endogenous fluctuations, but recent and across countries: profits, invest­ research. models again stress outside factors ment, interest rates, money, and and exogenous random shocks. credit. Leads and lags, nonlinearities The Value of Children and In an ideal world under the and asymmetries, also had demon­ Immigrants in a Pay-As­ assumptions of perfect competition, strably eminent roles, which they You-Go Pension System: flexible prices, national expectations, retain. Multiple-shock models are A Proposal for a Partial and money neutrality, there is room superior to Single-shock models. Transition to a Funded for real business cycles because of Finally, recessions have high social System shocks to technology and possibly costs in terms of unemployment and Hans-Werner Sinn also shocks to tastes, relative prices, depressed growth. Expansions can NBER Working Paper No. 6229 and fiscal variables. In the real world, also be costly by causing imbalances October 1997 however, there is ample evidence and excesses. Structural and policy JEL Nos. H55, J6 that many sticky prices and wages problems may seem to be separable Public Economics coexist with many flexible prices and from these cyclical problems but wages. Movements in levels of prices often are not. I show that the net fiscal external­ can be stabilizing, even while move­ ity created by an additional member ments in expected changes of prices of a pay-as-you-go pension system are destabilizing. Cyclical movements that is endowed with individual Trade versus Investment in nominal aggregates point to the accounts equals the gross contribu­ Liberalization role of money. The premise of pas­ tions of this member. In Germany, James R. Markusen sive money clashes with the popular this is an amount of about DM NBER Working Paper No. 6231 view that monetary policy is highly 175,000. I use this information to October 1997 important. The lesson of recent his­ design a hybrid funded system that JEL Nos. F12, F23 tory is that monetary factors influence avoids this externality and improves International Trade and Investment the course of economic activity along the public pension system under with real and expectational variables. Despite several theoretical contri­ equity and efficiency considerations. There are incentives for learning butions and considerable informal Business Cycles Observed how to avoid biased forecasts, but empirical evidence to the contrary, and Assessed: Why and data and models are not good enough the notion that trade and investment How They Matter to assure that rational expectations are substitutes persists in trade pol­ are achieved and maintained. Indi­ icy analysis. This paper considers the Victor Zarnowitz viduals' rational plans are not neces­ liberalization of commodity trade ver­ NBER Working Paper No. 6230 sarily collectively consistent. Unstable sus liberalization that allows direct October 1997 and adverse conditions can exist investment versus the two together. JEL Nos. E32, E37 even under the extreme assumptions For an economy with a relative scarcity Economic Fluctuations and Growth of perfect flexibility of all prices and of skilled labor, trade and investment Business cycles on the whole are perfect foresight. liberalization are quite different, and well defined and described, yet they Highly aggregative models, while the two together in a sense are com­ have no generally accepted explana­ attractively simple and articulated, plements. The intuition may be that tion. Whether in spite of, or because are not able to deal with the diver­ direct investment provides such a of, proliferation of abstract and frag­ sity of causes and consequences of country with crucial inputs (knowl­ mented models in this field, the the­ business cycles. Disaggregation is edge-intensive producer services) ory of business cycles is long on essential for theories that emphasize without which the country cannot questions but short on answers. endogenous processes of spending, effectively exploit its abundant fac­ Natural disasters and then financial saving, borrowing, and investing. tors in certain industries.

------NBER Reporter Winter 1997/8 47. Liberalization and January 1, 1999 choices equal closing operating losses. There is no evi­ Incentives for Labor market exchange rates on December dence that risk management and Migration: Theory with 31, 1998. Given that legal constraint, cash holdings are substitutes. Applications to NAFTA there still exist several strategies for James R. Markusen and choosing the relative prices of EMU Firm-level Evidence on Stephen Zahniser member currencies against the Euro. Productivity Differentials, NBER Working Paper No. 6232 Unfortunately, most of these have po­ Turnover, and Exports in October 1997 tentially damaging side effects. One Taiwanese Manufacturing JEL Nos. F22, F15 approach, based on official Stage 2 Bee Yan Aw, Xiaomin Chen, International Trade and Investment offers of contingent Euro-forward and MarkJ. Roberts contracts with value dates at the start NBER Working Paper No. 6235 One of the motivations for NAYfA of Stage 3, allows a highly credible October 1997 from the U.S. point of view was to pre announcement of the bilateral JEL Nos. 012, D24 reduce the incentives for Mexican currency conversion factors to be set Productivity migration into the United States. at the start of EMU. That approach Unskilled rural males are a primary The manufacturing sector in Taiwan assumes, however, that no prospec­ source of illegal immigration and are has a market structure with large tive EMU members can withdraw also Mexico's relatively abundant fac­ numbers of small finns, a heavy fo­ between their selection in May 1998 tor. This group therefore should be cus on less capital-intensive indus­ and the start of Stage 3. made better off by trade and invest­ tries, and a dense network of firms that specialize in subcontracting and ment liberalization, according to the The Determinants and traditional Heckscher-Ohlin model. trading services. These features may Implications of Corporate lower the start-up costs of new man­ Existing evidence, along with the Cash Holdings best guesses of many experts in the ufacturing firms. Recent theoretical Tim Opler, Lee Pinkowitz, Rene models of market evolution empha­ area, suggest that NAYfA is unlikely StuIz, and Rohan Williamson to have a significant positive impact size that low sunk costs of entry and NBER Working Paper No. 6234 on this group, at least not within the exit speed firm turnover by facilitat­ October 1997 ing entry and increasing the pressure time frame of several decades. We Corporate Finance draw on a number of recent theoret­ on inefficient firms to exit. As a ical contributions in order to offer We examine the determinants and result, low-cost entry and exit may reasons why NAFTA may not raise implications of holdings of cash and contribute to aggregate improve­ the wages of unskilled Mexican marketable securities by publicly ments in productivity by facilitating workers. traded U.S. firms in the 1971-94 the rapid transfer of resources from period. Firms with strong growth less to more efficient producers with­ A Strategy for Launching opportunities and riskier cash flows in an industry. Using comprehensive the Euro hold relatively high ratios of cash to firm-level panel data from the Taiwa­ Maurice Obstfeld total assets. Firrns that have the great­ nese Census of Manufactures for 1981, NBER Working Paper No. 6233 est access to the capital markets (for 1986, and 1991, we measure differ­ October 1997 example, large firms and those with ences in total factor productivity JEL Nos. F31, F33, E42 credit ratings) tend to hold lower among entering, exiting, and contin::' International Finance and ratios of cash to total assets. These uing firms, and quantify the contri­ Macroeconomics results are consistent with the view bution of firm turnover to improve­ that firms hold liquid assets in order ments in industry productivity. This paper analyzes the con­ to insure that they will be able to We find significant differences in straints that European Union (ED) law keep investing when cash flow is too productivity across manufacturing places on the January 1, 1999 choices low relative to planned investment firms that are reflected in turnover of irrevocably fixed conversion rates and when outside funds are expen­ patterns in both the domestic and between the Euro and the currencies sive. The short-run impact of excess export market. Cohorts of new firms of European Monetary Union (EMU) cash on capital expenditures, acqui­ have lower average productivity than member states. Current EU legisla­ sition spending, and payouts to share­ incumbents, but are themselves a tion, notably the Maastricht treaty, holders is small. The main reason that heterogeneous group. The more pro­ requires that the bilateral currency firms experience large changes in ductive members of the group, on conversion factors implie? by the excess cash is the Occurrence of average, survive and in many cases

48. NBER Reporter Winter 1997/8------their productivity converges to the comparisons, and we analyze both the view that fiscal institutions have level of incumbents. Exiting firms are macro- and microeconomic aspects real effects on fiscal policy outcomes. less productive than survivors. of financial intermediation. Exporters, including firms that exited Is Bank-Centered Corporate the export market recently, are more State Fiscal Institutions Governance Worth It? A productive than nonexporters. These and the u.S. Municipal Cross-Sectional Analysis patterns are consistent with the view Bond Market of the Performance of that both the domestic and the James M. Poterba and Japanese Firms during the export market sort out high-produc­ Kim S. Rueben Asset Price Deflation tivity from low-productivity firms and NBER Working Paper No. 6237 Jun-Koo Kang and Rene M. Stu1z that the export market is a tougher October 1997 NBER Working Paper No. 6238 screen. The productivity differential JEL Nos. H61, H74 October 1997 between entering and exiting firms is Public Economics Corporate Finance an important source of industry-level This paper presents new evidence productivity growth in Taiwanese We examine the determinants of on the effect of state fiscal institu­ manufacturing, accounting for as firm stock-price performance from tions, particularly balanced-budget much as half of industry improve­ 1990 to 1993 in Japan. During that rules a'nd restrictions on state debt period, the typical firm on the Tokyo ment in some industries and time issuance, on the yields on state gen­ periods. Stock Exchange lost more than half eral obligation bonds. We analyze its value, and banks experienced information from the Chubb Relative severe adverse shocks. We show that Finance and Development Value Survey, which contains relative in an Emerging Market: firms whose debt had a higher frac­ tax-exempt yields on the bonds tion of bank loans in 1989 performed Argentina in the issued by different states over the Interwar Period worse from 1990 to 1993. This effect period 1973-96. We find that states is statistically as well as economically Gerardo della Paolera and with tighter anti-deficit rules, and Alan M. Taylor significant and holds when we con­ more restrictive provisions on the trol for a variety of variables that af­ NBER Working Paper No. 6236 authority of state legislatures to issue October 1997 fect performance during this period. debt, pay lower interest rates on their We find that firms that were more International Finance and bonds. The interest rate differential Macroeconomics bank-dependent also invested less between a state with a very strict anti­ during this period than other firms. The long-run economic perform­ deficit fiscal constitution, and one This points to an adverse effect of ance of Argentina since World War I with a lax constitution, is between 15 bank-centered corporate governance, has been relatively disappointing un­ and 20 basis pOints. States with bind­ namely that firms suffer when their til recently. Yet in the interwar ing revenue limitation measures tend banks are experiencing difficulties. period, signs of future retardation to face higher borrowing rates by and recurring crises were not obvious. approximately the same amount, Higher Tariffs, Lower It is often claimed that the remarkably while states with expenditure limits Revenues? Analyzing the rapid growth of domestic financial face lower borrowing costs. Thus fis­ Fiscal Aspects of the "Great markets was an unmitigated success. cal restraints that control expendi­ Tariff Debate of 1888" In conventional models, such "finan­ tures are viewed favorably by bond Douglas A. Irwin cial deepening" would help to accel­ market participants, while those that NBER Working Paper No. 6239 erate development, especially in an restrict taxes, and therefore might October 1997 industrializing economy such as interfere with the state's ability to JEL Nos. N71, H6, F13 Argentina's. Yet the promise of this repay interest, result in higher bor­ Development of the American Economy trend was unfulfilled: first the out­ rowing costs. The effect of strict fiscal and International Trade and Investment break of World War I and then the institutions is particularly evident Great Depression proved setbacks when a state's economy is weak. After the Civil War, Congress justi­ for the fledgling financial system, and These results provide important evi­ fied high import tariffs (relative to a long-run deterioration set in after dence that bond market participants their prewar levels) as necessary in 1940. In this paper we trace the consider fiscal institutions in assess­ order to raise sufficient revenue to course of financial development ing the risk characteristics of tax­ payoff the public debt. By the early using historical and international exempt bonds, and further support 1880s, the federal government was

------NBER Reporter Winter 1997/8 49. running large and seemingly intract­ that this median household will still Impact of a Managed able fiscal surpluses: revenues ex­ need to save 16 percent of annual Behavioral Health Care ceeded expenditures (including debt earnings to preserve pre-retirement Carve-Out: A Case Study service and repurchases) by over 40 consumption. For retirement at age of One HMO percent during that decade. The 65, assets are expected to be about Anne E. Brisson, Richard G. political parties proposed alternative $420,000 and required additional sav­ Frank, Elizabeth S. Notman, plans to deal with the surplus: the ing totalling 7 percent of earnings per and Julie A. Gazmararian Democrats proposed a tariff reduc­ year. These summary statistics con­ NBER Working Paper No. 6242 tion to reduce customs revenue, the ceal extraordinary heterogeneity in October 1997 Republicans offered higher tariffs to both assets and saving needs in the JEL No. III reduce imports and customs revenue. older population. Older high wealth Health Care This paper examines this debate and households have 45 times more We examine a case study of a attempts to determine the revenue assets than the poorest decile, and carve-out for mental health (MH) effects of the proposed tariff changes. this disparity increases with age. and substance abuse (SA) services The results indicate that, given the There are also large differences in between a local plan of a national height of the tariff and the price elas­ prescribed saving targets, ranging HMO (N=120,213) and a local man­ ticity of import demand during from 38 percent of annual earnings u.s. aged behavioral health care vendor the 1880s, the actual tariff was below for those in the lowest wealth decile (MBHC). This is one of the first stud­ the maximum revenue rate, and to negative rates for the wealthiest ies which estimates the impact of an therefore a tariff reduction would decile. HMO carve-out on costs and patterns have reduced customs revenue. of MH/SA care. Three years of insur­ ance claims data 0993-5) were used Projected Retirement for the analyses, with a new carve­ Wealth and Savings Foreign Direct Investment out contract implemented in May Adequacy in the Health as a Catalyst for Industrial 1994. The new carve-out arrange­ and Retirement Study Development ment included a new vendor, a James F. Moore and James R. Markusen and change in the organizational structure Olivia S. Mitchell Anthony J. Venables of clinical services, and increased NBER Working Paper No. 6240 NBER Working Paper No. 6241 financial risk to the vendor for inpa­ October 1997 October 1997 tient care. Descriptive and empirical JEL Nos. J14, G23, H55 International Trade and Investment analyses are reported on a continu­ Aging How does a foreign direct invest­ ously enrolled population (N=49,529). Low saving rates raise questions ment (FDI) project affect local firms The analyses show that the new about Americans' ability to maintain in the same industry? Competition in carve-out arrangements had a signif­ consumption levels in old age. Using the product and factor markets tends icant impact on spending and utiliza­ the Health and Retirement Study, this to reduce profits of local firms, but tion of services. Enrollees were 20 paper explores asset holdings among linkage effects to supplier industries percent less likely to use MH/SA a nationally representative sample of may reduce input costs and raise services after the implementation of people on the verge of retirement. profits. This paper develops an ana­ the new carve-out, and inpatient Making reasonable projections about lytical framework for assessing these MH/SA utilization dropped 50 per­ asset growth, we assess how much effects. We establish circumstances in cent under the new carve-out. Over­ more people would need to save in which FDI is complementary to local all, MH/SA spending per enrollee order to preserve consumption levels industry, and show how FDI may dropped from approximately $4.90 after retirement. We find that the lead to the establishment of local per month to $2.20 per month. Out­ median older household has current industrial sectors. These sectors may patient MH/SA spending per user wealth of approximately $325,000 grow to the point where local pro­ dropped 35 percent after the imple­ including pensions, Social Security, duction overtakes and forces out FDI mentation of the new carve-out. housing, and other financial wealth, plants. Our results are consistent with Further research should be conducted an amount projected to grow to about the experience of a number of indus­ to evaluate the impact on access and $380,000 by retirement at age 62. trial sectors in the newly industrializ­ quality of care, given the substantial Nevertheless, our model suggests ing countries. decrease in utilization and spending.

