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 2008

NBER Reporter Online

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

Suggested Citation: National Bureau of Economic Research (NBER) (Ed.) (2008) : NBER Reporter Online, Volume 2008, NBER Reporter Online, National Bureau of Economic Research (NBER), Cambridge, MA

This Version is available at: http://hdl.handle.net/10419/61991

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 NBER Reporter NATIONAL BUREAU OF ECONOMIC RESEARCH

Reporter OnLine at: www.nber.org/reporter 2008 Number 2

Program Report

IN THIS ISSUE Program Report Law and Economics  Law and Economics

Research Summaries Happiness Economics 7 Understanding Social Networks  Christine Jolls*

NBER Profile 3 Conferences 4 NBER News 24 Program and Working Group Meetings 28 Bureau Books 46 The NBER’s Law and Economics Program studies the effects and causes of legal rules, with a special focus on the foundational legal sub- jects — property law, criminal law, contract law, and tort law — and on the operation of the legal process. In regard to the operation of the legal pro- cess, the Program examines topics such as labor markets for lawyers, litiga- tion dynamics, judicial and agency behavior, and the determinants of leg- islative action. In addition, the Program studies the effects and causes of legal rules across a range of legal subjects beyond property, criminal, con- tract, and tort law, including corporate law, employment law, health care law, social welfare law, family law, bankruptcy law, patent and copyright law, and antitrust law. The Program meets twice per year, once at a mid-year program meet- ing and again at the NBER’s Summer Institute. The Program’s recent spe- cial activities include a project in the area of corporate governance, under the direction of Research Associate Lucian Bebchuk; selected papers from that project are scheduled to appear in a forthcoming issue of the Review of Financial Studies. In this article, I first describe recent research in the foundational legal subjects and then turn to work on the operation of the legal process. I con- clude with an overview of work on the effects and causes of legal rules in corporate law, employment law, health care law, social welfare law, family law, bankruptcy law, patent and copyright law, and antitrust law.

In 2007, we converted from issuing quar- * terly, seasonal issues to a system of numbered Jolls directs the NBER’s Law and Economics Program and is a profes- issues. This is the second issue of 2008 under sor at Yale Law School. In this article, the numbers in parentheses refer to the new numbering system: 2008 Number 2. NBER Working Papers. Steven Shavell, the previous director of the Law and Economics Program, provided extremely helpful feedback on this Program Report.

NBER Reporter • 2008 Number 2  Property Law, Criminal Law, NBER Reporter Contract Law, and Tort Law A fundamental aspect of any legal system is the structure of property rights. Recent work by Oliver Hart (13540) is the latest in an impor- tant line of papers by Hart, Sanford Grossman, and John Moore that examine the effects of a The National Bureau of Economic Research is a private, nonprofit research organization founded in 1920 and devoted to objective quantitative analy- particular conception of property rights empha- sis of the American economy. Its officers and board of directors are: sizing “residual control.” In this view, the owner of an asset retains those rights to the asset that President and Chief Executive Officer — James M. Poterba are not specifically assigned by any existing con- Controller — Kelly Horak tractual commitment. While previous papers in BOARD OF DIRECTORS the series have examined the effects of the resid- ual control rights conception of property rights Chairman — Elizabeth E. Bailey Vice Chairman — John S. Clarkeson in the face of parties’ non-contractible up-front Treasurer — Robert Mednick investments, the recent paper by Hart studies the effects of property rights of this form in a DIRECTORS AT LARGE model with uncertainty of values and costs of Peter Aldrich George C. Eads Laurence H. Meyer Elizabeth E. Bailey Jessica P. Einhorn Michael H. Moskow a good to be traded. Hart’s model suggests that Richard Berner Martin Feldstein Alicia H. Munnell ex-ante contracting over asset ownership can John Herron Biggs Jacob A. Frenkel Rudolph A. Oswald reduce later incentives to engage in hold-up. A John S. Clarkeson Judith M. Gueron Robert T. Parry system that did not provide for property rights Don R. Conlan Robert S. Hamada Richard N. Rosett in the Grossman-Hart-Moore sense (reserving Kathleen B. Cooper Karen N. Horn Marina v. N. Whitman Charles H. Dallara John Lipsky Martin B. Zimmerman residual rights to the state rather than to the property owner) would be inferior within this DIRECTORS BY UNIVERSITY APPOINTMENT framework. George Akerlof, California, Berkeley Marjorie B. McElroy, Duke Also central to the structure of property Jagdish W. Bhagwati, Columbia Joel Mokyr, Northwestern Glen G. Cain, Wisconsin Andrew Postlewaite, Pennsylvania rights is the question of when government may Ray C. Fair, Yale Uwe E. Reinhardt, Princeton obtain ownership of assets from unwilling pri- Franklin Fisher, MIT Nathan Rosenberg, Stanford vate parties. In a recent paper (13564), Steven Mark Grinblatt, California, Los Craig Swan, Minnesota Shavell models the desirability of allowing gov- Angeles David B. Yoffie, Harvard ernment takings of private land by its eminent Saul H. Hymans, Arnold Zellner, Michigan Chicago domain power when the government’s infor- DIRECTORS BY APPOINTMENT OF OTHER ORGANIZATIONS mation about owners’ valuations is imperfect. Jean Paul Chavas, American Agricultural Economics Association Shavell shows that eminent domain becomes Gail D. Fosler, The Conference Board appealing if the number of property owners is Martin Gruber, American Finance Association Timothy W. Guinnane, Economic History Association large, in order to overcome a problem of “hon- Arthur B. Kennickell, American Statistical Association est” holdout. This conclusion holds regardless Thea Lee, American Federation of Labor and of whether the land that the government seeks Congress of Industrial Organizations is a parcel at a fixed location or instead is located William W. Lewis, Committee for Economic Development anywhere in a region. Robert Mednick, American Institute of Certified Public Accountants Angelo Melino, Canadian Economics Association A legal system determines what, if any, con- Harvey Rosenblum, National Association for Business Economics duct should be subject to criminal sanctions John J. Siegfried, American Economic Association and establishes the shape of those sanctions. An extensive recent law and economics litera- The Reporter is issued for informational purposes and has not been reviewed by the Board of Directors of the NBER. It is not copyrighted ture studies the effects of various forms of crimi- and can be freely reproduced with appropriate attribution of source. Please nalization. John Donohue and Justin Wolfers provide the NBER’s Public Information Department with copies of any- (11982) examine the potential deterrent effects thing reproduced. of the ultimate criminal sanction — the death Requests for subscriptions, changes of address, and cancellations should penalty, which is reserved predominantly for be sent to Reporter, National Bureau of Economic Research, Inc., 1050 homicides. While there is some variation in Massachusetts Avenue, Cambridge, MA 02138-5398. Please include the the use of capital punishment, both across time current mailing label. and across states with different legal regimes,