50. NBER Reporter Winter 1997/8------International Competition learn the same lesson from the lower for service stations and new and Exchange Rate Shocks: Kuznets Curve debate? No economist car dealers. Consumers' behavior is A Cross-Country Industry should expect an "unconditional" consistent with a learning model in Analysis of Stock Returns Kuznets Curve to emerge from the which they have diffuse initial priors John M. Griffin and growth experience of all countries regarding the probability that they Rene M. Stolz and at all times. The forces of the fail at individual firms, and "Bayesian NBER Working Paper No. 6243 industrial revolution thought to have update" using two to three inspection October 1997 an impact on inequality can be offset outcomes at each firm. Asset Pricing or reinforced by demography, skill supply, and globalization. This paper Explaining National It is widely accepted that, for assesses the role of globalization and Differences in the Size some industries, competition across demography via mass migrations. and Industry Distribution countries is economically important, Second, why has it taken econo­ of Employment and competition is affected strongly mists so long to learn that demogra­ StevenJ. Davis and by exchange rate changes. We ex­ phy influences growth? When treated Magnus Henrekson plore the validity of this view using properly, demography can be shown NBER Working Paper No. 6246 weekly stock return data on 320 to have a significant impact on the October 1997 industry pairs in six countries from per capita growth of GDP. I seek to JEL Nos. L52, ]21, H30 1975 to 1997. We find that common answer these two questions by look­ Labor Studies shocks to industries across countries ing at inequality and growth experi­ are more important than competitive What factors determine national ence in the Old world, the New shocks. Weekly exchange rate shocks differences in the size and industry world, and Asia over the last century explain almost none of the relative distribution of employment? We and a half. performance of industries. Based on stress the role of the economic policy returns measured over longer hori­ Consumer Beliefs and environment as determined by busi­ zons, the importance of exchange Buyer and Seller Behavior ness taxes, employment security laws, rate shocks increases slightly, and the in the Vehicle Inspection credit market regulations, the na­ importance of common shocks to Market tional pension system, wage-setting industries increases more substan­ institutions, and the size of the pub­ Thomas N. Hubbard tially. Both industry and exchange lic sector. We characterize these as­ NBER Working Paper No. 6245 rate shocks are more important for pects of the policy environment in October 1997 industries that produce goods traded Sweden prior to 1990-1 and compare ]EL Nos. 114, LIS, D83 internationally, but the importance of them to the situation in other Euro­ Industrial Organization these shocks is economically small pean countries and the United States. for those industries as well. The California vehicle emission in­ Our characterization and interna­ spection market offers a rare oppor­ tional comparisons show that Swed­ Growth, Distribution and tunity to examine how incentives ish policies strongly disfavored less Demography: Some operate in "diagnosis-cure" markets. I capital-intensive firms, smaller firms, Lessons from History investigate why sellers help vehicles entry by new firms, and individual pass inspections, focusing on multi­ and family ownership of business. Jeffrey G. Williamson period mechanisms such as those in We also compile evidence that NBER Working Paper No. 6244 reputation models. I show that the these policies affect outcomes. Tak­ October 1997 demand faced by individual firms is ing the U.S. industry distribution as a JEL Nos. D3, Fl, ]1, N3, 01 sensitive to inspection outcomes. benchmark that reflects a compara­ Development of the American Economy Consumers are 30 percent more tively neutral set of policies and insti­ If we have learned anything from likely to return to a firm at which tutions, Sweden's employment distri­ the recent outpouring of empirical they previously passed inspection bution in the mid-1980s is sharply growth equations, it is that life is than one at which they previously tilted away from low-wage industries far too complex to expect "uncon­ failed. If, over the long run, an inde­ and industries with greater employ­ ditional" convergence among all pendent garage fails one additional ment shares for smaller firms and countries and at all times. This fact vehicle per month, this decreases establishments. Compared to other motivates two questions. First, why demand by 5.6 inspections per European countries, Sweden has an has it taken economists so long to month on average. This figure is unusually high share of employment

------NBER Reporter Winter 1997/8 51. in large firms. Furthermore, the Swed­ model suggests a means of measur­ economy's long-run output gain and ish rate of self-employment in the ing changes in the cost of undertak­ harms initial older generations. 1970s and 1980s is the lowest among ing disguised capital flows, based on Insulating these older generations all OECD countries. the past history of differentials be­ through transition relief further The institutional and policy factors tween external interest rates (adjusted reduces the long-run gains from tax emphasized by Our study differ for exchange rate changes) and reform. Switching to a proportional greatly across countries. This fact domestic ceiling interest rates, pro­ income tax without deductions and suggests that our approach can be vided that the authorities' foreign exemptions hurts current and future fruitfully applied to other studies of exchange market activities are incor­ low lifetime earners, but helps every­ national differences in industry and porated into the analysis. Parameter one else. It also raises long-run out­ size structures and their evolution estimates for Korea, Mexico, and the put by over 5 percent. The X tax over time. As an example, the tax Philippines indicate that the real cost makes everyone better off in the long reform wave of the 1980s-which of undertaking disguised capital run, and raises long-run output by 7.5 largely evened out cross-country dif­ flows declined by nearly 70 percent percent, but harms initial older gen­ ferences in corporate taxation among on average between the early 1970s erations who bear its implicit wealth OECD countries - offers some basis and the late 1980s. tax. for projecting a movement towards greater Similarity among wealthy Simulating u.S. Tax Reform Persistence of Medicare countries in the size and industry dis­ David Altig, AlanJ. Auerbach, Expenditures Among tribution of employment. Laurence J. Kotlikoff, Kent A. Elderly Beneficiaries Smetters, and Jan Walliser Alan M. Garber, Thomas E. Capital Mobility and NBER Working Paper No. 6248 MaCurdy, and Mark C. McClellan Exchange Market October 1997 NBER Working Paper No. 6249 Intervention in JEL Nos. H20, C68 October 1997 Developing Countries Public Economics JEL Nos. Ill, J14, 118, H51 Michael P. Dooley, We use a new large-scale dynamic Aging and Health Care DonaldJ. Mathieson, and simulation model to compare the The highly uneven distribution of Liliana Rojas-Suarez equity, efficiency, and macroeco­ Medicare payments among elderly NBER Working Paper No. 6247 nomic effects of five alternatives to beneficiaries, combined with the pre­ October 1997 the current U.S. federal income tax: dictability of some of the expendi­ JEL Nos. F32, F36, F4 a proportional income tax; a pro­ tures, poses several challenges to the International Finance and portional consumption tax; a flat Medicare program. We present infor­ Macroeconomics tax; a flat tax with transition relief; mation about the distribution of This paper develops a new tech­ and a progressive variant of the flat Medicare expenditures among bene­ nique for measuring changes in the tax called the "X tax." Our model ficiaries in specific years, accompa­ degree of capital mobility confronting incorporates intragenerational het­ nied by new evidence on the extent a developing country that has restric­ erogeneity and kinked budget con­ to which Medicare payments for the tions on capital flows and official straints. We predict major macroeco­ care of individual beneficiaries per­ ceilings on domestic interest rates. nomic gains (including an 11 percent sist over long time periods. Our Because such official controls rule increase in long-run output) from analysis is based on a longitudinal out the use of traditional interest rate replacing the federal tax system with population of Medicare enrollees parity conditions to measure changes a proportional consumption tax. during the years from 1987 to 1995. in· the degree of capital mobility, the Future middle- and upper-income We find that high-cost users analysis first examines an intertem­ classes gain from this policy, but ini­ accounted for a disproportionate poral model of an open economy. tial older generations are hurt by the share of the growth of Medicare Part This model describes the linkages policy's implicit capital levy. Poor A (hospital) payments during this between the cost of undertaking dis­ members of current and future gen­ period, but that an increase in the guised capital flows, the current erations also lose. number of beneficiaries using cov­ account, capital controls, domestic The flat tax, which adds a stan­ ered services was largely responsible and external financial market condi­ dard deduction to the consumption for the growth of Medicare Part B tions, and the authorities' foreign tax, makes all members of future payments. Few beneficiaries are in exchange market interventions. The generations better off, but halves the the highest-cost categories for multi-

52. NBER Reporter Winter 1997/8------pIe years; the high mortality rates of given recursively and may be used to I Just Ran Four Million individuals who use medical services quantify the "degree" of market . Regressions heavily, whether the expenditures incompleteness. To investigate the xavier X. Sala-I-Martin occur in one year or repeatedly, lim­ practical significance of these E-arbi- \ NBER Working Paper No. 6252 its the extent of expenditure persis­ trage strategies, we consider several November 1997 tence. Even among survivors, it is numerical examples, including path­ JEL Nos. 051, 052, 053 unusual to remain in the highest-cost dependent options and options on Economic Fluctuations and Growth categories for multiple years. Never­ assets with stochastic volatility and theless, individuals with high expen­ jumps. In this paper, I try to move away ditures in one year are likely to have from the extreme bounds method of higher-than-average expenditures in identifying "robust" empirical rela­ other years, and expenditures are tions in the economic growth litera­ highly skewed even over a period of ture. Instead of analyzing the ex­ nine years. Any policy to reform treme bounds of the estimates of the coefficient of a particular variable, I Medicare will need to accommodate On the Superiority of the persistence of expenditures in analyze the entire distribution. My Corrective Taxes to claim in this paper is that, if we do order to provide adequate coverage Quantity Regulation for all beneficiaries. this, the picture emerging from the Louis Kaplow and Steven Shaven empirical growth literature is not the NBER Working Paper No. 6251 pessimistic "nothing is robust" that November 1997 we get with the extreme bound JEL Nos. H23, D62, K32, L51 analysis. Instead, we find that a sub­ Public Economics stantial number of variables can be The traditional view of economists found to be strongly related to growth. Pricing and Hedging has been that corrective taxes are Derivative Securities in superior to direct regulation of harm­ An Incomplete Markets: ful externalities when the state's Technology and E-Arbitrage Approach information about control costs is Bilateral Trade Dimitris Bertsimas, Leonid incomplete. In recent years, however, Jonathan Eaton and Kogan, and Andrew W. Lo Samuel Kortum many economists seem to have NBER Working Paper No. 6250 NBER Working Paper No. 6253 adopted the view that either correc­ November 1997 November 1997 tive taxes or quantity regulation JEL No. G13 JEL Nos. Fll, F17, 033 could be superior to the other. One Asset Pricing International Trade and Investment argument for this view, identified Given a European derivative secu­ with Weitzman [1974], holds only if We develop a Ricardian model to rity with an arbitrary payoff function the state is constrained to use a fixed explore the role of trade in spread­ and a corresponding set of underly­ tax rate (a linear tax schedule) even ing the benefits of innovation. The ing securities on which the derivative when harm is nonlinear. Corrective theory delivers an equation for bi­ security is based, we solve the dy­ taxes are indeed superior to quantity lateral trade that, on its surface, re­ namic replication problem: find a regulation if-as seems more plausi­ sembles a gravity specification, but self-financing dynamic portfolio strat­ ble -the state can impose a non­ identifies underlying parameters of egy-involving only the underlying linear tax equal to the schedule of technology. We estimate the equation securities-that most closely approx­ harm, or can adjust the tax rate upon using trade in manufactures among imates the payoff function at matu­ learning that it diverges from mar­ the OECD. The parameter estimates rity. By applying stochastic dynamic ginal harm. Another argument, asso­ allow us to simulate the model to programming to the minimization of ciated with Baumol and Oates [1988], investigate the role of trade in spread­ a mean-squared-error loss function is that quantity regulation gains ing the benefits of innovation and to under Markov state dynamics, we appeal when the state is uncertain examine the effects of lower trade derive recursive expressions for the about the harm caused by an exter­ barriers. Typically, foreigners benefit optimal replication strategy that are nality. In this case, however, a cor­ by only a tenth as much as the inno­ readily implemented in practice. The rective tax schedule (equal to the vating country, but in some cases the approximation error or "E" of the expected harm schedule) is superior benefits to close neighbors approach optimal replication strategy is also to quantity regulation. those of the innovator.

------NBER Reporter Winter 1997/8 53. Monetary Policy Rules foreign direct investment. The mea­ atic level. This casts doubt on the in Practice: Some sure of uncertainty is constructed view that large participants have bet­ International Evidence based on unpublished individual sur­ ter information about the future Richard Clarida, Jordi Gall, vey responses about levels of cor­ movement of exchange rates. It fur­ and Mark Gertler ruption in host countries. The result ther strengthens the case that the NBER Working Paper No. 6254 is striking: the effect is negative, sta­ large players trade on noise rather November 1997 tistically significant, and quantita­ than on information. JEL Nos. E58, F41 tively large. An increase in the uncer­ Economic Fluctuations and Growth, tainty level from that of Singapore to Econometric Models of International Finance and Macro­ that of Mexico, at the average level Limit-Order Executions economics, and Monetary Economics of corruption in the sample, is equiv­ Andrew W. 1.0, A. Craig MacKinlay, andJune Zhang We estimate monetary policy reac­ alent to raising the tax rate on multi­ NBER Working Paper No. 6257 tion functions for two sets of coun­ national firms by 32 percentage November 1997 tries: the G3 (Germany, Japan, and points. Hence, the second-moment JEL No. G23 the United States) and the E3 (United (uncertainty) effect can and does Asset Pricing Kingdom, France, and Italy). We find have first-order importance. that since 1979, each of the G3 cen­ Limit orders incur no price impact, tral banks has pursued an implicit The Big Players in the however their execution time is form of inflation targeting, which Foreign Exchange Market: uncertain. We develop several econo­ may account for the broad success of Do They Trade on metric models of limit-order execu­ monetary policy in those countries Information or Noise? tion times using survival analysis, and over this time period. The evidence Shang-Jin Wei andJungshik Kim then estimate them with actual limit­ also suggests that these central banks NBER Working Paper No. 6256 order data. We estimate models for have been forward looking: they re­ November 1997 time-to-first-fill and time-to-comple­ spond to anticipated inflation as op­ JEL No. F31 tion, and for limit-sells and limit­ posed to lagged inflation. As for the International Finance and buys, and we incorporate the effects E3, even prior to the emergence of Macroeconomics of explanatory variables, such as the the "hard ERM," their central banks We ask whether there is private limit price, the limit size, the bid/offer were influenced heavily by German information in the foreign exchange spread, and market volatility. We find monetary policy. Further, using the market, and whether speculation that execution times are very sensi­ Bundesbank's policy rule as a bench­ reduces or exacerbates volatility. We tive to limit price and several other mark, we find that at the time of the use a recent dataset on foreign cur­ explanatory variables, but are not EMS collapse, interest rates in each rency positions of large market par­ sensitive to limit size. We also show of the E3 countries were much higher ticipants that includes their positions that hypothetical limit-order execu­ than domestic macroeconomic con­ on options and other derivatives. tions, constructed either theoretically ditions warranted. Taken together, This is the first dataset that describes from first-passage times or empiri­ the results support the view that comprehensive currency positions of cally from transactions data, are very. some form of inflation targeting market participants. We find first, that poor proxies for actual limit-order under certain circumstances may be not only the absolute value of the executions. superior to fixing exchange rates, as options position, but also that of Generating Real Persistent a means of gaining a nominal anchor spot, forward, and futures positions for monetary policy. Effects of Monetary Shocks: of large participants "Granger-causes" How Much Nominal Why is Corruption So Much exchange rate volatility. This suggests Rigidity Do We Really Need? that the large participants' currency More Taxing Than Tax? Olivier Jeanne speculation does not stabilize ex­ Arbitrariness Kills NBER Working Paper No. 6258 change rate volatility. Second, our Shang-Jin Wei November 1997 regression analyses do not find any NBER Working Paper No. 6255 JEL Nos. E1, E3 positive association between large November 1997 Monetary Economics participants' position in a foreign cur­ JEL Nos. F20, F23, H3, 037 rency and its subsequent apprecia­ This paper asks whether money International Trade and Investment tion. A non-parametric approach can generate persistent economic This paper examines the effect of finds some weak support for a posi­ fluctuations in dynamic general equi­ corruption-induced uncertainty on tive association but not on a system- librium models of the business cycle.

54. NBER Reporter Winter 1997/8------I show that a small nominal friction Congress is increasingly unable to its specific role. The studies generally in the goods market can make the increase general taxes to pay for hypothesize that discretion either response of output to monetary social programs. Thus we fund these serves as the means by which chang­ shocks large and persistent if it is programs from taxes on specific sec­ ing broad social norms against crime amplified by real wage rigidity in the tors of the economy. In this paper, I causes changes in sentencing pat­ labor market. I also argue that, given consider the Congressional legisla­ terns; or it serves as the means by the level of real wage rigidity that is tion that established a program for all which internal social nOrms of the observed in developed countries, a public schools and libraries in the criminal justice system prevent the small degree of nominal stickiness United States to receive subsidized implementation of formal changes in might be sufficient for money to pro­ service to the Internet. The cost of laws. duce economic fluctuations as per­ the program currendy is estimated to We reject both of these hypothe­ sistent as those observed in the data. be $2.25 billion per year. Congress ses, using data on the sentencing of passed legislation that directed all California prisoners before and after On the Disutility users of interstate telephone service the passage of Proposition 8, which and Discounting of to pay for the program. provided for sentence enhancements Imprisonment and the Using analytical methods from for those convicted of certain "seri­ Theory of Deterrence public finance, I calculate the effi­ ous" crimes with "qualifying" crimi­ A. Mitchell Polinsky and ciency cost to the economy of the nal histories. We find that an increase Steven Shavell increased taxation of interstate tele­ in the statutory sentence for a given NBER Working Paper No. 6259 phone services to fund the Internet crime cannot only increase sentence November 1997 access discounts. I estimate the cost length for those charged with the JEL No. K14 to the economy of raising the $2.25 crime, but also for those charged Law and Economics billion per year to be at least $2.36 with factually "similar" crimes: a "sim­ This article studies the implica­ billion (in addition to the $2.25 bil­ ilar" crime is defined as one that has tions for the theory of deterrence of: lion of tax revenue), or the efficiency legal elements in common with the 1) the manner in which individuals' loss to the economy for every $1 given crime. These spillovers are disutility from imprisonment varies raised to pay for the Internet access consistent with neither broad social with the length of the imprisonment discounts is an additional $1.05 to norms nor internal social norms. term; and 2) discounting the future $1.25 beyond the money raised for Thus, we conclude that discretion disutility and future public costs of the Internet discounts. This cost to takes a less-well studied form, which imprisonment. We address two ques­ the economy is extraordinarily high we call "prosecutorial maximization:' tions: Is deterrence enhanced more compared to other taxes used by the by increasing the length of imprison­ Federal government to raise rev­ International Trade and ment terms, or by raising the likeli­ enues. I discuss an alternative Labor-Demand Elasticities hood of imposing imprisonment? method by which the FCC could Matthew J. Slaughter What is the optimal combination of have raised the revenue for the NBER Working Paper No. 6262 the severity and probability of impris­ Internet discounts which would have November 1997 onment sanctions? a near zero cost to the economy. JEL Nos. F1, J3 International Trade and Investment Taxation by The Role of Discretion in In this paper, I try to determine Telecommunications the Criminal Justice System whether international trade has been Regulation Daniel P. Kessler and increasing the own-price elasticity of Anne Morrison Piehl Jerry Hausman demand for U.S. labor in recent dec­ NBER Working Paper No. 6261 NBER Working Paper No. 6260 ades. There are three main results. November 1997 November 1997 First, from 1960 through 1990, JEL Nos. K41, D73 JEL Nos. H21, L51 demand for U.S. production labor Law and Economics Industrial Organization and became more elastic in manufactur­ Public Economics Although a substantial body of ing overall, and in five of eight indus­ Telecommunications regulation in research suggests that individuals' tries within manufacturing. Second, the United States is replete with a discretion in the criminal justice sys­ during this time the demand for U.S. system of subsidies and taxes. Be­ tem is important, there is disagree­ nonproduction labor did not become cause of budgetary spending limits, ment in the existing literature about more elastic in manufacturing over-

------NBER Reporter Winter 1997/8 55. all, or in any of the eight industries which includes corporate income Previous studies tend to find that within manufacturing. If anything, and corporate property taxes, and immigration has a weak negative demand seems to be growing less taxes on stock- and bondholders, effect on the employment and earn­ elastic over time. Third, the hypothe­ averages 42.1 percent of pretax prof­ ings of native-born workers. These sis that trade contributed to increased its. The average pretax rate of return studies generally overlook the effect of elasticities has mixed support, at best. for 1990-6 is 8.6 percent, and the immigration on an important sector For production labor, many trade average aftertax return is 5 percent. of the labor force, the self-employed. variables have the predicted effect for Although the accounting return to Anecdotal evidence suggests that specifications with only industry con­ corporate capital has been higher in immigrants, especially those from trols. But these predicted effects dis­ the mid-1990s than at any previous Asian countries, may displace black­ appear when time controls are point in the last two decades, the owned business owners. We use included as well. For nonproduction substantial volatility in the return Census of Population micro data to labor, things are somewhat better, but series makes it premature to con­ see whether black self-employment time continues to be a very strong clude that these years represent a levels are lower in labor markets with predictor of elasticity patterns. Thus departure from past experience. a higher share of immigrants. We the time series of labor-demand elas­ define labor markets as metropolitan ticities is explained largely by a resid­ Aggregate Investment areas (MAs), and we use the varia­ ual, time itself. This result parallels Ricardo J. Caballero tion across 94 MAs in the United the common finding in studies of ris­ NBER Working Paper No. 6264 States to examine the relationship ing wage inequality. Just as there November 1997 between black self-employment and appears to be a large unexplained JEL Nos. E2, D9 immigration in 1980 and 1990. To residual for changing factor prices control for permanent differences The 1990s have witnessed a re­ over time, there also appears to be a across MAs in other influences, we vival in economists' interest in and large unexplained residual for chang­ also estimate the effect of the change hope of explaining aggregate and ing factor demand elasticities over in immigration from 1980 to 1990 on microeconomic investment behavior. time. the change in black self-employment New theories, better econometric over this period. In general we find procedures, and more detailed panel The Rate of Return to that immigration has no effect, or datasets are behind this movement. Corporate Capital and only a small negative but statistically Much of the progress has occurred at Factor Shares: New insignificant effect, on black male or the level of microeconomic theories Estimates Using Revised female self-employment. Our find­ and evidence; however, progress in National Income Accounts ings are similar if we weight immi­ aggregation and general equilibrium and Capital Stock Data gration rates by the propensity of im­ aspects of the investment problem James M. Poterba migrant groups to be self-employed; also has been significant. The con­ NBER Working Paper No. 6263 if we limit our sample of immigrants cept of sunk costs is at the center of November 1997 to those from only Asian countries; modern theories. The implications of JEL Nos. D24, D33, G31 and if we try other alternative esti­ these costs for investment go well mation techniques and specifications. This paper presents new evidence beyond the neoclassical response to on the rate of return on tangible the irreversible-technological friction assets in the United States, incorpo­ they represent, for they also can lead Slowdowns and rating the recently-revised national to first-order inefficiencies when Meltdowns: Postwar accounts as well as new estimates of interacting with informational and Growth Evidence from the replacement cost of the repro­ contractual problems. 74 Countries ducible physical capital stock. The Dan Ben-David and David H. Papell pretax return on capital in the nonfi­ Does Immigration Hurt NBER Working Paper No. 6266 nancial corporate sector has averaged Mrican-American November 1997 8.5 percent over the 1959-96 period. Self-Employment? JEL Nos. C22, 01, OS, 047 The paper also presents new esti­ Robert W. Fairlie and mates of the total tax burden on non­ Bruce D. Meyer We propose an explicit test for financial corporate capital, which NBER Working Paper No. 6265 determining the significance and the averages 54.1 percent over this time November 1997 timing of slowdowns in economic period. For the 1990s, this tax rate, JEL Nos. F22, J15 growth during the postwar period.

56. NBER Reporter Winter 1997/8------We examine a large sample of coun­ Demographic Transitions Strategic Bidding in a tries (both industrialized and devel­ and Economic Miracles Multi-Unit Auction: An oping) and find that a majority­ in Emerging Asia Empirical Analysis of though not all-exhibit a significant David E. Bloom and Bids to Supply Electricity structural break in their postwar Jeffrey G. Williamson in England and Wales growth rates. In nearly all of these NBER Working Paper No. 6268 Catherine D. Wolfram cases, the break was followed by a November 1997 NBER Working Paper No. 6269 slowdown in growth. The breaks fall JEL Nos. J1, 01, 04, 053 November 1997 into two primary periods which JEL Nos. D44, L94 The demographic transition - a delineate countries by developmen­ change from high to low rates of This paper considers the bidding tal and regional characteristics as well mortality and fertility - has been behavior of participants in the daily as by the magnitude of the subse­ more dramatic in East Asia during auction for supplying electricity in quent slowdowns. We find that most this century than in any other region England and Wales. Every day, own­ industrialized countries experienced or historical period. By introducing ers of generating capacity submit postwar growth slowdowns in the demographic variables into an empir­ bids reflecting a price for power from early 1970s; the United States, Can­ ical model of economic growth, we their plants. The bid made by the last ada, and the United Kingdom did show that this transition has con­ plant that is used to meet electricity not; and developing countries (and in tributed substantially to East Asia's needs in a given time period is the particular, Latin American countries) so-called economic miracle. The price paid for capacity from all plants. tended to experience much more "miracle" occurred in part because Theoretical work on such uniform­ severe slowdowns which, in contrast East Asia's demographic transition price multi-unit auctions suggests that with the more developed countries, resulted in its working-age popula­ bidders selling more than one unit of began nearly a decade later. tion growing at a much faster pace a good have an incentive to increase than its dependent population during the prices they bid at high quantities. Convergence Clubs and 1965-90, thereby expanding the per If a bid sets the equilibrium price, Subsistence Economies capita productive capacity of East then the bidder receives a higher Asian economies. This effect was not price for that unit, as well as for all Dan Ben-David inevitable; rather, it occurred because inframarginal units. I find evidence of NBER Working Paper No. 6267 East Asian countries had social, eco­ strategic increases in bids. Plants that November 1997 nomic, and political institutions and are likely to be used only after a JEL Nos. E1, E2, 04 policies that allowed them to realize number of other plants are already This paper focuses on one possi­ the growth potential created by the operating will bid more. The larger ble explanation for the empirical evi­ transition. The empirical analyses supplier will submit higher bids, all dence of income convergence among indicate that population growth has else equal. Finally, there is some evi­ the world's poorest and wealthiest a purely transitional effect on eco­ dence that bids for given plants are countries, and income divergence nomic growth; this effect operates higher when the suppliers have more among most of the remaining coun­ only when the dependent and work­ available capacity. tries. The model incorporates the ing-age populations are growing at assumption of subsistence consump­ different rates. An important implica­ Geographic Concentration tion into the neoclassical exogenous tion of these results is that future as a Dynamic Process growth model, yielding outcomes demographic change will tend to Guy Dumais, Glenn Ellison, and that are consistent with the conver­ depress growth rates in East Asia, Edward L. Glaeser gence-divergence empirical evi­ while it will promote more rapid eco­ NBER Working Paper No. 6270 dence. While subsistence consump­ nomic growth in Southeast and South November 1997 tion can lead to negative saving and Asia. JEL Nos. L60, Rll disaccumulation of capital, it also can be associated with positive saving The degree of geographic concen­ and the accumulation of capital. The tration of individual manufacturing model predicts that the poorer the industries in the United States has country, the lower its saving rate. declined only slightly in the last 20 This result also appears to be borne years. At the same time, new plant out by the evidence I provide. births, plant expansions, contrac-