2 NBER Reporter • 2008 Number 2 Donohue and Wolfers conclude that this to deter. Civil liability creates obligations for understanding and improving diverse variation is small when compared with of one private party to another rather forms of regulation of risky products. the large swings in the homicide rate; than liability of an individual to the state, even when the use of the death penalty as in criminal law. Much recent work in The Operation of the increases, the absolute number of execu- tort law has been in the area of prod- Legal Process tions remains quite small. Thus, exist- ucts liability — the liability of firms to ing results linking capital punishment to consumers for injuries sustained in using The legal process involves many play- reductions in the homicide rate prove to a firm’s products. Firms’ liability gen- ers, including lawyers, litigants, judges, be extremely fragile, and the data that are erally will affect the price of products, agencies that administer federal and state presently available do not allow any strong and in recent work Andrew Daughety statutes, and legislatures. In regard to inference about even the sign of the deter- and Jennifer Reinganum examine con- lawyers, a number of recent papers have rent effect of capital punishment. sumer inferences about product qual- analyzed labor markets for lawyers. In Current law subjects criminal sex ity from price variation.1 In Daughety one paper, Jesse Rothstein and Albert offenders to a variety of registration and Reinganum’s model, price is a sig- Yoon provide empirical evidence on the and notification laws that, respectively, nal of product quality. Instead of signal- so-called “mismatch hypothesis” — that require convicted sex offenders to provide ing quality through price, though, firms affirmative action in law school -admis valid contact information to law enforce- may choose to disclose quality directly. sions hurts minority students who attend ment authorities and mandate that infor- Daughety and Reinganum show that more selective schools than they other- mation about sex offenders be made pub- firms may inefficiently choose signaling wise would have and, as a result, experi- lic. In recent work, J.J. Prescott and Jonah over disclosure of product quality when ence lower graduation rates and less suc- E. Rockoff (13803) offer new evidence marginal cost, including the cost of legal cess in passing the bar.2 According to on these laws’ effects. Using fine-grained liability, is increasing in product safety. Rothstein and Yoon, mismatch effects in information on state registration and A. Mitchell Polinsky and Shavell fact are observed only for minority stu- notification laws, Prescott and Rockoff (12776) also examine the relationship dents with the weakest entering creden- present evidence that sex offenses against between products liability law and firms’ tials — students who, without affirmative neighbors declined with the adoption disclosure behavior. In the absence of lia- action, often would not have been admit- of registration laws and that notification bility for product harms, there is a trad- ted to any law school. For minority stu- laws deter potential offenders with no eoff between forced sharing of informa- dents with moderate or strong entering prior record while increasing recidivism tion and firms’ willingness to accumulate credentials, Rothstein and Yoon find no among those who have previously com- information in the first place. When firms evidence of mismatch effects in either mitted sex offenses. face liability irrespective of their degree of graduation or bar passage rates. As the property rights work described fault for the harms caused by products, by Christopher Avery, Richard A. earlier reveals, the value of property rights contrast, mandatory disclosure does not Posner, Alvin E. Roth, and I (13213) is intertwined with the ability of par- affect the information that is disclosed examine another aspect of the labor mar- ties to enter into contracts governing and, thus, does not affect the information ket for lawyers — the market for federal the use of their property. Contract law that is accumulated. judicial law clerks. The hiring process for determines whether and how agreements In many circumstances, products lia- law clerks has long been characterized among parties will be legally enforced. A bility law does make firms, irrespective by the type of unraveling of transaction major part of the function of contract law of their degree of fault, liable to consum- times that has also been observed in many is to determine whether and how to fill ers for the harms caused by products — a other entry-level labor markets. Avery, gaps in parties’ contracts. Are the exist- regime called “strict liability.” This form Posner, Roth, and I surveyed both federal ing gap-filling rules of contract law effi- of liability for product harms is ordinar- appellate judges and clerkship applicants cient? Work by Surajeet Chakravarty and ily justified by reference to informational and found clear evidence of substantial Bentley MacLeod (13960) provides an and cognitive failures of consumers, but in non-adherence to official judicial timing affirmative answer with respect to a range recent work Cass Sunstein and I (11738) guidelines intended to prevent the hir- of contract law rules — including the explore the potential effects of respond- ing of clerks before a designated time. We important rule setting “expectation dam- ing to these failures, not with strict lia- describe, however, ways that judges and ages” as the standard measure of dam- bility, but rather with requirements to clerks might settle at an equilibrium level ages for breach of contract — in a model engage in “debiasing” communications of imperfect, but still meaningful, adher- informed by standard industry contract- designed to reduce consumer errors. We ence to the designated start date regime. ing practices. distinguish such debiasing communica- Lawyers, along with the clients they Tort law imposes civil — as distin- tions from conventional informational represent, resolve many of the lawsuits guished from criminal — liability on cer- mechanisms and suggest that legal debi- in which they are involved without any tain forms of behavior that society wishes asing strategies hold significant promise recourse to the courtroom. In fact, the

NBER Reporter • 2008 Number 2 3 overwhelming majority of lawsuits are set- activity and stock market development parency of corporate financial informa- tled by the parties prior to trial. An impor- across nations. Using measures of the legal tion. Benjamin Hermalin and Michael tant set of law and economics papers has powers of government agencies charged Weisbach (12875) offer a model in which modeled the bargaining process between with enforcement of securities laws, they such increased transparency may, con- opposing parties over whether to agree find little relationship. Howell Jackson trary to its presumed intent, reduce firm to a pretrial settlement. In recent work, and Mark Roe, by contrast, use measures profits and increase executive compensa- Yasutora Watanabe fits a dynamic model of budgets and staffing of such securities tion. Increased transparency in Hermalin of such litigant bargaining to data on the enforcement agencies in several samples and Weisbach’s model may also increase time, mode, cost, and terms of settlement of about 40 nations and find significant the rate of turnover of chief executive of a large set of legal disputes.3 The mod- association between enforcement agen- officers. el’s fit with the data suggests that asym- cies’ resources and those nations’ financial Backdating of stock option grants has metric initial beliefs and the opportunity market outcomes.4 been a source of substantial Securities and for learning over time are important fea- A third type of body engaged in Exchange Commission activity, as well tures of litigation-settlement bargaining. law making and law enforcement, along- as private litigation, in recent years. In a Turning to judicial behavior, a line side courts and government agencies, is pair of studies (12771, 12811), Bebchuk, of recent papers by Andrei Shleifer and the legislature. Legislative behavior may Yaniv Grinstein, and Urs Peyer docu- coauthors has examined common law be influenced by, among other things, ment both the frequency and the corpo- decisionmaking in a system of judge- campaign contributions and the informa- rate governance correlates of option grant made law. The first paper in this series, by tion that legislators receive from lobbyists backdating and other forms of opportu- Shleifer and Nicola Gennaioli, models the and other interest-group actors. Recent nistic option grant timing. The correlates evolution of legal rules in common law work by Charles Cameron and John de of such behavior include a smaller fraction courts (11265); in the most recent work, Figueiredo offers empirical evidence on of independent members of the board of Shleifer, Anthony Nisbett, and Richard the second, information-mediated type directors (for both CEO option grants Posner turn to a specific set of decided of influence.5 They find strong evidence and director option grants) and longer cases to examine the evolution of a par- that interest-group expenditures on infor- CEO tenure (for CEO option grants). ticular legal rule in action (13856). The mational lobbying vary with legislative Corporate law structures the exer- theoretical model provides a foundation budget cycles — rather than electoral cise of shareholder votes, and recent work for the evolutionary adaptability of com- cycles — as well as with the ideological by Yair Listokin studies the effects of the mon law, while in the actual set of decided distance between the interest group and existing structure on voting outcomes.6 cases there was no evidence of conver- the party in legislative power. Cameron Listokin examines the results of share- gence to any stable resting point. and de Figueiredo’s empirical results are holder voting on management-sponsored Judges often make decisions in groups, consistent with leading theoretical mod- resolutions in “close” cases, in which man- and in recent work Edward Glaeser and els on informational lobbying. agement’s share of the vote is within 10 Sunstein (13687) model the evolution of percentage points of the cutoff point for group members’ views as a result of delib- Corporate Law, Employment success (which is typically 50 percent). He erations. One common effect of group Law, Health Care Law, finds that management overwhelmingly deliberation is polarization, in which indi- Social Welfare Law, Family wins these close votes. His conclusion is viduals’ pre-deliberation views become that the existing structure of shareholder more extreme as a result of the delibera- Law, Bankruptcy Law, voting is not effective in producing voting tions. Glaeser and Sunstein show that Patent and Copyright outcomes that mirror underlying share- polarization may, but need not, follow Law, and Antitrust Law holder preferences. from rational Bayesian updating by group Alongside the study of the effects of members. An extremely active area of research corporate law, recent research examines Glaeser and Sunstein’s analysis has considers the effects and causes of corpo- the causes of corporate law. In a model of implications for decisionmaking not rate law. Papers in this area have been fea- interest group lobbying in the context of only by judges but also, as they note, tured both in regular Law and Economics corporate lawmaking (13702), Bebchuk by members of decisionmaking bod- Program sessions and in sessions con- and Zvika Neeman identify a range of cir- ies within government agencies. Other ducted by the Corporate Governance cumstances under which such lobbying recent work in law and economics has Project. leads to an inefficiently low level of inves- studied a variety of effects of government In studying the effects of corpo- tor protection. Their model indicates that agency action. For instance, Rafael La rate law rules, much recent attention observed correlations between countries’ Porta, Florencio Lopez-de-Silanes, and has been paid to legal rules — many of levels of investor protection and these Shleifer (9882) examine the relationship them enacted in the wake of the collapse countries’ economic performance may between government agency enforcement of Enron — requiring increased trans- reflect the effects of the second factor on

 NBER Reporter • 2008 Number 2 the first as well as (what has been empha- ing field spanning an enormous range of ity, have not achieved financial indepen- sized by the existing literature) the effects research questions. One important ques- dence. Steven D. Levitt and Joseph J. of the first factor on the second. tion concerns the effects of legal rules Doyle (12519) analyze the likely effects In addition to managers and share- intended to increase the availability to of an important mandate intended to holders, firms are populated by employ- patients of health-care-related informa- protect children — the requirement that ees, whose relationship with their firm tion. Recent work by M. Kate Bundorf, children under specified ages (which vary is regulated by employment law. Many Natalie Chun, Gopi Shah Goda, and across states in the United States) ride provisions of employment law mandate Daniel P. Kessler (13888) studies the in child safety seats. Using several datas- that particular benefits be provided to effects of mandated health care provider ets containing information on auto acci- employees. One very economically sig- “report cards” that include detailed infor- dent injuries and types of child restraint nificant mandate in the employment con- mation on medical outcomes. Using a in use, Levitt and Doyle find that stan- text is the Social Security program, which unique identification strategy, the authors dard lap-and-shoulder seat belts perform requires payroll deductions to fund gov- find that mandated birth-rate information as well as child safety seats in prevent- ernment benefits upon disability or retire- for infertility treatment centers increases ing serious injury for children aged 2 ment of employees. Research by David the market share of centers with high suc- through 6. Child safety seats, however, Autor and Mark Duggan (12436) ana- cess rates. are more effective in preventing less seri- lyzes upward trends in the level of Social Medical malpractice is another major ous injury for this age group. Thus, Levitt Security disability payments in recent aspect of health care law (as well as an and Doyle’s findings suggest that existing decades. Autor and Duggan link these aspect of tort law). In a recent paper, Janet mandates of child safety seats have some increases to changes in legal qualification Currie and MacLeod (12478) study the effect in reducing injury, though perhaps standards, in real benefit levels, and in the effects of medical malpractice law reform not the primary sort of effect legislators size of the workforce. within the field of obstetrics — a branch intended. Recent work by Louis Kaplow of medicine thought to have been par- Family law is another active area of (12452) analyzes Social Security using a ticularly hard hit by the “liability crisis” law and economics research, and changes model that incorporates myopia on the in medical malpractice law. Thus, Currie in divorce law in recent decades have part of employees. Kaplow studies the and MacLeod examine how reform in this been a particular focus of study. Betsey effect of Social Security on labor sup- area has affected the types of procedures Stevenson examines the effect on invest- ply in the presence of myopic employ- performed in childbirth and the ensuing ment in marriage-specific capital of the ees who give excessive weight to present health outcomes. They introduce a model move to divorce “on demand”.8 She finds payroll deductions that finance distant of physician behavior that allows for dif- that this change reduced the number future benefits. Kaplow’s model shows ferential effects across patient characteris- of children produced by the marriage, that even with myopia, Social Security tics and that provides a new way to model spouses’ willingness to invest in their part- may cause labor supply either to rise or to alternative liability rules. Empirically, ners’ education, and spouses’ likelihood of fall depending on the curvature of indi- Currie and MacLeod find that while some choosing to have one partner remain out viduals’ utility as a function of consump- medical malpractice reforms have posi- of the labor force. tion and on whether individuals’ myopia tive effects on health outcomes, imposing The effects of bankruptcy law have extends to labor supply choice as well as caps on non-economic damages has nega- been the subject of several recent papers. savings decisions. tive effects. Within the United States, both federal Another important federal employ- Patricia Born, W. Kip Viscusi, and and state law are relevant to the bank- ment mandate involves medical leave Tom Baker (12086) also examine the ruptcy process. Recent work by Edward under the Family and Medical Leave Act effects of medical malpractice law reform. Morrison shows that most distressed of 1993 (FMLA), which I study in a While a substantial literature examines firms use state law to liquidate or reor- recent paper.7 Exploiting variation across the effects of such reform on insurers’ ganize and that the attractiveness of state states in the presence or absence of man- incurred losses, Born, Viscusi, and Baker law varies in predictable ways with its dated medical leave at the state level prior are able to look at longer-term effects. nature — and particularly with the degree to the FMLA’s enactment, I find positive They find these longer-term effects to be of transparency of the state insolvency employment effects of mandated medi- especially significant for non-economic process.9 Because the choice between fed- cal leave for individuals with disabili- damage caps. eral and state regimes turns on a compari- ties. I suggest that this result may reflect Another important category of law is son of the two, reform of federal bank- the absence of hiring disincentives from social welfare law, which mandates ben- ruptcy law may have unanticipated effects mandated medical leave given the limited efits for children, for individuals who if some firms switch away from the federal observability of many leave-necessitating at the age of majority are not able to regime in response to the reform. medical conditions at the time of hiring. live independently, and for adults who, Around the world, bankruptcy Health care law is a rapidly grow- though free of any recognized disabil- regimes exhibit a range of features, and a

NBER Reporter • 2008 Number 2 5 recent paper by Simeon Djankov, Caralee (12132). Philipson and Posner address a 3 Y. Watanabe, “Learning and McLiesh, and Shleifer explores how the basic question: should antitrust law apply Bargaining in Dispute Resolution: Theory legal rights of creditors in bankruptcy in the same way to non-profit firms as to and Evidence from Medical Malpractice relate to the level of credit extended in for-profit firms? In their model, not only Litigation,” NBER Summer Institute Law a country. In a sample of 129 countries is there no ground for lesser antitrust scru- and Economics Workshop 2006. over a quarter century, the authors find a tiny of non-profit firms, but also in some 4 H. Jackson and M. Roe, “Public and positive effect of legal protection of credi- circumstances the welfare gains from anti- Private Enforcement of Securities Laws: tors on the level of credit extended. That trust regulation are greater than in the Resource-Based Evidence”, NBER Summer effect is observed both in the cross section case of for-profit firms. Institute Law and Economics Workshop and longitudinally when a given country In antitrust law as in many other areas 2008. expands the legal protection it affords to of law, the social welfare effects of legal 5 J. de Figueiredo and C. Cameron, creditors. regulation depend in part on the character “Endogenous Cost Lobbying: Theory and Patent and copyright law govern the of market relationships in the absence of Evidence,” NBER Law and Economics circumstances under which inventors and regulation. Within the antitrust domain, Program Meeting 2007. creators will be awarded special prop- a major question for researchers has been 6 Yair Listokin, “Management Always erty rights. With respect to inventions, the degree to which incumbent monop- Wins the Close Ones,” NBER Summer patent law requires that an invention be olists can succeed in excluding rivals Institute Law and Economics Workshop “non-obvious” in order to receive patent through contracts with downstream buy- 2007. protection. Nisvan Erkal and Suzanne ers. Complementing a line of important 7 C. Jolls, “Mandated Medical Leave in Scotchmer suggest that models in which theoretical models analyzing this ques- the Workplace,” NBER Summer Institute invention is a product solely of invest- tion, Claudia Landeo and Kathryn Spier Law and Economics Workshop 2006. ments in research and development fail offer recent experimental evidence on the 8 B. Stevenson, “The Impact of Divorce to capture the full scope of the non-obvi- use of exclusionary contracts.11 They find Laws on Marriage-Specific Capital,” ousness requirement.10 In their analysis, that such exclusionary contracts do occur NBER Summer Institute Law and invention is a product not only of finan- both when downstream buyers are not Economics Workshop 2006. cial investments but also of scarce, cre- able to communicate with one another 9 E. Morrison, “Bargaining Around ative ideas. When an idea comes along, and when they can engage in such com- Bankruptcy: Small Business Distress and a potential innovator faces a choice of munication. Their results complement State Law,” NBER Law and Economics whether to invest in the idea or to take the theoretical models suggesting that regula- Program Meeting 2007. chance that the market niche in question tion of exclusive dealing contracts may be 10 N. Erkal and S. Scotchmer, “Scarcity may be filled by someone else who comes welfare enhancing. of Ideas and Options to Invest in R&D,” along later with a substitute idea. In Erkal NBER Law and Economics Program and Scotchmer’s model, conditioning the 1 A. Daughety and J. Reinganum, Meeting 2008. reward for innovation on the level of non- ”Products Liability, Signaling, and 11 C. Landeo and K. Spier, “Naked obviousness is shown to be optimal. Disclosure,” NBER Summer Institute Law Exclusion: An Experimental Study of A longstanding focus within law and Economics Workshop 2007. Contracts with Externalities,” NBER and economics has been antitrust law. 2 J. Rothstein and A. Yoon, “Mismatch Summer Institute Law and Economics In recent work, Tomas Philipson and in Law School,” NBER Law and Workshop 2008. Richard Posner continue in this line Economics Program Meeting 2007.