------NBER Reporter Winter 1997/8 57. tions, and closures have shifted large interacting, and dense urban areas Survival of the Fittest or the quantities of employment across increase the speed of interaction. The Fattest? Exit and Financing plants, firms, and locations. This model predicts that cities will have a in the Trucking Industry paper uses data from the Census higher mean and variance of skills. Luigi Zingales Bureau's Longitudinal Research Data­ Cities will attract young people who NBER Working Paper No. 6273 base to examine how relatively sta­ are not too risk averse and who ben­ November 1997 ble levels of geographic concentra­ efit most from learning (for example, tion emerge from this dynamic more patient people). Older, more This paper studies the impact of process. While industries' agglomera­ skilled workers will stay in cities only imperfections in the capital market tion levels tend to remain fairly con­ if they can internalize some of the on the natural selection of the most stant, there is greater variation in the benefits that their presence creates efficient firms. I estimate the effect of location of these agglomerations. for young people. The level of leverage before deregulation on the We decompose aggregate changes urbanization will rise when the survival of trucking firms after the in concentration into those portions demand for skills rises, when the Carter deregulation, and find that attributable to plant births, expan­ ability to learn by imitation rises, or highly leveraged carriers are less sions, contractions, and closures, and when the level of health in the econ­ likely to survive the deregulation find that the location choices of new omy rises. Empirical evidence on shock, even after I control for vari­ firms and the differences in growth urban wages supports the learning ous measures of efficiency. This rates have played the most significant view of cities and I corroborate effect is stronger in the imperfectly role in reducing levels of geographic empirically a variety of other impli­ competitive segment of the motor concentration, while plant closures cations of the theory. carrier industry. High debt seems to have tended to reinforce agglomer­ affect survival by curtailing invest­ ation. Finally, we look at coagglomer­ ments and reducing the price per­ ation patterns to test three of Marshall's Exceptional Exporter ton-mile that a carrier can afford to theories of industry agglomeration: 1) Performance: Cause, charge after deregulation. agglomeration saves transport costs Effect, or Both? by proximity to input suppliers or Andrew B. Bernard and Power in a Theory final consumers; 2) agglomeration J. BradfordJensen of the Firm allows for labor market pooling; and NBER Working Paper No. 6272 Raghuram G. Rajan and 3) agglomeration facilitates intellec­ November 1997 Luigi Zingales tual spillovers. While there is some JEL Nos. FlO, D21, 160 NBER Working Paper No. 6274 truth behind all three theories, we A growing body of empirical work November 1997 find that industrial location is driven has documented the superior per­ Transactions take place in the firm far more by labor mix than by any of formance characteristics of exporting rather than in the market because the the other explanatory variables. plants and firms relative to those that firm offers power to agents who do not export: employment, ship­ Learning in Cities make specific investments. The liter­ ments, wages, productivity, and cap­ ature has emphasized the allocation Edward L. Glaeser ital intensity are higher at exporters at of ownership as the primary mecha­ NBER Working Paper No. 6271 any given moment. We ask whether nism by which the firm does this. November 1997 good firms become exporters or in­ Within the contractibility assumptions JEL Nos. J24, 015, RIO stead that exporting improves firm of this literature, we identify a poten­ Alfred Marshall argues that indus­ performance. The evidence is quite tially superior mechanism, the regu­ trial agglomerations exist in part clear on one point: good firms be­ lation of access to critical resources. because individuals can learn skills come exporters. Both growth rates Access can be better than ownership from each other when they live and and measures of levels of success are because: 1) the power agents get work in close proximity. An increas­ higher ex-ante for the exporters. The from access is more contingent on ing amount of evidence suggests that benefits of exporting for the firm are them making the right investment; the informational role of cities is a less clear. Employment growth and and 2) ownership has adverse effects primary reason for their continued the probability of survival are higher on the incentive to specialize. The existence. This paper formalizes for exporters; however, productivity theory explains the importance of Marshall's theory using a model in and wage growth is not superior, par­ internal organization and third party which individuals acquire skills by ticularly over longer horizons. ownership.

58. NBERReporter Winter 1997/8------Job Destruction and were in their mid-20s were not likely We find little evidence that the Propagation of Shocks to gain such coverage later. Declines MSP rules affected the wages or Wouter J. den Haan, Garey in coverage are sharpest among the employment of covered workers. We Ramey, and Joel Watson least educated cohorts of young men. find weak evidence suggesting that NBER Working Paper No. 6275 We show that most of the decline is the MSP shifted the composition of November 1997 attributable to the substantial increase employment of older workers toward JEL Nos. E24, E32 in health insurance costs during the MSP-exempt jobs. We find strong evi­ 1980s. By contrasting young men's dence of low compliance with the We develop and quantitatively experiences with pension receipt MSP rules. implement a dynamic general equi­ with their health insurance experi­ Our results cast doubt on the effi­ librium model with labor market ences, we show that structural changes cacy of provisions designed to reduce matching and endogenous job in the labor market cannot explain crowd-out in new health insurance destruction. The model produces a any of the decline in coverage within programs. close match with data on job creation cohorts. and destruction. Cyclical fluctuations Our results suggest that the exist­ in the job destruction rate serve to A New Bankruptcy ing system of employer-sponsored magnify the effects of productivity Procedure that Uses health insurance subsidies did not shocks on output, as well as making Multiple Auctions compensate for the declines in earn­ those effects much more persistent. Oliver Hart, Rafael La Porta ings and increases in health insur­ Interactions between decisions about Drago, Florencio Lopez-de­ ance costs faced by young men be­ household saving and decisions Silanes, and John Moore tween 1981 and 1996. about separation in employment rela­ NBER Working Paper No. 6278 tionships playa key role in propa­ November 1997 Avoiding Health Insurance JEL Nos. G33, K22 gating shocks. Crowd-Out: Evidence from the Medicare-as-Secondary­ We develop a new bankruptcy Graduation to Health Payer Legislation procedure that makes use of multiple Insurance Coverage: Sherry Glied and Mark Stabile auctions. The procedure is designed 1981-1996 NBER Working Paper No. 6277 to work even when capital markets Sherry Glied and Mark Stabile November 1997 do not function well (for example in NBER Working Paper No. 6276 developing economies, or in econo­ The cost of efforts to expand November 1997 mies in transition), although it can be health insurance coverage to the cur­ used in all economies. The health insurance experience rently uninsured increases when peo­ of young men, who are new entrants ple who would otherwise purchase to the labor market, provides an early private insurance obtain subsidized Employer Learning and indicator of the strengths and weak­ public coverage. Legislators increas­ Statistical Discrimination nesses of the employer-sponsored ingly are interested in mechanisms Joseph G. Altonji and Charles R. Pierret health insurance system. Insurance that target insurance benefits to those NBER Working Paper No. 6279 coverage for these men has fallen who need them most. This paper November 1997 sharply over the past 15 years. We investigates the effects of one of the JEL Nos. D83, J31 examine patterns of health insurance first such targeting efforts, the 1982 coverage for cohorts of young men Medicare as Secondary Payer (MSP) We provide a test for statistical dis­ using successive cross-sectional sur­ provisions. crimination, or "rational" stereotyp­ veysand longitudinal data. The MSP rules require employers ing, in environments in which agents We find that coverage declines who offer insurance coverage to their learn over time, and apply it to the persist and are exacerbated as young employees under 65 to offer cover­ labor market. If profit maximizing men age. Not only did cohorts of age on the same terms to their Medi­ firms have limited information about men born during the 1950s fail to age care-eligible employees. This coverage the general productivity of new into employer-sponsored coverage as then becomes "primary" to Medicare. workers, then they may choose to they reached their 30s and 40s, but We examine the incidence of this use easily observable characteristics, they actually lost such coverage as implicit tax, the magnitude of tax such as years of education, to "statis­ they grew older. Furthermore, young avoidance efforts, and the extent of tically discriminate" among workers. men who lacked coverage when they tax compliance. As firms acquire more information

------NBERReporter Winter 1997/8 59. about a worker, pay will become sumer; in the latter there is a distribu­ of differences in tax rates and tax­ more dependent on actual produc­ tion of individuals across preference back schemes in transfer programs. tivity and less dependent on easily parameters. In the discrete model, We evaluate household demands for observable characteristics or creden­ taxes induce a large response from a leisure and consumption goods tials that predict productivity. subset of the population, while the numerically, using optimization tech­ Consider a wage equation that majority of the population shows niques within a larger equilibrium contains both the interaction be­ unchanged behavior. Welfare costs of structure including the production tween experience and a hard-to­ similar taxes in continuous models side of the economy, since demands observe variable that is positively can substantially exceed those in dis­ are nonanalytic. Our results suggest related to productivity and the inter­ crete models or vice versa, depend­ that a money-metric measure of the action between experience and a ing upon the formulation used. We utility equivalent of transfers received variable that firms can observe eas­ also report on experiments for a two­ by the bottom deciles of U.K. house­ ily, such as years of education. We labor type household model with holds in the early 1990s was only 32 show that the wage coefficient on the one continuous variable (secondary percent of cash transfers received unobservable productivity variable labor) and one discrete variable (pri­ because of the conditionality in these should rise with time in the labor mary labor), and do calculations programs. market, and the wage coefficient on using an empirically-based model education should fall. We investigate specification calibrated to UK. data. Apocalypse Now? this proposition using panel data on Our model clearly shows that dis­ Fundamental Tax Reform education, the AFQT test, father's crete choice matters in the assess­ and Residential Housing education, and wages for young men ment of the cost of labor supply tax Values and their siblings from NLSY. We also distortions. Donald Bruce and examine the empirical implications of Douglas Holtz-Eakin statistical discrimination on the basis The Redistributive Effects NBER Working Paper No. 6282 of race. Our results support the hy­ of Transfers November 1997 pothesis of statistical discrimination, Keshab Bhattarai and JEL No. H24 although they are inconsistent with John Whalley the hypothesis that firms fully utilize NBER Working Paper No. 6281 Using a simulation model crafted the information on race. Our analysis November 1997 to integrate the short- and long-run has wide implications for the analysis effects of tax reform on the housing Existing literature assessing the of the determinants of wage growth market, we find modest impacts from impacts of transfers on low income and productivity, and for the analysis fundamental reform of the federal households assumes that participants of statistical discrimination in the income tax. These results suggest that in transfer programs benefit by the labor market and elsewhere. concerns about the impact of tax full amount of cash transfers re­ reform on housing values and ceived. We argue here that because household net worth are overstated. Discreteness and the tax-back arrangements accompany Welfare Cost of Labor To the extent that reform is otherwise such transfer programs, and endog­ desirable, fears of drastic effects on Supply Distortions enous participation decisions (that Keshab Bhattarai and the housing market should not stand is, regime choices) are involved, a John Whalley as an impediment to reform. money-metric measure of the utility NBER Working Paper No. 6280 generated by transfers typically will November 1997 Costly Capital Reallocation be substantially less than the cash and the Effects of We discuss the effect of discrete value of transfers received. We use a choice of labor supply (or, leisure conditional choice general equilib­ Government Spending Valerie A. Ramey and consumption) on measures of the rium model of the United Kingdom, Matthew D. Shapiro welfare cost of labor supply tax dis­ calibrated to literature-based labor NBER Working Paper No. 6283 tortions. We construct comparable supply and labor demand elasticities, November 1997 continuous- and discrete-choice with a leisure-consumption choice JEL Nos. E62, E13, E22 models, each calibrated to have sim­ for household and production involv­ ilar aggregate (uncompensated) labor ing heterogeneous labor inputs. In Changes in government spending supply elasticities. In the former, the model, households face non­ often lead to significant shifts in there is a single representative con- convex budgets that are set because demand across sectors. We analyze