 NBER Reporter • 2008 Number 2 Research Summaries

Happiness Economics

David G. Blanchflower*

Despite Thomas Carlyle’s claim, when research, primarily with Andrew Oswald interpersonal comparisons of well-being. he was arguing that slavery was morally at the University of Warwick, but also with Psychologists, however, view it as natural superior to the market, economics is no David Bell at the University of Stirling; that a concept such as happiness should longer the dismal science.1 A growing body Chris Shadforth at the Bank of England, be studied in part by asking people how of literature in the economics of happi- and Richard Freeman from Harvard. they feel. One definition of happiness is ness and mental well-being has emerged. Typical questions are: 1) Happiness the degree to which an individual judges It is now fashionable to try to understand — for example from the General Social the overall quality of his or her life as the pursuit of happiness, and, after a long Survey, Taken all together, how would you favorable. As a validation of the answers delay, the ideas promoted originally by say things are these days — would you say to recorded happiness levels, it turns out Richard Easterlin are attracting worldwide that you are very happy, pretty happy, or not that answers to happiness and life satis- attention.2 There is even a World Database too happy? 2) Life satisfaction — for exam- faction questions are correlated with: 1) of Happiness. ple from the European Eurobarometer objective characteristics such as unemploy- Anyway, I came to the topics of hap- Surveys, On the whole, are you very satis- ment; 2) assessments of the person’s hap- piness and well-being as a labor economist fied, fairly satisfied, not very satisfied, or not piness by friends and family members; 3) who had mostly worked on wages, and at all satisfied with the life you lead? 3) assessments of the person’s happiness by who early on was struck by the stability Psychological health and mental his or her spouse; 4) heart rate and blood- of the Mincerian earnings function across strain — for example from the British pressure measures of response to stress; 5) time and space. The basic structure of a log Household Panel Survey, Such as the GHQ the risk of coronary heart disease; 6) dura- earnings equation, no matter what dataset score, which amalgamates answers to ques- tion of authentic or so-called Duchenne was used and what country it is estimated tions about how well people have been sleep- smiles (a Duchenne smile occurs when for, has a similar structure. It turns out that ing, their level of confidence, feelings of both the zygomatic major and obicularus there are patterns in the well-being data. I depression, among others. 3 orus facial muscles fire, and human beings am struck by the fact that there is a great The micro data on happiness are eas- identify these as “genuine” smiles); 7) skin- deal of stability in happiness and life-satis- ily obtainable from most data archives resistance measures of response to stress; faction equations, no matter what country including ICPSR for the GSS and the 8) electroencephelogram measures of pre- we look at, what dataset or time period, Eurobarometers, the Data Archive at frontal brain activity. whether the question relates to life satisfac- the University of Essex and ZACAT in It is apparent that most people are tion or happiness, and how the responses Germany for the Eurobarometers, ISSP, happy (or, more precisely, mark themselves are coded whether in three, four, five or European Social Survey, BHPS, GSOEP, fairly high up on a scale). This finding even as many as ten categories. European Q uality of Life Survey, European has subsequently been replicated in many In general, economists have focused Social Surveys and so on. Life satisfac- datasets over many time periods and for on modeling three fairly simple questions tion data are also now available annually numerous countries. For example, in the on life satisfaction and happiness, and from the Latinobarometers, while happi- United States in 2006 only 13 percent of that is what I have done mostly in my ness data are also available annually in the people in the GSS said they were not very Asianbarometers. Several of the data series happy, 56 percent were pretty happy, and *Blanchflower is a Research Associate in the extend back at least to the early 1970s. 31 percent very happy. In Eurobarometer NBER’s Program on Labor Studies and the Some are panels (BHPS, GSOEP). 67.2 for April-May 2007 (ICPSR#21160) Bruce V. Rauner Professor of Economics at Economists have had longstand- for the European Union (EU) 15 in 2007, Dartmouth College. His Profile appears later in this issue. ing reservations about the reliability of for example, 3 percent said they were not

NBER Reporter • 2008 Number 2 7 at all satisfied, while 12 percent were not an unpleasant state. Encouragingly, there on happiness of GDP per capita and inter- very satisfied, 60 percent fairly satisfied, are many similarities in the findings. For est rates.15 I find, conventionally, that both and 24 percent very satisfied. In the 2005 example, both the U-index and conven- higher unemployment and higher infla- Latinobarometers, which also asked the tional measures show higher levels of well- tion lower happiness. Interest rates are same 4-step life satisfaction question in being among wealthier, higher educated also found to enter happiness equations eighteen Latin American countries, 4.6 and older individuals. One attraction of negatively. Changes in GDP per capita percent said they were not at all satisfied, the evaluated time use data, though, is that have little impact on more economically 25.4 percent were not very satisfied, 39.7 it can be used to understand why some developed countries, but do have a positive percent were fairly satisfied, and 30.3 per- groups are happier than others. impact in the poorest countries — consis- cent very satisfied. In the United States happiness has tent with the Easterlin hypothesis. I find There are a number of common pat- trended downward over time, but the pic- that unemployment depresses well-being terns in the determinants of happiness, ture is rather more mixed in Europe: using more than inflation. The least educated which have been replicated in a number the Eurobarometer life satisfaction ques- and the old are more concerned about of other papers. Happiness and life sat- tions, when an ordered logit is estimated unemployment than inflation. Conversely, isfaction tends to be higher among a) with controls for education, age, gender, the young and the most educated are more women, b) people with lots of friends, c) schooling, marital status, and labor mar- concerned about inflation. An individu- the young and the old, d) married and ket status, nine countries have positive al’s experience of high inflation over their cohabiting people, e) the highly educated, time trends (Denmark; Finland; France; adult lifetime lowers their current happi- f ) the healthy, g) those with high income, Italy: Luxembourg; Netherlands; Spain; ness over and above the effects from cur- h) the self-employed, i) people with low Sweden; and the United Kingdom.) rent inflation and current unemployment. blood pressure, j) those who have sex at Austria and Ireland have no significant Unemployment appears to be more costly least once a week with the same partner, time trend while Belgium, Germany, than inflation in terms of its impact on k) right-wing voters, l) the religious, m) Greece, and Portugal have significant wellbeing. members of non-church organizations, n) downward trends. There is evidence from Oswald and I found that the well-being volunteers, o) those who take exercise, and developing countries of a steeper upward of the young had risen both in the United p) those who live in western countries.4 trend in happiness. This is especially appar- States and Europe.16 Explaining why was The self-employed especially value their ent in Latin America and among the eight difficult. This is still the case: in the GSS the independence.5 It turns out that the main Eastern European countries that joined time trend is positive for those aged under findings from responses on both happiness the European Union in 2004.11 30 and flat for those aged 30 and higher. and life satisfaction are also broadly repli- There is evidence though of a down- In Europe using pooled Eurobarometers cated in data on unhappiness, hyperten- ward trend in happiness for women and there is a significant upward time trend for sion, stress, depression, anxiety, and pain an upward trend for men in the United both but much stronger for the young. We from a considerable number of cross-coun- States12. Betsey Stevenson and Justin ruled out that it was explained by a decline try data sources.6 Happy people are less Wolfers confirmed this finding for the in the chance of war with the Eastern bloc, likely to commit suicide.7 United States and also found that wom- falling discrimination, changing education There is also evidence of adaptation. en’s declining relative well-being is found and work, or the rise in youth-oriented Good and bad life events wear off, at in multiple countries, datasets, and mea- consumer goods. The paper demonstrated least partially, as people get used to them. sures of subjective well-being, and is perva- that most of the increase is to be found in Oswald and Powdthavee provide longitu- sive across demographic groups.13 Relative the group who were unmarried. dinal evidence that people who become declines in female happiness have eroded a In a 2004 paper, Oswald and I esti- disabled go on to exhibit considerable gender gap in happiness in which women mated the dollar values of events like unem- recovery in mental well-being.8 In fixed- in the 1970s typically reported higher sub- ployment and divorce.17 They are large. effects equations they estimate the degree jective well-being than did men. These For the typical individual, a doubling of of hedonic adaptation at — depending on declines have continued and a new gender salary makes a lot less difference than life the severity of the disability — approxi- gap may be emerging, one with higher sub- events like marriage or unemployment. mately 30 percent to 50 percent. jective well-being for men. A lasting marriage (compared to widow- I recently compared the results using A small literature, begun by Rafael Di hood as a “natural” experiment), for exam- data on life satisfaction and happiness with Tella and his co-authors, has found that ple, is estimated to be worth $100,000 those from evaluated time use9 and in par- both unemployment and inflation lower a year. Further calculation suggests that ticular the U-index (for “unpleasant” or happiness.14 This is true even after con- to “compensate” men exactly for unem- “undesirable”) as propounded by Krueger trolling for country fixed effects and year ployment would take a rise in income of et al for use in National Time Accounting.10 effects. My paper extends the literature by $60,000 per annum, and to “compensate” The U-index is designed to measure the looking at more countries over a longer for being black would take $30,000 extra proportion of time an individual spends in time period. It also considers the impacts per annum. Oswald and Powdthavee find

 NBER Reporter • 2008 Number 2 that using GHQ mental distress as the per capita, the unemployment rate, or the In a pair of articles Oswald and I stud- measure of well-being, the hedonic com- employment rate.24 ied well-being in Australia along with a pensation annual amount in the first year Consistent with other research, published response by Andrew Leigh and for the death of a child might be of the Oswald and I recently documented, using Justin Wolfers.27 According to the HDI, order of £100,000 ($200,000).18 These are evidence on over two million people, Australia now ranks third in the world. all large sums, and in a sense reflect the low that psychological well-being is U-shaped That places the country above all the other (happiness) value of extra income. Money, through life.25 A difficulty with research on English-speaking nations. This article it seems, does not buy more sexual partners this issue is that there are likely to be omit- raises questions about that assessment. It or more sex.19 ted cohort effects (earlier generations may reviews the new work on happiness eco- The data show that richer people are have been born in, say, particularly good nomics, considers implications for poli- happier and healthier, but can we be sure or bad times). First, using data on 500,000 cymakers, and examines where Australia about causality? As in other areas of eco- randomly sampled Americans and West lies in international subjective well-being nomics, this is both of central importance Europeans, the paper designs a test that rankings. Using new ISSP data on approx- and difficult to establish beyond all doubt. can control for cohort effects. Holding imately 50,000 randomly sampled indi- One attempt, by Jonathan Gardner and other factors constant, we show that a typ- viduals from 35 nations, the article shows Oswald, found that Britons who receive ical individual’s happiness reaches its min- that Australia lies close to the bottom of lottery wins of between £1,000 and imum on both sides of the Atlantic and an international ranking of job satisfaction £120,000 go on, compared to those who for both males and females in middle age. levels. Among a sub-sample of English- receive tiny wins, to exhibit better psycho- Second, evidence is provided for the exis- speaking nations, where a common lan- logical health.20 But Oswald and I found tence of a similar U-shape through the life- guage should help subjective well-being that individuals in the United States were course in East European, Latin American, measures to be reliable, Australia performs found to be less happy if their incomes and Asian nations. Third, a U-shape in age fairly poorly on a range of happiness indi- are far above those of the poorest people. is found in separate well-being regression cators. Moreover, Australia has the high- However, people do appear to compare equations in 72 developed and developing est overall suicide rate. This appears to be themselves more with well-off families, so nations. Fourth, using measures that are a paradox. that perhaps they get happier the closer closer to psychiatric scores, we document a The literature on the economics their income comes to that of rich people comparable well-being curve across the life of well-being is currently growing at a around them. Relative income certainly cycle in two other datasets: 1) in GHQ- remarkable rate. If one takes the view appears to matter. Erzo Luttmer finds, for N6 mental health levels among a sample that human happiness is ultimately the the United States, that higher earnings of of 16,000 Europeans, and 2) in reported most important topic in social science, neighbors are associated with lower lev- depression-and-anxiety levels among one perhaps — after decades where economists els of self-reported happiness, controlling million UK citizens. Fifth, we discuss some lagged behind other social scientists — this for an individual’s own income.21 Alberto apparent exceptions, particularly in devel- should not be surprising. However, we still Alesina and his co-authors find, using a oping nations, to the U-shape. Sixth, we have a great deal to learn. Is the Easterlin sample of individuals across the United note that American male birth-cohorts Paradox — that happiness at a national States (1981–96) and Europe (1975–92) seem to have become progressively less level does not increase with wealth once that individuals have a lower tendency to content with their lives. More, truly lon- basic needs are fulfilled — correct or only report themselves as happy when inequal- gitudinal research on this topic would be an approximation to the truth? Are rela- ity is high, even controlling for individual valuable. tive-wage effect comparisons always harm- income. The effect is stronger in Europe In surveys of well-being, countries ful to people? How should we think about than in the United States.22 such as Denmark and the Netherlands the connections between mental health Life satisfaction scores predict reason- emerge as particularly happy while nations and what economists have traditionally ably well the scale of the flow of workers like Germany and Italy report lower lev- called utility? What are the deep links coming to the United Kingdom from the els of happiness. But are these kinds of between money and happiness? The next eight Accession countries of the Czech findings credible? Oswald and I provide decade is likely to see a great deal of work Republic, Estonia, Hungary, Latvia, some evidence suggesting that the answer on these important topics. Lithuania, Poland, Slovakia, and Slovenia. is yes.26 Using data on 16 countries, we These countries joined the European show that happier nations report systemat- 1 “…I should say, like some we have heard Union in 2004 and only the United ically lower levels of hypertension. As well of, no, a dreary, desolate, and indeed, quite Kingdom and Ireland gave them the right as potentially validating the differences abject and distressing one, what we might to work. In these countries the levels of in measured happiness across nations, call, by way of eminence, the dismal science” happiness are low.23 The propensity to this suggests that blood-pressure readings Thomas Carlyle, 1849. migrate is more highly correlated with might be valuable as part of a national 2 R.A. Easterlin, “Does Economic life satisfaction scores than it is with GDP well-being index. Growth Improve The Human Lot?