60. NBER Reporter Winter 1997/8------the effects of sector-specific changes Population and Ideas: equilibrium paths can occur starting in government spending in a two­ A Theory of at the same initial level of govern­ sector dynamic general equilibrium Endogenous Growth ment debt. rnodel in which the reallocation of Charles I. Jones capital acrOss sectors is costly. The NBER Working Paper No. 6285 The Market and the tWO-sector model leads to a richer November 1997 Estimators: Forecasting array of possible responses of aggre- JEL Nos. 041,030 the Cost of Medicare ate variables than the one-sector Catastrophic Coverage Why do economies exhibit sus­ ~odel. The empirical part of the Sherry Glied and Tama Brooks tained growth in per capita income? I aper estimates the effects of military NBER Working Paper No. 6287 p . f argue that endogenous fertility and buildups on a vanety 0 macroeco- November 1997 increasing returns to scale are the nomic variables using a new measure fundamental ingredients in under­ As part of the process of enacting of rnilitary shocks. The behavior of standing endogenous growth. Endog­ the Medicare Catastrophic Coverage rnacroeconomic aggregates is consis­ enous fertility leads the scale of the Act (MCCA) in 1988, both the Con­ tent with the predictions of a multi­ economy to grow over time. Increas­ gressional Budget Office (CBO) and sector neoclassical model. In particu­ ing returns translate this increase in the Department of Health and Human lar, consurnption; real product wages, scale into rising per capita income. Services (HHS) developed estimates and rnanufacturing productivity fall in According to this view, a justification of the cost of the pharmaceutical response to exogenous military build­ for increasing returns rather than component of the proposal. These ups in the post-World War II United linearity in the equation for techno­ estimates varied substantially. For States. logical progress is the fundamental some benefit years, cost estimates insight of the idea-based growth lit­ differed by a factor of more than two. erature. Endogenous fertility together This paper uses data from the with the increasing returns associated stock market to measure how market with the nonrivalry of ideas generates participants gauged the likely conse­ endogenous growth. quences of the MCCA, and to com­ pare the market estimate with those of the CBO and HHS estimators. We The Upcoming Slowdown Debts and Deficits with examine the market response to key in U.S. Economic Growth Fragmented Fiscal events linked to passage and repeal charles I. Jones Policymaldng of the MCCA for brand name and Andres Velasco NBER Working Paper No. 6284 generic pharmaceutical producers. NBER Working Paper No. 6286 November 1997 We find that on event days associated November 1997 JEL Nos. 040, 030 with passage of the MCCA, generic JEL Nos. H3, H6, E6 At least since 1950, the United pharmaceutical firms had positive States has been stimulated by in­ I develop a political-economic and significant excess stock market creases in educational attainment, model of fiscal policy in which gov­ returns. On early event days associ­ increases in research intensity, and ernment resources are a "common ated with passage, brand name pro­ the increased openness and devel­ property" out of which interest ducers had smaller positive returns o rnent of the world economy. Con- groups can finance expenditures on and on later days, brand name pro­ traryp to the conventlOna. l' View, such their preferred items. This setup has ducers had small negative returns. changes suggest that the U.S. econ- striking macroeconomic implications. On event days associated with repeal rny is far from its steady state bal­ Transfers are higher than a benevo­ of the MCCA, brand name producers ~nced groWth path. The theoretical lent planner would choose; fiscal had small positive excess returns and frarnework analyzed here provides a deficits emerge even when there are generic producers had zerO or small coherent interpretation of this evi­ no reasons for intertemporal smooth­ negative returns. dence, and indicates that when these ing and, in the long run, government The effect of the MCCA on the increases cease and the U.S. econ­ debt tends to be excessively high; stock price of pharmaceutical firms ornY reaches its steady state, then peculiar time profiles for transfers would depend on the elasticity of U.S. per capita growth can be expected can emerge, with high positive net demand for pharmaceuticals. Dif­ to fall to a rate of approximately one transfers early on giving way to high ferences in assumptions about this quarter its post-war average. taxes later on; and multiple dynamic elasticity were a key component of

_------NBER Reporter Winter 1997/8 61. the differences between CBO and "Make Us a King": Anarchy, The Alleged Instability of HHS estimates. Using the market Predation, and the State Nominal Income Targeting returns to evaluate these elasticities, Herschel I. Grossman Bennett T. McCallum we find that market participants NBER Working Paper No. 6289 NBER Working Paper No. 6291 shared the CBO's view that demand November 1997 November 1997 responses to the legislation would be JEL Nos. D60, D70, H40 JEL Nos. E52, E32 small. We also find that the market anticipated that the MCCA would In order to enforce a collective It has been argued recently that a strongly favor generic manufacturers. choice to allocate resources to guard­ monetary policy of nominal income ing against predators, producers must targeting would result in dynamically subject themselves to the state's sov­ unstable processes for output and The Sources of Regional ereign power to tax and spend. But, inflation. That result holds in a theo­ Variation in the Severity with these sovereign powers in hand, retical model that includes backward­ of the Great Depression: the state can exploit the producers by looking IS and Phillips-curve rela­ Evidence from u.S. taxing and spending for its own pur­ tions, but these are rather special and Manufacturing, 1919-37 poses. Using a general equilibrium theoretically unattractive. I demon­ Joshua L Rosenbloom and model in which people can choose strate that replacement of the special william. A. Sundstrom to be producers or predators, this pa­ Phillips curve with one of several NBER Working Paper No. 6288 per rationalizes the biblical request, more plausible specifications over­ November 1997 "Make us a king:' The analysis shows turns the instability result, whether or JEL Nos. N12, N62, E32 that, if the technology of predation is not the IS equation is replaced with a The severity of the Great Depres­ good enough, then having a "king" is forward-looking version. Thus, the sion in the United States varied by better than not having a king for instability result is quite fragile and region. Most notably, compared with everyone-including both producers provides almost no basis for a nega­ the rest of the country, the South and potential predators - even tive judgment on nominal income Atlantic states experienced a milder though the king maximizes the con­ targeting. contraction, while the Mountain sumption of a ruling elite. states suffered more severely. The Transition Issues for the impact of the contraction was more European Monetary Union uniform across other regions of the Willem H. Buiter and Anne C. Sibert country-surprisingly so, considering The Past, Present, and NBER Working Paper No. 6292 the large regional differences in Future of Macroeconomic November 1997 industrial structure. We use data from Forecasting Francis X. Diebold JEL Nos. F31, F33, F36, E58, E63 the biennial Census of Manufactures NBER Working Paper No. 6290 on 20 individual manufacturing in­ If Stage Three of EMU starts on November 1997 dustries disaggregated by state to January 1, 1999, then transition issues analyze the relative contributions of Broadly defined, macroeconomic will remain on two time scales. Until industry mix and location to regional forecasting is alive and well. Non­ July 1, 2002, national currencies and variations in economic performance structural forecasting, which is based the euro will co-exist as legal tender. during the period 1919-37. Industrial largely on reduced-form correlations, We argue that in principle intra-EMU composition had a significant impact always has been well and continues currency risk exists during that pe­ on regional employment growth, to improve. Structural forecasting, riod, but that no EMU member can be with regions that concentrated on the which aligns itself with economic forced out through speculative attacks. production of durable goods or in­ theory, and hence rises and falls with Cohabitation of "Ins" and "Outs" puts to the construction sector tend­ theory, receded following the decline has an open-ended time scale. We dis­ ing to fare worse than other regions. of Keynesian theory. In recent years, cuss the effect of EMU on incentives Long-run regional trends also played however, powerful new dynamic sto­ for both Ins and Outs to undertake an important role in regional varia­ chastic general eqUilibrium theory structural reform, and the coordina­ tion, and explain much of the South has been developed, and structural tion problems associated with the Atlantic region's more favorable per­ macroeconomic forecasting is poised distribution of seigniorage revenue formance over the cycle. for resurgence. and the Stability and Growth Pact.

62. NBERReporter Winter 1997/8------The Competitive Effects of of at least one party. It is possible for The Enforcement of Transmission Capacity in the interests of the two sides to con­ Intellectual Property a Deregulated Electricity flict. Conflict is certain if the excess Rights: A Survey of the ~;. ,l~; Industry demand for exported goods does not Empirical Literature i: Severin Borenstein, James respond to changes in the prices of Jean o. Lanjouw andJosh Letner ," ~. Bushnell, and Steven Stoft imported goods. In this case, any NBER Working Paper No. 6296 0,NBER Working Paper No. 6293 policy that raises imports must also December 1997 . , ~:November 1997 reduce welfare. JEL Nos. K41, 032, 034 }-:I:;' In an unregulated electricity gen­ We examine several recent ave­ ti eration market, the degree to which nues of empirical research into the " generators in different locations com- Implications of Rising enforcement of intellectual property pete with one another depends on Personal Retirement Saving rights. To frame these issues, we start the capacity to transmit electricity James M. Poterba, Steven F. Venti, with a stylized model of the patent between locations. We study the and David A. Wise litigation process. The bulk of the impaCt of transmission capacity on NBER Working Paper No. 6295 paper is devoted to linking the em­ competition among generators. We November 1997 pirical literature on patent litigation show that there may be no relation­ JEL Nos. J14, J26 to the parameters of this model. The ship between the effect of a trans­ Retirement savings accounts, par­ four major areas we consider are: 1) mission line in spurring competition ticularly employer-provided 401(k) how the propensity to litigate patents and the actual electricity that flows plans, have expanded rapidly in the varies with the expected benefits of on the line. We then investigate the last decade. More than 40 percent of litigation; 2) the ways in which the equilibriums that are likely to result workers are currently eligible for cost of litigation affects the willing­ as transmission lines between previ­ these plans, and over 70 percent of ness to enforce patents; 3) how the ously unconnected locations are built eligibles participate in these plans. cost of enforcing patents changes the and expanded. We demonstrate that The substantial and ongoing accu­ private value of patent rights; and 4) limited transmission capacity can give mulation of assets in these plans has the impact of intellectual property lit­ a firm the incentive to restrict its out­ the potential to significantly alter the igation on the innovation process put in order to congest transmission financial preparations for retirement itself. into its area of dominance. We apply by future retirees. We use data on cur­ this analysis to a model of California's rent age-specific patterns of 401(k) Stylized Facts of Patent forthcoming deregulated electricity participation, in conjunction with So­ Litigation: Value, Scope, market. Our results indicate that at cial Security earnings records which and Ownership least one firm could have an incen­ provide detailed information on age­ Jean O. Lanjouw and tive to strategically induce transmis­ Mark Schankerman earnings profiles Over the lifetime, to sion congestion, and that relatively NBER Working Paper No. 6297 project the 401(k) balances of future small investments in transmission December 1997 retirees. The results, which are illus­ may yield surprisingly large payoffs JEL Nos. 034, K41 trated by reference to individuals in terms of increased competition. who were 27 and 37 in 1996, demon­ We investigate the characteristics Welfare and Market strate the growing importance of of litigated patents by combining for Access Effects of Piecemeal 401(k) saving. The projected mean the first time information about Tariff Reform 401(k) balance at retirement for a patent case filings from the U.S. dis­ current 37 year old is $91,600, assum­ trict courts and detailed data from the JiandongJu and Kala Krishna ing that the 401(k) plan assets are U.S. Patent and Trademark Office. NBER Working Paper No. 6294 invested half in stocks and half in We construct a series of indicators for November 1997 bonds. For a current 27 year old, the the factors which the theoretical lit­ JEL Nos. Fl, F13 projected balance is $125,500. These erature suggests contribute to litiga­ In a situation where tariff reforms results support the growing impor~ tion: the frequency of disputes; the are being negotiated between two tance of personal saving through size and asymmetry of stakes; the parties, one of which aims to raise its retirement saving accounts in con­ structure of information; and costs. exports and the other to raise its wel­ tributing to financial well-being in Compared to a random sample of fare, tariff cuts must be in the interest old age. U.S. patents from the same cohorts

------NBER Reporter Winter 1997/8 63. and technology areas, we find, more Environmental Regulation do not seem consistent with impor­ valuable patents and those with and Labor Demand: tant empirical regularities surround­ domestic owners are considerably Evidence from the ing recent crises in emerging mar­ more likely to be involved in litiga­ South Coast Air Basin kets. This has generated considerable tion. Patents owned by individuals are Eli Berman and Linda T. Bui interest in models that associate at least as likely to be the subject of NBER Working Paper No. 6299 crises with self-fulfilling shifts in pri­ a case as corporate patents, and liti­ December 1997 vate expectations. gation is particularly frequent in new JEL Nos. C2, C8, H2, J23, L5, Q2 In this paper, I develop a first gen­ technology areas. We interpret the eration model based on an alternative results with reference to theoretical The nature of environmental reg­ policy conflict. Credit constrained gov­ models of litigation and settlement, ulation, by generating rich variation ernments accumulate reserve assets and discuss what they suggest about across regions and over time, pro­ in order to self-insure against shocks the effect of patent litigation on the vides an excellent opportunity for to national consumption. Govern­ incentives to invest in Rand D. estimating the effects of regulation on ments also insure domestic financial employment. We use direct measures markets that are poorly regulated. of regulation and plant data to esti­ Given this policy regime, a variety of mate the employment effects of an internal and external shocks gener­ unprecedented increase in air quality ate capital inflows to emerging mar­ regulation in the Los Angeles region, kets, followed by successful and Death to the Log-Linearized using for comparison unregulated anticipated speculative attacks. Consumption Euler plants in other regions, industries, I argue that a common external Equation! (And Very Poor and years. While environmental reg­ shock generated capital inflows to Health to the Second-Order ulation generally is thought to reduce emerging markets in Asia and Latin Approximation) employment, economic theory is am­ America after 1989. Country-specific Christopher D. Carroll biguous on this point, since pollution factors determined the timing of NBER Working Paper No. 6298 abatement technologies may use la­ speculative attacks. Lending policies December 1997 bor. We find that air quality regulation of industrial country governments JEL Nos. C6, D91, E21 induced very expensive investments and international organizations ex­ in abatement capital for individual plain contagion-that is, a bunching This paper shows that standard plants, especially for oil refineries. of attacks over time. empirical methods for estimating log­ Despite these high costs, we find no linearized consumption Euler equa­ evidence that environmental regula­ Predation, Efficiency, tions cannot successfully uncover tion decreased labor demand, even and Inequality structural parameters, like the coeffi­ when we allow for induced plant exit Herschel I. Grossman and cient of relative risk aversion, from and dissuaded plant entry. If any­ Minseong Kim the dataset of simulated consumers thing, air quality regulation probably NBER Working Paper No. 6301 behaving exactly according to the increased employment slightly. December 1997 standard model. Furthermore, con­ JEL Nos. D60, D74, D31, D50 sumption growth for the simulated A Model of Crises in consumers is very highly statistically We show how predation breaks Emerging Markets related to predictable income growth. the links between an economy's ag­ Michael P. Dooley Thus standard "excess sensitivity" gregate resource endowment and NBER Working Paper No. 6300 tests would reject the hypothesis that consumption and between the inter­ December 1997 consumers are behaving according to personal distribution of endowments JEL Nos. F32, F34 the standard model. Results are not and consumption. We construct a much better for the second-order First-generation models show that general-equilibrium model in which approximation to the Euler equation. apparently random speculative at­ some people (the privileged) are well I conclude that empirical estimation tacks on policy regimes can be fully endowed with resources and other of consumption Euler equations consistent with rational and well­ people (the unprivileged) are poorly should be abandoned, and discuss informed speculative behavior. Un­ endowed with resources. Each per­ some alternative empirical strategies fortunately, models driven by a con­ son can choose to be a producer or that are not subject to the problems flict between exchange rate policy a predator. In this model, the choice of Euler equation estimation. and other macroeconomic objectives to be a predator decreases aggregate