NBER Reporter • 2008 Number 2 9 Some Empirical Evidence” in Nations Accounting and Subjective Well-Being, United States.” and Households in Economic Growth: A.B. Krueger, ed. University of Chicago 18 A.J. Oswald and N. Powdthavee, Essays in Honor of Moses Abramowitz, Press. “Death, Happiness, and the Calculation of P. A. David and M. W. Reder, eds. New 7 H.H. Koivuma, R. Honkanen, H. Compensatory Damages”, Journal of Legal York: Academic Press (1974), and R.A. Viinamaeki, K. Heikkalae, J. Kaprio, Studies, forthcoming. Easterlin, “Will Raising the Incomes of All and M. Koskenvuo, “Life Satisfaction 19 D.G. Blanchflower and A.J. Increase the Happiness of All?” Journal of and Suicide: a 20-Year Follow-Up Study”, Oswald, “Money, Sex and Happiness”, Economic Behavior and Organization, 27 American Journal of Psychiatry, 1589(3), Scandinavian Journal of Economics, (June1995), pp.35–48. 2001, pp. 433–9. 106(3), 2004, pp 393–415. 3 For economists who want a readable 8 A.J. Oswald and N. Powdthavee, 20 J. Gardner and A.J. Oswald, “Money way into the literature, introductions to it “Does Happiness Adapt? A Longitudinal And Mental Well-Being: A Longitudinal can be found in sources such as A.J. Oswald, Study Of Disability with Implications for Study of Medium Sized Lottery Wins”, “Happiness And Economic Performance”, Economists and Judges”, Journal of Public Journal of Health Economics, 26, 2007, Economic Journal, 107 November1997, Economics, 92, (2008), pp.1061–77. pp.49–60. pp. 1815–31; B.S. Frey and A. Stutzer, 9 D.G. Blanchflower, “International 21 E. Luttmer, “Neighbors As Negatives; Happiness and Economics, Princeton Evidence on Well-Being.” Relative Earnings and Well-Being”, and Oxford: Princeton University Press, 10 A.B. Krueger, D. Kahneman, D. Quarterly Journal of Economics, August (2002); and A.E. Clark, P Frijters, and Schkade, N. Schwarz, and A.A. Stone, 2005, 120(3), pp. 963–1002. M.A. Shields, “Relative Income, Happiness, “National Time Accounting: The Currency 22 A. Alesina, R.Di Tella, and R.J. and Utility: An Explanation for the of Life”, in National Time Accounting and MacCulloch, “Inequality and Happiness: Easterlin Paradox And Other Puzzles”, Subjective Well-Being, A.B. Krueger, ed. Are Europeans and Americans Different?” Journal of Economic Literature, 46(1), University of Chicago Press, forthcoming. Journal of Public Economics, 88, 2004, (2008), pp.95–144. 11 D.G. Blanchflower and C. Shadforth, pp. 2009–42. 4 D.G. Blanchflower and A.J. Oswald, “Fear, Unemployment, and Migration”, 23 D.G. Blanchflower, “Unemployment, “Well-Being Over Time in Britain and NBER Working Paper No. 13506, Well-Being and Wage Curves in Eastern the United States”, Journal of Public September 2007; and D.G. Blanchflower, and Central Europe”, Journal of Japanese Economics, Volume 88, Issues 7–8, July “International Evidence on Well-Being.” and International Economies, 15(4), 2004, pp. 1359–86; and P. Dolan, T. 12 D. G. Blanchflower and A.J. Oswald December 2001, pp. 364–402. Peasgood, and M. White, “Do We Really “Well-Being Over Time in Britain and the 24 D.G. Blanchflower and C. Shadforth Know What Makes Us Happy? A Review United States.” “Entrepreneurship in the UK”. of the Economic Literature on the Factors 13 B. Stevenson and J. Wolfers, “The 25 D.G. Blanchflower and A.J. Oswald, “Is Associated with Subjective Well-Being”, Paradox Of Declining Female Happiness”, Well-Being U-Shaped over the Life Cycle?” Journal of Economic Psychology, 29, mimeo, University of Pennsylvania, 2007. Social Science and Medicine, 66(8), pp. (2008), pp. 94–122. 14 R. Di Tella, R.J. MacCulloch, and A.J. 1733–1749, April 2008. 5 D.G. Blanchflower and A.J. Oswald, Oswald, “Preferences Over Inflation and 26 D.G. Blanchflower and A.J. Oswald, “What Makes An Entrepreneur?” Unemployment: Evidence from Surveys “Happiness and Hypertension Across Journal of Labor Economics, January of Happiness”, American Economic Nations”, Journal of Health Economics, 1998, 16(1), pp. 26–60; D.G. Review, 91, 2001, pp. 33–41. See also 27(2), March 2008, pp. 218–33. Blanchflower, “Self-Employment: More J. Wolfers, “Is Business Cycle Volatility 27 D.G. Blanchflower and A.J. Oswald, May Not Be Better”, Swedish Economic Costly? Evidence from Surveys of Subjective “Happiness and the Human Development Policy Review, 11(2), Fall 2004, pp. Wellbeing”, International Finance, 6:1, Index: the Paradox of Australia,” The 15–74; and D.G. Blanchflower and 2003, pp.1–26. Australian Economic Review, 38(3), C. Shadforth, “Entrepreneurship in 15 D.G. Blanchflower, “Is Unemployment September 2005, pp. 307–18; D.G. the UK”, Foundations and Trends in More Costly Than Inflation?” NBER Blanchflower and A.J. Oswald, “On Leigh- Entrepreneurship, 3(4), (2007), pp. Working Paper No.13505, September Wolfers and Well-Being in Australia”, The 1–108. 2007. Australian Economic Review, 39(2), 6 D.N.F. Bell and D.G. Blanchflower, 16 D.G. Blanchflower and A.J. Oswald, June 2006, pp. 185–6; and A. Leigh and “The Scots May Be Brave But They Are “The Rising Well-Being Of The Young”, J. Wolfers, “Happiness and the Human Neither Healthy Nor Happy”, Scottish in Youth Employment and Joblessness in Development Index: Australia is not a Journal of Political Economy, 54(2), Advanced Countries, D.G. Blanchflower Paradox”, Australian Economic Review, pp. 151–307, May 2007; and D.G. and R. B. Freeman, eds., University of 39(2), June 2006, pp. 176–84. Blanchflower, “International Evidence Chicago Press, 2000. on Well-Being”, presented to an NBER 17 D. G. Blanchflower and A.J. Oswald conference, forthcoming in National Time “Well-Being Over Time in Britain and the

10 NBER Reporter • 2008 Number 2 Understanding Social Networks

Dora L. Costa*

Heroes and Cowards: income inequality were less likely to be civ- and leadership. But loyalty to comrades The Social Face of War ically engaged. These results contributed extended only to men like themselves — in to a growing literature in economics docu- ethnicity, social status, and age. When are men willing to sacrifice for menting the disturbing fact that people are Sacrifices for the common good are the common good? What are the ben- less likely to be “good citizens” when they costly. Standing by their comrades raised efits to men of friendship? How do com- live in more diverse communities. men’s chances of dying. What then are munities deal with betrayal? And what Our early work on community par- the benefits to men of friendship? We are the costs and benefits of being in a ticipation attracted academic and popular can reply by looking at who survived the diverse community? Matthew Kahn and media attention. Although we were flat- extreme conditions of Civil War POW I answer these questions in an interdisci- tered, we were aware that our measures camps.3 We can see the effects of age, social plinary book, Heroes and Cowards: The of “civic engagement” bordered on “small status, rank, camp population, and the Social Face of War (forthcoming, Princeton potatoes.” We were examining low stakes presence of own officers on survival. We University Press for NBER). Building on a outcome measures such as entertaining in can also see that the fellowship of their series of joint NBER Working Papers, we the household, joining neighborhood asso- comrades helped soldiers survive POW weave a single narrative from the life histo- ciations, and volunteering for local clubs. camps and, the deeper the strength of ties ries of 41,000 Union Army soldiers, diaries In the summer of 2001, we realized between men, the higher their probabil- and letters, and government documents. that the American Civil War, 1861 to ity of survival. Ties between kin and ties One summer we both read Robert 1865, provided the ideal “laboratory.” The between comrades of the same ethnicity Putnam’s thought-provoking book setting was high stakes — roughly one out were stronger than ties between other men Bowling Alone (2000). We were fasci- of every six Union Army soldiers died dur- from the same company. nated by Putnam’s account of the decline ing the war. Unlike people in civilian life If loyalty toward your own kind is in American civic engagement over time. today, Union Army soldiers could not pick admirable, how do communities deal with Putnam emphasized the growing popu- and choose their communities. Even when betrayal? In the Civil War companies were larity of television as a pivotal cause of the they signed up with friends, some men raised locally and hometowns were well decline in community participation, but ended up in homogenous units and oth- aware of who was a “coward” and who we wondered whether an unintended con- ers in heterogenous units and they could was a “hero” during the war. Some towns sequence of the rise of women working in not leave their units unless they deserted. were pro-war and others anti-war. Men the paid labor market was that PTAs and Their “communities” were the roughly 100 who betrayed their pro-war neighbors by neighborhood associations lost their “vol- men in their units — men they lived with deserting moved away, driven out by shame unteer army.” We started to write a paper 24 hours a day. and ostracism.4 Community codes of con- testing whether the rise in women’s labor We answer the question of when men duct are re-enforced not just by loyalty but force participation explained the decline are willing to sacrifice for the common also by punishments. in residential community participation.1 good by examining why men fought in By examining men’s lives during the To our surprise, we found little evidence the Civil War.2 During this war most sol- war we saw that more diverse communities supporting this claim. Instead, our analy- diers stood by their comrades even though are less cohesive. Their members are less sis of long-run trends in volunteering, join- a rational soldier would have deserted. willing to sacrifice and derive fewer ben- ing groups, and trust suggested that, all else Punishments were too rare and insuffi- efits from being part of the community. equal, people who live in cities with more ciently severe to deter men from deserting. Are there then any benefits to being in a * What then motivated these men to stand diverse community? When we look at the Dora L. Costa is a Research Associate their ground? Was it their commitment lives of black soldiers after the Civil War we in the NBER’s Programs on Development of the American Economy, Aging, and to the cause, having the “right stuff,” high can understand the tensions between the Children. She is also the Director of the morale, officers, or comrades? We examine short-run costs of diversity and its long- NBER Working Group on Cohort Studies all of these explanations and find that loy- run benefits. Men did not like to serve with and a professor of economics at UCLA. alty to comrades trumped cause, morale, those who were different from them, so

NBER Reporter • 2008 Number 2  much so that they were more likely to des- same ethnicity, occupation, and age group. nounced among Union Army veterans ert, but in the long-run the ex-slaves who For black soldiers, similar men were those than it is among more recent populations, joined the Union Army learned the most from the same state or even plantation and probably because of the elimination of the from being in units with men who were from the same slave or free background. summer-time diarrheal deaths with the different from themselves.5 But, in the long run (and studies of col- improvement in sanitation and because of Whether diversity fosters understand- lege roommates have never been able to the improvement in vitamin levels among ing or distrust is a long standing question examine the long run), Union Army sol- women pregnant during the long winter in the social sciences that has become par- diers benefited from their interactions with months.8 Lorens Helmchen, Sven Wilson, ticularly timely with rising immigration men who were different. Freemen taught and I found that being born in the second and growing income inequality. We find the former slaves to write and helped them relative to the fourth quarter predicted the that the same types of social network vari- forge a freeman’s identity. Both slaves and probability of Union Army veterans devel- ables that determined who deserted from freemen first learned of new cities and oping arteriosclerosis at older ages, perhaps the Union Army and who survived POW states from their comrades who had come because nutritional deprivation in utero camps predict commitment to organiza- from those places. leads to compromised immune function tions in civilian life today. Organizational There is increasing interest in build- and higher inflammation rates.9 membership is lower in metropolitan areas ing “good” communities today. The World If older age health and longevity is with greater racial and ethnic diversity Bank, on its social capital Web site, writes partially determined by earlier life condi- and higher income inequality; support for “Increasing evidence shows that social tions, then senescent processes may be plas- income redistribution is higher when the cohesion — social capital — is critical for tic and highly controllable and life expec- aid recipients are from the same racial and poverty alleviation and sustainable human tancy rates are likely to continue to rise ethnic group; and laboratory games show and economic development” (http://web. until improvements in early life conditions that trust is higher when the players look worldbank.org). This social capital has have been exhausted. Older age mortality like each other.6 both positive and negative consequences. rates may therefore not decrease as rap- Our work emphasized the importance Union Army deserters were never re-inte- idly for the post-baby-boom cohort as they of ethnicity, state of birth, occupation, age, grated into their communities, not because have for earlier cohorts. However, because and kinship for the formation of social ties of legal punishments, but because of shame of rising incomes, the value of even mar- in the past. We are not claiming these were and ostracism. ginal improvements in life expectancy is the only factors that influenced the for- We have highlighted the tensions now higher than the value of the very large mation of social ties among Union Army between cohesion and diversity. A commu- increases in life expectancy experienced at soldiers. Nor are we claiming that these nity of similar people is likely to be cohe- the beginning of the twentieth century.10 factors are as important now as they were sive and its members are likely to sacrifice Early life conditions may also partially in the past. Race and ethnicity no longer time, effort, and even their lives for each explain observed racial health disparities predetermine friendships and marriages. other. But in a diverse community mem- in modern populations and among Union Although racial and ethnic diversity still bers can learn from one another. Army veterans. Data on Union Army vet- affect community participation, they have erans reveal very high rates of arterioscle- become less important relative to income.7 Health and Social Networks rosis among black relative to white veter- Although people want to be friends ans, differences that can be explained by with others they can relate to, they may Our research agenda on social net- blacks’ greater life long burden of infec- learn the most from those who are dif- works does not end with our book. Our tious disease.11 The effects of infectious ferent. In recent Supreme Court cases a new NIA-funded research combines disease rates are clearly observed in black- brief filed by eight universities empha- our interests in social networks with my white differentials in birth weights, prema- sized that students educate each other, research agenda on the determinants of turity rates, and stillbirths among children that cross-racial learning takes place, and health at older ages. Using data on Union born in the first third of the twentieth cen- that this learning is valued by students Army veterans, I find that a large pro- tury under the auspices of Johns Hopkins and by the labor market. Nevertheless, portion of the poor health and longev- University Hospital.12 few large-scale studies actually measure the ity experience of men in the past relative Economic and epidemiological benefits of diversity in either a university or to men in more recent cohorts can be research has linked social networks to an employment setting and campus news- explained by such early life experiences as health. People who report themselves to paper accounts suggest large amounts of infectious disease, poor nutritional intake, be socially isolated, both in the number racial self-segregation. and physical job demands. For example, and quality of their personal relationships Like college students, Civil War sol- Joanna Lahey and I found that season face a higher mortality risk from all causes diers preferred to interact with others who of birth had a strong effect on the older and from several infectious, neoplastic, and looked like them. For white Union Army age mortality rates of men who lived to cardiovascular diseases. A large body of lit- soldiers, similar men were those of the age 60. This effect was much more pro- erature links stress, whether in the form

12 NBER Reporter • 2008 Number 2 of war, natural disasters, divorce, lack of in Social Capital, 1952–1988”, NBER 8 D.L. Costa and J.N. Lahey, “Becoming control on the job, or even disrupted sleep Working Paper No. 8295, May 2001, and Oldest-Old: Evidence from Historical patterns, to cardiovascular disease. Social Kyklos, 56(1) (Fasc 1 2003), pp. 17–46. U.S. Data,” NBER Working Paper No. networks could either mitigate or accen- 2 D.L. Costa and M.E. Kahn, “Cowards 9933, September 2003, and Genus, tuate the effects of stress. They could miti- and Heroes: Group Loyalty in the 61(1) (2005), pp. 21–38, and D.L. Costa gate the effects of stress through benefi- American Civil War,” NBER Working and J.N. Lahey, “Predicting Older Age cial effects on psychological and physical Paper No. 8627, December 2001, and Mortality Trends,” Journal of the European well-being. But, they could accentuate the Quarterly Journal of Economics, 118(2) Economic Association, 3(2–3) (March- effects of stress if the stressor leads to the (May 2003), pp. 519–48. April 2005). loss of friends or family (for example the 3 D.L. Costa and M.E. Kahn, “Surviving 9 D.L. Costa, L. Helmchen, and well-established effect of death of a spouse Andersonville: The Benefits of Social S. Wilson, “Race, Infection, and on the mortality of a survivor). Networks in POW Camps”, NBER Arteriosclerosis in the Past”, Proceedings of By studying how the interactions Working Paper No. 11825, December the National Academy of Sciences, 104 between unit cohesiveness and combat 2005, and American Economic Review, (2007), pp. 13219–24. Written for the mortality, and between POW camp expe- 97(4) (2007), pp. 1467–87. NBER conference, Economics of Health and rience and the strength of social networks 4 D.L. Costa and M.E. Kahn, “Shame Mortality, http://www.nber.org/books.html within the POW camp, affected older age and Ostracism”, NBER Working Paper No. 10 D.L. Costa and M.E. Kahn, “Changes mortality and morbidity, we can inves- 10425, April 2004, and “Deserters, Social in the Value of Life, 1940–1980,” NBER tigate the interaction between stress and Norms, and Migration,” Journal of Law Working Paper No. 9396, December 2002, social networks in one of the few human and Economics, 50 (2007), pp. 323–53. and Journal of Risk and Uncertainty, populations to provide us with measures of 5 D.L. Costa and M.E. Kahn, “Forging 29(2) (September 2004), pp. 159–80. stress, of long-run outcomes, and of exog- a New Identity: The Costs and Benefits 11 D.L. Costa, L. Helmchen, and enous social networks. We are finding that of Diversity in Civil War Combat Units S. Wilson, “Race, Infection, and being in a more cohesive company reduced for Black Slaves and Freemen”, NBER Arteriosclerosis in the Past”. the negative, long-term consequences of Working Paper No. 11013, December 12 D.L. Costa, “Race and Pregnancy wartime stress on older age mortality and 2004, and Journal of Economic History, Outcomes in the Twentieth Century: A morbidity, particularly from cardiovascu- 66(4): pp. 936–62. Long-Term Comparison,” NBER Working lar causes.13 Focusing on the role of stress 6 For a summary of social capital Paper No. 9593, March 2003, and Journal factors in health and mortality may be research see D. L. Costa and M.E. Kahn, of Economic History, 64(4) (December a fruitful line of research, particularly as “Civic Engagement and Community 2004), pp. 1056–86. we exhaust the gains from public health Heterogeneity,” Perspectives on Politics, 13 D.L. Costa and M.E. Kahn, “Health, advances. 1(1) (March 2003). Stress, and Social Networks: Evidence from 7 D.L. Costa and M.E. Kahn, Union Army Veterans,” NBER Working 1 D.L. Costa and M.E. Kahn, “Understanding the American Decline in Paper No. 14053, June 2008. “Understanding the American Decline Social Capital, 1952–1988”.