64. NBER Reporter Winter 1997/8------consumption, because predators' re­ This paper is concerned with inte­ waves of the 1984 Survey of Income SOurces are wasted by not being used gration between the economies of Program Participation (SIPP), and productively, and because producers the COre and periphery in the twenti­ find that high SSI benefits reduce sav­ sacrifice production by allocating re­ eth century world capital market. It ing among households with heads sources to guarding against predators. proceeds with some general obser­ who are approaching the SSI eligibil­ Analyzing this model, we find that vations and a special focus on the ity age and are likely to participate in the minimum equilibrium ratio of case of Argentina. I argue that under­ the program. predators to producers depends only standing the changing relations in on the technology of predation. In international capital markets offers Patent Buy-Outs: addition, the equilibrium ratio of important insights into the growth A Mechanism for predators to producers equals its and development process, especially Encouraging Innovation minimum value if and only if the for the countries of the periphery. Michael Kremer ratio of unprivileged to privileged Moreover, studying the extent of mar­ NBER Working Paper No. 6304 people is not larger than this mini­ ket integration in history informs cur­ December 1997 mum value. These properties imply rent conditions in the relationship JEL Nos. 031, 030, N40 that, in contrast to a model that ab­ between capital-scarce economies, In 1839, the French government stracts from predation, the fully egal­ like Argentina, and the global capital purchased the patent on the Da­ itarian distribution of resources does market as a whole. Looking to the fu­ guerreotype process and placed it in not satisfy the Rawlsian criterion of ture, the repercussions of economic the public domain. This paper exam­ maximizing the consumption of the reform and demographic change sug­ ines one mechanism under which person with the lowest consumption. gest likely implications for future sav­ governments would use an auction (In fact, the fully egalitarian distribu­ ing, investment, and international to estimate the private value of tion is not even Pareto efficient). capital flows. patents, and then offer to buyout Instead, the Rawlsian criterion selects patents at this private value times a an unequal distribution of resources The Effect of Means-Tested fixed markup. The markup would in which the ratio of unprivileged to Income Support for the correspond to the typical estimated privileged people equals the mini­ Elderly on Pre-Retirement ratio of the social and private values mum ratio of predators to producers, Saving: Evidence from the of inventions: for example, two. Most and in which unprivileged people SSI Program in the u.S. patents purchased would be placed have only the minimum possible David Neumark and in the public domain. But in order to endowment of resources. In the Elizabeth Powers induce bidders to reveal their valua­ resulting Rawlsian equilibrium, only NBER Working Paper No. 6303 tions, a few patents would be sold to unprivileged people choose to be December 1997 the highest bidder. These patent buy­ predators rather than producers; JEL Nos. 138, J14, J26 outs could eliminate monopoly price because both the ratio of predators We attempt to draw inferences distortions and incentives for waste­ to producers and the amount of about potential behavioral responses ful reverse engineering, while raising resources that predators waste are to means-tested income support for private incentives for original re­ minimized, aggregate consumption is the elderly by examining the effects search. However, patent buyouts are maximized. In addition in the Rawl­ on saving of the Supplemental Secu­ potentially vulnerable to collusion. sian equilibrium, predation equalizes rity Income (SSI) program for the Patent buyouts may be particularly the consumption of the privileged aged in the United States. Part of the appropriate for pharmaceuticals. and the unprivileged. SSI program provides payments to the poor elderly, thus operating as a Are Public Housing Argentina and the World means-tested public retirement pro­ Projects Good for Kids? Capital Market: Saving, gram. The federal government sets Janet Currie and Aaron Yelowitz Investment, and Inter­ eligibility criteria and benefit levels NBER Working Paper No. 6305 national Capital Mobility for the federal component of the pro­ December 1997 in the Twentieth Century gram, but many states supplement JEL Nos. H5, 13, JO, RO Alan M. Taylor federal SSI benefits substantially. We NBER Working Paper No. 6302 exploit the state-level variation in SSI One of the goals of federal hous­ December 1997 benefits to estimate the effects of SSI ing policy is to improve the prospects JEL Nos. F41, F43, N16, N26, 011, 054 on saving. We use data from selected of children in poor families. But little

------NBERReporter Winter 1997/8 65. research has been conducted into the The design of rules for central cies. The model can be solved in effects on children of housing pro­ bank policy has been a subject of closed form and illustrated in terms grams, perhaps because it is difficult increasing interest to many monetary of the simplest graphical apparatus, to find datasets with both information economists. This essay first presents so as to provide the analysis of mac­ about participation and interesting an analytical structure in which a pol­ roeconomic interdependence, struc­ measures of outcome. This paper com­ icymaker is presumed to formulate a tural spillovers, policy links, and bines data from several sources in or­ rule based on the solution to an opti­ strategic complementarities with rig­ der to provide a first look at the effect mal control problem, and then to orous but intuitive micro-founda­ of participation in public housing examine a number of issues that are tions. In contrast with the traditional projects on housing quality and on the germane to the current debate on the literature, our findings emphasize the educational attainment of children. nature of such rules. These issues positive externalities of foreign mon­ We use administrative data from include the implication for policy­ etary expansions and foreign fiscal the Department of Housing and Ur­ making of the slope of the output­ contractions on domestic welfare, ban Development to impute the prob­ inflation variability frontier, the while highlighting the ambiguous ability that a Census household lives importance of various types of uncer­ effects on welfare of domestic policy in a public housing project. We find tainty, the consequences of a zero shocks. that a higher probability of living in a nominal interest rate floor, and the project is associated with poorer out­ possible reasons for interest rate comes. We then use two-sample smoothing. Social Security and instrumental variables (TSIV) tech­ Although this essay is intended to Retirement in Canada niques to combine information on raise, rather than resolve, key ques­ Jonathan Gruber the probability of living in a project, tions concerning policy rules, it does NBER Working Paper No. 6308 based on the 1990 to 1995 Current offer fairly compelling evidence on December 1997 Population Surveys, with information one point: the potential conse­ JEL Nos. J14, J26, H55 on outcomes that we obtained from quences of the move by many central the 1990 Census. The common instru­ banks toward some form of price­ In Canada, government transfers ment in both samples is an indicator level or inflation targeting. In adopt­ to older persons are one of the larg­ that is set equal to one if the house­ ing such an approach, central banks est and fastest growing components hold is entitled to a larger housing implicitly are changing the relative of the government budget. I provide project unit because of the sex com­ importance of output and inflation an overview of the interaction be­ position of the children in the house­ variability in their objective function. tween these transfer programs and hold. Those families entitled to a The robustness of the policy rule, retirement behavior. I begin by doc­ larger unit are 24 percent more likely however, may depend on the shape umenting historical trends in labor to live in projects. After controlling of the output-inflation variability force participation and program for omitted variables bias using TSIV, tradeoff. The data indicate that this receipt, and contemporaneous pat­ we find that project households are tradeoff is extremely steep: small terns of work and income receipt for less likely to suffer from overcrowding decreases in inflation variability are the current cohort of older persons. I and less likely to live in high-density associated with very large increases then describe the structure of this complexes. Project children are also in output variability. This finding sug­ system of Canadian transfer pro­ 12 to 17 percentage points less likely gests that pure inflation targeting may grams. Finally, I present the results of to have been held back in school one have very undesirable side effects. a simulation model which measures or more grades, although this effect is the implicit tax/subsidy rate on work after age 55 within this system. I find confined to boys. Thus, most families Welfare and do not face a tradeoff between hous­ that, for married male workers, there Macroeconomic are modest taxes on work through ing quality and child outcomes: the Interdependence age 64 which rise to fairly high levels average project improves both. Giancarlo Corsetti and thereafter. But these taxes are sub­ Paolo Pesenti Central Bank Policy Rwes: stantially lower for single workers, NBER Working Paper No. 6307 Conceptual Issues and since they do not have wives who December 1997 Practical Considerations are eligible for means-tested trans­ Stephen G. Cecchetti We develop a simple choice-theo­ fers, and for workers with other sub­ NBER Working Paper No. 6306 retic model for carrying out welfare stantial sources of income, since the December 1997 analyses of the international trans­ family is not at all eligible for means­ JEL Nos. E52, E58 mission of monetary and fiscal poli- tested transfers.

66. NBER Reporter Winter 1997/8------

------~------Corporate Governance Managing the Public Debt impact of capital market liberaliza­ Luigi Zingales in Fiscal Stabilizations: tions on the correlation between NBER Working Paper No. 6309 The Evidence emerging equity market returns and December 1997 Alessandro Missale, Francesco the world market return. Our empir­ JEL No. G3 Giavazzi, and Pierpaolo Benigno ical approach is designed to control NBER Working Paper No. 6311 for other economic events which This essay summarizes my own December 1997 might confound the impact of foreign personal view of what corporate gov­ JEL Nos. E31, E41, E43, G12 speculators on local equity markets. ernance is about. I argue that it makes Whatever the empirical specification, sense to discuss corporate gover­ This paper provides evidence on the cost of capital always decreases nance only in an incomplete-contract the behavior of public debt managers after capital market liberalization, but world. In this world, the notion of during fiscal stabilizations in OEeD the effect is economically and statis­ corporate governance is intrinsically countries over the last two decades. tically weak. The effects on volatility related to the definition of the firm. We find that debt maturity tends to and correlation are less robust. In this respect, I review the short­ lengthen: the mOre credible the pro­ comings of the existing definitions of gram; the lower the long-term interest the firm and the possible applications rate; and the higher the volatility of of the idea that the firm is a "nexus of short-term interest rates. We show that specific investments" introduced by this debt issuing strategy is consistent Unequal at Birth: Rajan and Zingales [19971. I conclude with optimal debt management if A Long-Term Comparison by discussing the limitations of the information between the government of Income and Birth incomplete contracts approach to COr­ and private investors is asymmetric, Weight porate governance. as is usually the case at the outset of Dora L. Costa a stabilization attempt, when private NBER Working Paper No. 6313 Sharp Reductions in investors may lack full confidence in December 1997 Current Account Deficits: the announced budget cuts. JEL Nos. 112, N30 An Empirical Analysis I demonstrate that although socio­ Gian Maria Milesi-Ferretti economic differences in birth weight and Assaf Razin have always been fairly small in the NBER Working Paper No. 6310 United States, they have narrowed December 1997 since the beginning of this century. I JEL Nos. F32, F34 Foreign Speculators and argue that maternal height, and there­ We study the determinants and Emerging Equity Markets fore the mother's nutritional status consequences of sharp reductions in Geert Bekaert and during her growing years, accounted Campbell R. Harvey current account imbalances (rever­ for most of the socioeconomic differ­ NBER Working Paper No. 6312 sals) in low- and middle-income ences in birth weight in the past, but December 1997 countries. We try to answer two ques­ not today. This implies that in the JEL Nos. F3, GO, G1 tions: First, what triggers reversals? past health inequality was transmit­ Second, what factors explain the A number of countries have de­ ted across generations. I also show costliness of reversals? We find that layed the opening of their capital that children born at the beginning both domestic variables, such as the markets to international investment of this century compared favorably to current account balance, openness, because of reservations about the modern populations in terms of birth and the level of reserves, and exter­ impact of foreign speculators on ex­ weights, but suffered higher fetal and nal variables, such as the terms of pected returns and market volatility. neonatal death rates because obstet­ trade shocks, U.S. real interest rates, We propose a cross-sectional time­ rical and medical knowledge was and growth in industrial countries, series model that attempts to assess poorer. In addition, by day ten of life, play an important role in explaining the impact of market liberalizations children of the past were at a disad­ reversals in current account imbal­ - in the form of offering depositary vantage relative to children today ances. We also find that countries receipts, country funds, and other because "best practice" resulted in in­ with a less appreciated real exchange financial instruments, in an extrana­ sufficient feeding. The poor average rate, higher investment, and open­ tional market-on the cost of capital health of past populations therefore ness prior to the reversal tend to and market volatility in emerging originated in part in the first days grow faster after a reversal occurs. equity markets. We also examine the of life.

------NBER Reporter Winter 1997/8 67. NIMBY Taxes Matter: quite persistent, even conditional on restructure their old-age pension sys­ State Taxes and Interstate sales. tems, and we highlight available evi­ Hazardous Waste Shipments To explain the two facts, we pre­ dence on the performance of the Arik Levinson sent a linear-quadratic model. The system in each case. NBER Working Paper No. 6314 model can rationalize the facts in a December 1997 number of ways, but two stylized Myopia and Inconsistency JEL Nos. H73, H23 explanations have the virtue of rela­ in the Neoclassical tive simplicity and support from a Growth Model This paper examines the extent to number of papers. Both assume that Robert J. Barro which state taxes have inhibited in­ there are persistent shocks to de­ NBER Working Paper No. 6317 terstate transport of hazardous waste mand for the good in question, and December 1997 for disposal in the United States. I use that the marginal production cost JEL Nos. 040, E21 panel data from the Toxics Release curve slopes up. The first explanation Inventory (TRI) and the Resource Con­ also assumes that there are highly I modify the neoclassical growth servation and Recovery Act (RCRA) persistent shocks to the cost of pro­ model to allow for a varying rate of on interstate shipments of waste, and duction. The second assumes that time preference. If the household analyze them in conjunction with a there are strong costs of adjusting cannot commit on future choices of set of state characteristics, including production and a strong accelerator consumption, and if utility is loga­ hazardous waste disposal taxes and motive. rithmic, then the equilibrium will disposal capacity. I use four ap­ However, research to date has not resemble the standard results of the proaches to deal with the potential reached a consensus on whether one neoclassical model. In this solution, endogeneity of taxes and unobserved of these two, or some third alterna­ the effective rate of time preference heterogeneity among states: a "nat­ tive, provides a satisfactory expla­ is high, but constant. Although this ural experiment" based on the retal­ nation of inventory behavior. We model has potentially important iatory nature of some state tax laws; suggest several directions for future implications for institutional design a fixed-effects model; a two-stage research that promise to improve our and other policies-because house­ least squares estimator; and a rein­ understanding of inventory behavior holds would benefit from an ability terpretation of the coefficient on the and thus of business cycles. to commit future consumption­ distance among states. I conclude there is a sense in which the results that hazardous waste taxes are a sta­ After Chile, What? Second­ are observationally equivalent to tistically and economically significant Round Pension Reforms in those of the conventional model. deterrent to interstate waste trans­ Latin America When I extend the framework to port, that taxes are being imposed by Olivia s. Mitchell and allow for partial ability to commit, large-capacity and large-import Flavio Ataliba Barreto some testable hypotheses emerge states, and that these taxes therefore NBER Working Paper No. 6316 concerning the link between this have had a decentralizing effect on December 1997 ability and the rates of saving and the national pattern of hazardous JEL Nos. J26, J14, H55, G23 growth. I obtain steady-state results waste transport and disposal. for general concave utility functions, The apparent success of Chile's and work out some properties of the pension reform catalyzed a number dynamic paths for the case of iso­ Inventories of subsequent reforms in sister Latin Valerie A. Ramey and elastic utility. American nations. The "Chilean Kenneth D. West model" now has captured the atten­ NBER Working Paper No. 6315 Unions and Managerial Pay tion of policymakers and researchers December 1997 John DiNardo, Kevin Hallock, in the GECD as well. In this paper, JEL Nos. E22, E32 and Jorn-Steffen Pischke we identify six critical elements of NBER Working Paper No. 6318 We review and interpret recent old-age pension reform, and exam­ December 1997 work on inventories, emphasizing ine how these elements differ across JEL Nos. J31, J44, J51 empirical and business cycle aspects. the Chilean reform, and in several We begin by documenting two em­ other Latin nations that followed in Unions compress the wage distri­ pirical regularities about inventories: Chile's footsteps. We emphasize how bution among workers covered by that inventories move procyclically; these other Latin American nations union contracts. We ask whether and that inventory movements are adopted different mechanisms to unions also have an effect on the

68. NBERReporter Winter 1997/8------managers of unionized firms. To this other possible costs of currency uni­ volatility is less than -1 for reason­ end, we collect and assemble data on fication that are associated with a able parameter values (that is, a 1 unionization and managerial pay reduced number of asset markets. On percent increase in the coefficient of within firms and industries in the the benefit side, my theories of the variation of productivity shocks would United States and across countries. efficiencies attributable to a common reduce welfare by more than 1 per­ Generally, we find a negative COrre­ currency remain unsatisfactory, cent). We suggest that legal and infor­ lation between executive compensa­ despite recent advances. A key moti­ mation problems in the credit market tion and unionization in Our cross­ vation for the choice of a common may then underlie why volatility has section data, but no relationship currency rather than a fixed ex­ profound effects on emerging market between changes in unionization and change rate among national curren­ economies. the growth of compensation of exec­ cies is the fear of speculative attack. utives over time. Using data on NLRB I conclude by showing how self­ Market Forces and elections, we find that a loss of union fulfilling currency crises can occur, Sex Discrimination members attributable to decertifica­ and describe recent progress in nar­ Judith K. Hellerstein, tion elections is associated with rowing the range of multiple equilib­ David Neumark, and higher CEO pay, although our esti­ riums in adjustable-peg regimes. Kenneth R. Troske mates are imprecise. With CPS data, NBER Working Paper No. 6321 we consistently find that where Volatility and Financial December 1997 unions are stronger, fewer managers Intermediation JEL Nos. ]71, J16, J24 are employed. Joshua Aizenman and We report new evidence on the Andrew Powell existence of sex discrimination in Open-Economy NBER Working Paper No. 6320 wages and whether competitive mar­ Macroeconomics: December 1997 ket forces act to reduce or eliminate Developments in JEL Nos. F32, F34, E44 Theory and Policy discrimination. Specifically, we use Following the Tequila period, its plant- and firm-level data to examine Maurice Obstfeld after effects in Latin America, and the relationships among profitability, NBER Working Paper No. 6319 recent events in South East Asia, the changes in growth and ownership, December 1997 effect of volatility on emerging mar­ product market power, and the sex JEL No. F41 ket economies has become an impor­ composition of a plant Or firm's work­ This paper surveys recent research tant topic of research. The domestic force. Our strongest finding is that, in open-economy macroeconomics, financial intermediation process is among plants with high levels of using questions raised by European being advanced as one of the most product market power, those that economic and monetary unification important transmission mechanisms. employ relatively more women are to guide the topics discussed. A strik­ At the same time, there has been more profitable. No such relationship ing empirical regularity is the ten­ continued interest in the issues of exists for plants with apparently low dency for changes in the nominal imperfect information and rationing levels of market power. This is con­ exchange rate regime to systemati­ in credit markets. In this paper, we sistent with sex discrimination in cally affect the variability of nominal consider an economy where risk wages in the short run in markets and real exchange rates. This regu­ neutral banks provide intermediation where plants have product market larity (which disappears in high-infla­ services and risk neutral producers power. We also examine evidence on tion conditions) can be explained by demand credit to finance their work­ the longer-run effects of market sticky-price theories or by models of ing capital needs. Our model blends forces on discrimination, asking asset-market liquidity effects. But costly state verification with imper­ whether discriminatory employers plausible liquidity models have diffi­ fect enforcement power. In this con­ with market power are punished culty generating enough persistence text of costly financial intermediation, over time through lower growth than (in output and real exchange rates, in we show that a weak legal system non-discriminatory employers, or particular) to match the data. Thus combined with high costs of infor­ whether discriminatory employers the macroeconomic costs of giving mation verification leads to large, are bought out by non-discriminators. up the exchange-rate realignment first-order effects of volatility on pro­ We find little evidence that this oc­ option, emphasized in Mundell's duction, employment, and welfare. A curs over a five-year period, as growth optimum currency area concept, calibration illustrates that the semi­ and ownership changes for plants seem empirically relevant. I discuss elasticity of welfare with respect to with market power are generally not

------NBERReporter Winter 1997/8 69. significantly related to the sex com­ trial organization models (with prod­ The Central Tendency: position of a plant's workforce. uct differentiation and students being A Second Factor in both consumers of, and inputs into, Bond Yields higher education) would predict: Pierluigi Balduzzi, Sanjiv Ranjan Regions, Resources, and higher average college quality and Das, and Silverio Foresi Economic Geography: tuitions; greater between-college NBER Working Paper No. 6325 Sources of U.S. Regional variation in tuition and student qual­ December 1997 Comparative Advantage, ity; less within-college variation in JEL No. G12 1880-1987 student quality; higher average sub­ We assume that the instantaneous SukkooKim sidies to students; and greater riskless rate reverts towards a central NBER Working Paper No. 6322 between-college variation in subsi­ tendency which, in turn, is changing December 1997 dies. Changing market structure can randomly over time. As a result, cur­ JEL Nos. R12, Fl explain real tuition increases of rent short-term rates are not sufficient approximately 50 percent for selec­ This paper estimates the Rybczyn­ to predict future short-term rate tive private colleges. I use panel data ski equation matrix for the 20 two­ movements, as would be the case if from 1940 to 1991 on 1121 baccalau­ digit U.S. manufacturing industries the central tendency were constant. reate-granting colleges, including for various years between 1880 and However, since longer-maturity bond data on students' home residences, 1987. As predicted by the standard prices incorporate information about for this study. general equilibrium theory of interre­ the central tendency, longer-maturity gional trade, the regression estimates bond yields can be used to predict show that a consistent set of factor future short-term rate movements. endowments explains a significant We develop a two-factor model of amount of the geographic distribu­ A Unified Theory of the term structure which implies that tion of manufacturing activities over Underreaction, Momentum a linear combination of any two rates time. Although these results do not Trading, and Overreaction can be used as a proxy for the central rule out the importance of increasing in Asset Markets tendency. Based on this central ten­ returns, they do suggest certain limits Harrison Hong and dency proxy, we estimate a model of on how increasing returns affect U.S. JeremyC. Stein the one-month rate that performs economic geography. NBER Working Paper No. 6324 better than models which assume the December 1997 central tendency to be constant.

How the Changing Market We model an asset market popu­ Pecuniary Incentives to Structure of U.S. Higher lated by two grou ps of boundedly Work in the U.S. During rational agents-"newswatchers" and Education Explains College World War II "momentum traders" -each of whom Tuition C~ey B. Mulligan can process only a subset of all avail­ Caroline M. Hoxby NBER Working Paper No. 6326 able information. The bounded ration­ NBER Working Paper No. 6323 December 1997 ality of the news watchers creates a December 1997 tendency for prices to underreact to I argue that changes in workers' This paper presents theoretical private information in the short run. budget sets cannot explain the dra­ and empirical evidence demonstrat­ This underreaction in turn means that matic increases in civilian work in the ing how the changing market struc­ the momentum traders can make United States during World War II. ture of American higher education money by chasing trends. However, Although money wages grew during from 1940 to the present affected col­ if they are restricted to following sim­ the period, wartime aftertax real lege prices and quality. Over this ple (that is, univariate) strategies, wages were lower than either before period, the market for baccalaureate their best attempts at arbitrage, al­ or after the war. Evidence from the education became significantly more though profitable, inevitably must 1940s also appears to be inconsistent competitive, as it was transformed lead to overreaction at long horizons. with other pecuniary explanations, from a collection of local autarkies to In addition to providing a unified ac­ such as wealth effects of government a nationally integrated market. I count of asset market under- and over­ policies, intertemporal substitution demonstrate that the results of in­ reactions, the model generates a num­ induced by asset prices, unfulfilled creased competition were what indus- ber of distinctive testable implications. expectations, and changes in the

70. NBERReporter Winter 1997/8------nonmarket price of time. Although An additional year of general educa­ Has Job Stability Declined untested and relatively undeveloped, tion results in a 5 percent increase in Yet? New Evidence for nonpecuniary models of behavior are teacher wages, and a 2.5 percent the 1990s tempting explanations for wartime increase in aide wages. Specialized David Neumark, Daniel PoJsky, work. training influences teacher wages, and Daniel Hansen but has less impact on aide wages. NBER Working Paper No. 6330 Whither the World Bank Unionization has a large impact on December 1997 and the IMF? both wages and compensation of JEL Nos. e80, ]21, ]63 Anne o. Krueger teachers and aides. Alternative wages In earlier work, we examined the NBER Working Paper No. 6327 of the workers are related positively evolution through the 1980s of job December 1997 to teacher and aide wages. An increase stability in U.S. labor markets, using ]EL No. F33 in local unemployment decreases data assembled from a sequence of aides' wages, but has a positive im­ On their fiftieth anniversary, the Current Population Survey tenure pact on the teachers' wages. There is International Monetary Fund and the supplements. We found little or no evidence of profit sharing in the case World Bank were reviewed exten­ change in aggregate job stability in of aides, but not teachers. An increase sively, both to mark the occasion and the U.S. economy. In addition, older in center size raises teacher wages. to consider, often critically, their cur­ and more-tenured workers experi­ This body of evidence indicates rent roles and performance. This pa­ enced increases in job stability in the that both teacher and aide remuner­ per reviews the functions of the two latter part of the 1980s. ation have non-competitive flavors, institutions, in light of their evolution In this paper, we update the evi­ where the case is more compelling over the past 50 years and of changes dence on changes in job stability for aides. in the international economic system. through the rnid-1990s, using recently­ It then evaluates and assesses some released CPS data for 1995 that par­ of the criticisms and proposals for allel the earlier job tenure supple­ reform of the two institutions. Political Economics and ments. Updating the evidence from Macroeconomic Policy systematic random samples of the The Determinants of Torsten Persson and population and workforce through Child Care Workers' Guido Tabellini this period is especially important, Wages and Compensation: NBER Working Paper No. 6329 because the media have painted a Sectoral Difference, Human December 1997 particularly stark picture of declining Capital, Race, Insiders ]EL Nos. E5, E6, H2, H3, 01 job stability in the 1990s. and Outsiders This paper surveys the recent lit­ In the aggregate, there is some H. Naci Mocan and erature on the theory of macroeco­ evidence that job stability declined Deborah Viola nomic policy. We study the effect of modestly in the first half of the 1990s. NBER Working Paper No. 6328 various incentive constraints, such as Moreover, the relatively small aggre­ December 1997 lack of credibility, political opportu­ gate changes mask rather sharp ]EL Nos.]13,]31,]41,]44,]51, L3,]24 nism, political ideology, and divided declines in stability for workers with This paper investigates the deter­ government, on the policymaking more than a few years of tenure. minants of wages and compensation process. The survey is organized in Nonetheless, the data available up to for teachers and aides in child care three parts. Part I deals with mone­ this point do not support the conclu­ centers. Non profit status has no tary policy in a simple Phillips curve sion that the downward shift in job across-the-board impact on wages. model: it covers credibility issues, stability for more-tenured workers, The extent of the wage premium political business cycles, and optimal and the more modest decline in enjoyed by some nonprofit workers design of monetary institutions. Part aggregate job stability, reflect long­ depends on the category of the non­ II deals with fiscal policy in a dy­ term trends. profit center, the occupation of the namic general equilibrium set up: the workers, and their race. The rate of main topics are credibility of tax return to an additional year of tenure policy and political determinants of is 2 percent for both teachers and budget deficits. Part III studies eco­ aides; for prior experience, it is 1 per­ nomic growth in models with endog­ cent for teachers and zero for aides. enous fiscal policy.

------NBERReporter Winter 1997/8 71. The Careers of Modern tion of business regulation in Hong options show six times less respon­ Artists: Evidence from Kong, analyze the salient features of siveness to taxation. Other types of Auctions of Contemporary its scheme of control regulation, and compensation, such as salary and Paintings evaluate the impact of transition from bonus or nontaxed income, are either David W. Galenson regulation to competition. To sharply not responsive to tax rates or not NBER Working Paper No. 6331 contrast the difficulties of the tradi­ large enough to make a difference. December 1997 tional approach to regulation with The estimated elasticities indicate that JEL No. J24 the benefits of introducing competi­ the dead weight loss of recent tax tion, we focus on the cases of elec­ increases was around 15 to 25 per­ Using transactions from fine art tricity and telecommunications. We cent of the revenue generated. auctions for 42 leading American also discuss the direction for future contemporary artists, I estimate the changes. The Mexican Peso Crisis: relationship between the value of a How Much Did We Know? painting and the artist's age at the What Happens When When Did We Know It? date of its creation. I show that artists You Tax the Rich? Sebastian Edwards born before 1920 were likely to have Evidence from Executive NBER Working Paper No. 6334 done their most valuable work late in Compensation December 1997 their careers, while artists born in the Austan Goolsbee JEL Nos. F3, F4, 054, 023 1920s and 1930s were more likely to NBER Working Paper No. 6333 have done their most valuable work The Mexican crisis of 1994 raised a December 1997 at an early age. Comparing these re­ number of questions, throughout the JEL No. H24 sults with evidence drawn from art world, regarding the sustainability­ history textbooks and museum exhi­ This paper reexamines the respon­ and even the merits-of the market bitions indicates that these artists' siveness of taxable income to changes oriented reform process in Latin most valuable work also has been in marginal tax rates, using detailed America and other regions. Under­ the most highly regarded by scholars. compensation data on several thou­ standing the way events unfolded in I argue that the shift across genera­ sand corporate executives from 1991 Mexico during the early 1990s con­ tions in the shape of these artists' to 1995. The data confirm that the tinues to be fundamentally important age-price profiles was a result of both higher marginal rates of 1993 led to a for assessing the mechanics of cur­ the evolution of modern painting and significant decline in taxable income. rency crises. More importantly, per­ a growth in the demand for contem­ Indeed, this small group of execu­ haps, the eruption of the East Asian porary American art during the 1950s tives can account for as much as 20 currency crises in the summer and and 1960s. percent of the aggregate change in fall of 1997 has raised the question of wage and salary income for the mil­ whether the lessons from Mexico lion richest taxpayers, and one per­ indeed have been learned by policy­ Hong Kong's Business son alone can account for more than makers, private sector analysts, and Regulation in Transition 2 percent. The decline, however, is international civil servants. More Changqi Wu and almost entirely a short-run shift in the specifically, as a result of the recent Leonard K. Cheng timing of compensation, rather than a events in South East Asia, many NBER Working Paper No. 6332 permanent reduction in taxable observers have argued that the inter­ December 1997 income. The short-run elasticity of national financial organizations-the JEL Nos. L43, LSI, L94, L96 taxable income with respect to the IMF and the World Bank - and the The transition of Hong Kong's net-of-tax share exceeds one, but the governments of the advanced coun­ main economic activities from manu­ elasticity after one year is at most 0.4 tries have failed to revamp the early facturing to services is accompanied and probably close to zero. The re­ warning system that was supposed to by gradual changes in the regulatory sponse comes almost entirely from a prevent a repetition of a Mexico-style regimes for monopolies. The local large increase in the exercise of stock crisis. This paper analyzes the causes telecommunication services industry options in the year before the tax behind the Mexican crisis, emphasiz­ has been liberalized; deregulation of change, followed by a decline in the ing the role of capital inflows, infla­ public transport is taking shape; and year of the tax change, and the tionary inertia, and real exchange the schemes of control for electricity change is concentrated among exec­ rate overvaluation. I also ask a num­ suppliers are candidates for reform. utives at the top of the income distri­ ber of questions regarding the pre­ In this paper, we review the evolu- bution. Executives without stock dictability of the crisis: Should Wall

72. NBERReporter Winter 1997/8------Street analysts have known that things which government reSOurces are a that could be transferred to the next were getting out of hand? And if they "common property" out of which generation with a program of "planned did, why didn't they alert their interest groups can finance expendi­ giving": using the $10,000 per donor clients? And, how much did officials tures on their preferred items. This per donee exemption from the gift at the U.S. Treasury know about the setup has striking macroeconomic tax. While roughly one quarter of po­ depth of the Mexican problems? implications. First, fiscal deficits and tentially taxable assets could be trans­ What was the role of the media? I debt accumulation occur even when ferred in this way, the actual levels of conclude that although the U.S. Treas­ there are no reasons for intertempo­ inter vivos giving are much lower ury was fully aware of what- was ral smoothing. Second, those deficits than the levels that one would expect going on, most private sector analysts can be eliminated through fiscal re­ if households were taking full advan­ were unaware of the seriousness of form, but such a reform may only tage of this tax avoidance strategy. the situation. take place after a delay, during which government debt is built up.

Economic Integration and Channeling Domestic Convergence: u.s. Regions, The Estate Tax and After­ Savings into Productive 1840-1987 Tax Investment Returns Investment under SukkooKim Jrunes Poterba Asymmetric Information: NBER Working Paper No. 6335 NBER Working Paper No. 6337 The Essential Role of December 1997 December 1997 Foreign Direct Investment JEL Nos. N31, N32, 040, R12, F1 JEL Nos. H24, J14 Assaf Razin, Efraim Sadka, and Cbi-Wa Yuen This paper explores the effect of Despite the recent inroads made NBER Working Paper No. 6338 estate and gift taxes on the aftertax by models of interregional trade December 1997 rate of return earned by savers. The based on external economies, the JEL Nos. F21, F35, H25, H30 analysis of the long-run trends in U.S. estate tax affects only a small fraction regional specialization in agriculture, of households - taxable decedents Foreign direct investment (FDI) is manufacturing, wholesale trade, retail represented only 1.4 percent of all a predominant form of capital flow trade, services, and all economic deaths in 1995 - but the affected to low- and middle-income countries activities indicate that these trends households account for a substantial with insufficiently developed capital are more consistent with explana­ fraction of household net worth. The markets. This paper analyzes the tions based on the neoclassical estate tax can be viewed as a tax on problem of channeling domestic sav­ Heckscher-Ohlin model. Further­ capital income, with the effective rate ings into productive investment in more, while the long-run trends in depending on the statutory tax rate the presence of asymmetric informa­ U.S. regional industrial structures do as well as the potential taxpayer's tion between the managing owners not explain all the variations in mortality risk. Because mortality rates of firms and other portfolio stake­ regional income per capita, they play rise with age, the effective estate tax holders. We emphasize the crucial an important role in causing U.S. burden is greater for older than for role played by FDI in sustaining regional incomes to diverge and then younger individuals. The estate tax equity-financed capital investment for converge between the nineteenth adds approximately 0.3 percentage economies plagued by such informa­ and the twentieth centuries. points to the average tax burden on tion problems. Similar problems also capital income for households headed exist for foreign portfolio debt flows. by individuals between the ages of We identify how, in the presence of A Model of Endogenous 50 and 59. For households headed by information asymmetry, different cap­ Fiscal Deficits and individuals between the ages of 70 ital market structures may lead to for­ Delayed Fiscal Reforms and 79, however, the estate tax in­ eign over- or underinvestment and to creases the tax burden on capital domestic under- or oversaving, and Andres Velasco income by approximately 3 percent­ thus to inefficient equilibriums. We NBER Working Paper No. 6336 age points. The effects are even show how corrective tax-subsidy December 1997 larger for older households. policies, consisting of taxes on cor­ JEL Nos. E6, H3, H4, H6, H7 The paper also explores the frac­ porate income and on the capital This paper develops a political­ tion of the net worth held by house­ incomes of both residents and non­ economic model of fiscal policy, in holds that are subject to the estate tax residents, can restore efficiency.

------NBER Reporter Winter 1997/8 73. Quantitative Implications This paper replicates a classic whether they were exploitative or of the Home Bias: Foreign study of the American business elite. instead helped to resolve problems Underinvestment, The older study, done a half century with transactions costs. Finally, re­ Domestic Oversaving, ago, reported the composition of busi­ formers pressed for workers' compen­ and Corrective Taxation ness leaders one hundred years ago. sation and laws regulating women~s Assaf Razin, Efraim Sadka, I have drawn a sample of business hours, child labor, and workplace and Cbi-Wa Yuen leaders today to discover how much safety. I examine the impact of pro­ NBER Working Paper No. 6339 the composition of the American busi­ greSSive legislation and discuss the December 1997 ness elite has changed. As in the political economy of its passage. JEL Nos. F21, F25, H25, H30 earlier study, the business elite is compared to a sample of political There is strong evidence about a leaders. I find that democratization of home-court advantage in international the business elite has progressed portfolio investment. One explana­ Historical Perspectives only slightly in the past century, tion for the bias is an information on the Economic nowhere near as much as democrati­ asymmetry between domestic and Consequences of zation of the political elite. Despite foreign investors about the economic Immigration into the the myriad things that have changed performance of domestic firms. This United States in America in the last century, the asymmetry causes two types of distor­ Susan B. Carter and composition of the business elite tions: an aggregate production ineffi­ Richard Sutch appears recognizably the same. ciency, and a production-consump­ NBER Historical Paper No. 106 tion inefficiency, leading to foreign December 1997 underinvestment and domestic over­ JEL Nos. J61, Nll, N12, N31, N32 saving respectively. Such market fail­ Operations of "Unfettered" This paper highlights the distinc­ ures are quite severe, slightly more tive features of the theoretical ap­ so with equity flows than with debt Labor Markets: Exit and Voice in American Labor proach of scholars who analyzed the flows. These inefficiencies nonethe­ impacts of mass migration into the less can be corrected by a mix of tax­ Markets at the Turn of the Century United States in the two decades pre­ subsidy instruments, consisting of ceding World War 1. Broadly speak­ Price V. Fishback taxes on corporate income and on ing, this literature was couched in NBER Historical Paper No. 105 the capital incomes of both residents terms of the "aggregate production November 1997 and nonresidents. When only a par­ function;' and emphasized advancing JEL Nos. N3, J3, K2 tial set of instruments is available, technology, and changes in produc­ however, the prescription for each The American economy at the tivity and factor proportions. There tax instrument can change radically, tum of the century offers an excel­ was also a focus on the close interre­ and even may be reversed, although lent opportunity for studying rela­ latedness among the many diverse the welfare gains can be fairly sub­ tively unregulated labor markets. In elements in the economy. stantial and sometimes close to the this essay I di$cUSS the operation of One notable difference between first best optimum. This partial set of labor markets in the early 1900s. Af­ these historical studies and the recent instruments appears to be more ter examining the mobility of work­ literature on the impacts of immigra­ effective in handling the market fail­ ers, the integration of geographically tion is that the current literature tends ure in the case of equity flows than in dispersed labor markets, and a case to concentrate only on the first-round the case of debt flows. study of the extent of employer consequences. It is easy to show that monopsony, I examine the extent to these will be harmful to resident NBER Historical Papers which workers received compensat­ workers who face direct competition. ing differentials for workplace dis­ Economic historians writing about The American Business amenities, and the extent to which the earlier period of high immigration Elite in Historical competition among employers re­ went beyond the first-round effects. Perspective duced discrimination. During this Taking a long-run perspective, they PeterTemin period, institutions including the identified many aspects of the mass NBER Historical Paper No. 104 company town, company union, and immigration that were beneficial October 1997 share cropping developed. I reexam­ from the point of view of the resident Development of the American Economy ine these institutions to determine population.

74. NBER Reporter Winter 1997/8------NBER Technical Papers (0) = O. It extends the well-known re­ Algorithms for Solving sult that investors are myopic in this Dynamic Models with Horizon Length and model if and only if the utility func­ Occasionally Binding Portfolio Risk tion exhibits constant relative risk Constraints Christian Gollier and aversion. Lawrence J. Christiano and BichardJ.Zeckhauser Jonas D. M. Fisher NBER Technical Working Paper No. 216 NBER Technical Working Paper No. 218 October 1997 Cointegration and Long­ October 1997 JEL Nos. Gll, GI0, D81 Horizon Forecasting JEL Nos. C6, C63, C68 Asset Pricing Peter F. Christoffersen and Economic Fluctuations and Growth Francis X. Diebold We compare the attitude towards We describe and compare several NBER Technical Working Paper No. 217 current risk of two expected-utility­ algOrithms for approximating the so­ October 1997 maximizing investors who are identi­ lution to a model in which inequality JEL No. C5 cal except that the first investor will constraints occasionally bind. We then Economic Fluctuations and Growth live longer than the second. In one evaluate and compare their perform­ of the models we consider, there are We consider the forecasting of co­ ance using various parameterizations two assets at every period. The first integrated variables and show that, at of the Single sector growth model asset has a zero sure return, whereas long horizons, nothing is lost by ig­ with irreversible investment. We the second asset is risky without ser­ noring cointegration when forecasts develop parameterized expectation ial correlation of yields. It is often are evaluated using standard multi­ algorithms which, on the basis of suggested that the young investor variate measures of forecast accuracy. speed, accuracy, and convenience of should purchase more of the risky In fact, simple univariate Box-Jenkins implementation, appear to dominate asset than the old investor under such forecasts are just as accurate. Our re­ the other algOrithms. circumstances. We show that a nec­ sults highlight a potentially important essary and sufficient condition to get deficiency of standard forecast accu­ this property is that the Arrow-Pratt racy measures-they fail to value the index of absolute tolerance (TJ be maintenance of cointegrating rela­ convex. If we allow for a positive tionships among variables-and we risk-free rate, the necessary and suf­ suggest alternatives that explicitly ficient condition is Tu convex, plus Tu do so.

------NBERReporter Winter 1997/8 75. Nonprofit Org. U.S. Postage NBERReporter Paid Boston, MA NATIONAL BUREAU OF ECONOMIC RESEARCH Permit No. 55932

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

Address Correction Requested

L_~"""_ --.--~