NBER Profile: David G. Blanchflower

David G. Blanchflower is a Research an honorary Doctor of Science degree at Associate in the NBER’s Program on Labor the University of London in 2009. Studies and the Bruce V. Rauner Professor His main areas of research interest of Economics at Dartmouth College. He is are happiness economics; wage determina- also a Visiting Professor at the University tion; unemployment; discrimination; trade of Stirling and a member of the Monetary unions, and entrepreneurship. His book Policy Committee at the Bank of England. The Wage Curve with Andrew Oswald won Blanchflower obtained his BA at the Princeton University’s Richard A. Lester University of Leicester, a Masters from the prize. University of Wales, and PhD from the Blanchflower has two daughters and a University of London. He received an hon- son. orary Doctor of Letters in 2008 from the University of Leicester and will be awarded

NBER Reporter • 2008 Number 2 3 Conferences

Twenty-third Annual Conference on Macroeconomics

The NBER’s twenty-third Annual Area” Markus K. Brunnermeier, Princeton Conference on Macroeconomics, orga- Discussants: Lucrezia Reichlin, European University and NBER; Stefan Nagel, nized by NBER Research Associates Central Bank, and Harald Uhlig, Stanford University and NBER; and Daron Acemoglu of MIT, Kenneth University of Chicago Lasse H. Pedersen, New York University Rogoff of Harvard University, and Michael and NBER, “Carry Trades and Currency Woodford of Columbia University, took Quamrul H. Ashraf and Ashley Lester, Crises” place in Cambridge on April 4 and 5. The Brown University, and David N. Weil, Discussants: A. Craig Burnside, Duke program was: Brown University and NBER, “When University and NBER, and Hanno Lustig, Does Improving Health Raise GDP?” University of California, Los Angeles and Giuseppe Moscarini, Yale University Discussants: Hoyt Bleakley, University of NBER and NBER, and Fabien Postel-Vinay, Chicago and NBER, and Simon Johnson, University of Bristol, “The Timing of IMF and NBER Andrew Atkeson, University of Labor Market Expansions: New Facts and California, Los Angeles and NBER, and a New Hypothesis” Jeremy Greenwood, University of Patrick J. Kehoe, Federal Reserve Bank Discussants: Robert E. Hall, Stanford Pennsylvania and NBER, and Nezih of Minneapolis and NBER, “On the University and NBER, and Robert Guner, Universidad Carlos III de Need for a New Approach to Analyzing Shimer, University of Chicago and NBER Madrid, “Marriage and Divorce since Monetary Policy” (No abstract is available World War II: Analyzing the Role of for this paper.) Jean Boivin, HEC Montréal and NBER; Technological Progress on the Formation Discussants: John H. Cochrane, Marc P. Giannoni, Columbia University of Households” University of Chicago and NBER, and and NBER; and Benoit Mojon, Discussants: Stefania Albanesi, Columbia Bennett T. McCallum, Carnegie Mellon Federal Reserve Bank of Chicago, University and NBER, and Justin Wolfers, University and NBER “Macroeconomic Dynamics in the Euro University of Pennsylvania and NBER

Moscarini and Postel-Vinay docu- wages respond little on impact, and start of the euro. In particular, they find that ment three new facts about aggregate rising only when firms run out of cheap German interest-rate shocks triggered dynamics in U.S. labor markets over the unemployed hires and start competing stronger responses of interest rates and last 15 years, drawing in part from newly to poach and retain employed workers. consumption in some countries, such as available datasets. The new facts con- Aggregate shocks thus are propagated Italy and Spain, than in Germany itself. cern a strong co-movement between the by the hiring behavior of large firms. A According to their estimates, the creation employer-to-employer worker transition calibrated example shows that the model of the euro has contributed to a greater rate, various measures of wages, and the qualitatively captures the essence of the homogeneity of the transmission mech- share of employment at large firms. All three facts. anisms across countries and to an over- three remain below trend several years Boivin and his co-authors character- all reduction in the effects of this shock. into the expansion. Then, simultane- ize the transmission mechanism of mon- Using a structural open-economy model, ously, large firms take over employment, etary and oil-price shocks across coun- they argue that the combination of a workers start quitting more from job to tries of the euro area, document how this change in the policy reaction function, job, and wages accelerate. The research- mechanism has changed with the intro- mainly toward a more aggressive response ers investigate whether this new view of duction of the euro, and explore some to inflation and output, and the elimina- how business cycles evolve and mature is potential explanations. The framework tion of an exchange rate risk can explain consistent with the transitional dynam- they use allows them to jointly model the the evolution of the monetary transmis- ics of the Burdett and Mortensen (1998) euro area dynamics and permit the trans- sion mechanism observed empirically. equilibrium search model, analyzed in a mission of shocks to be different across Ashraf and his co-authors quantita- companion paper (Moscarini and Postel- countries. They find important hetero- tively assess the effect of exogenous health Vinay, 2008). In their model, following a geneity across countries in the effect of improvements on output, through demo- positive aggregate shock to labor demand, macroeconomic shocks before the launch graphic channels and changes in worker

14 NBER Reporter • 2008 Number 2 productivity. They consider both changes ical progress in the household sector can occur in periods in which investor risk in general health, proxied by changes in explain these facts. This makes it more appetite and funding liquidity decrease. life expectancy, and changes in the preva- feasible for singles to maintain their own Carry-trade losses reduce future crash lence of two particular diseases: malaria home, and for married women to work. risk, but increase the price of crash risk. and tuberculosis. In general, they find that To address this question, the authors The researchers also document excess co- the effects of health improvements on develop a search model of marriage and movement among currencies with similar income are substantially lower than those divorce, which incorporates household interest rates. Their findings are consistent that are often quoted by policymakers, production. An extension of the model with a model in which carry traders are and may not emerge at all for a third of a looks back at the prewar era. subject to funding liquidity constraints. century or more after the initial improve- Brunnermeier and his co-authors These papers will appear in an annual ment in health. document that carry traders are subject volume published by the University of Since World War II there has been: to crash risk — that is, that exchange rate Chicago Press. Its availability will be a rise in the fraction of time that married movements between high interest rate announced in a future issue of the Reporter. households allocate to market work; an and low interest rate currencies are nega- They will also be found at “Books in increase in the rate of divorce; and a decline tively skewed. They argue that this nega- Progress” on the NBER’s website. in the rate of marriage. Greenwood and tive skewness is attributable to sudden Guner argue that labor-saving technolog- unwinding of carry trades, which tend to

Demography and the Economy

NBER Research Associate John B. of California, Berkeley and NBER “Demographic Trends, Housing Equity, Shoven of Stanford University organized and the Financial Security of Future a conference on “Demography and the Gopi Shah Goda and John B. Shoven, Retirees” Economy” that took place in California “Adjusting Government Policies for Age Discussant: Thomas Davidoff, on April 11 and 12. These papers were Inflation” University of California, Berkeley discussed: Discussant: Warren Sanderson, SUNY Stony Brook Sylvester J. Schieber, Social Security Sam H. Preston and Caroline R. Sten, Advisory Board, “Richer or Poorer University of Pennsylvania, “The Future Axel Börsch-Supan, University of When Death Do Us Part?” of American Fertility” Mannheim and NBER, and Alexander Discussant: Steven F. Venti Discussant: Gopi Shah Goda, Harvard Ludwig, University of Mannheim, “Old University Europe Ages. Can it Still Prosper?” Orazio Attanasio, University College Discussant: Alan J. Auerbach, London and NBER; Sagiri Kitao, Larry E. Jones, University of Minnesota University of California, Berkeley and University of Southern California; and NBER; Alice Schoonbroodt, NBER and Giovanni L.Violante, New York Southampton University; and Michele University and NBER, “Financing Tertilt, Stanford University and NBER, Shripad Tuljapurkar, Stanford Medicare: A General Equilibrium “Fertility and Income in the Cross University, “The Final Inequality: Analysis” Section: Evidence and Theory” Changes in the Variance in Age at Discussant: Moshe Bushinsky, Discussant: Amalia Miller, University of Death over Time” University of California, Los Angeles Virginia Discussant: Victor R. Fuchs, Stanford and NBER University and NBER Adam Isen and Betsey Stevenson, Francesco C. Billari and Guido University of Pennsylvania, “Women’s James M. Poterba, MIT and NBER; Tabellini, Universita’ Bocconi, “Italians Education and Family Behavior: Trends Steven F. Venti, Dartmouth College are Late: Does it Matter?” in Marriage, Divorce, and Fertility” and NBER; and David A. Wise, Discussant: Luigi Pistaferri, Stanford Discussant: Enrico Moretti, University Harvard University and NBER, University

The level of fertility is the principal age structure, which in turn is a criti- pay-as-you-go system of public pensions. determinant of the shape of a population’s cal factor in the terms-of-trade within a Social Security Administration simula-

NBER Reporter • 2008 Number 2 5 tions show that the 75-year actuarial bal- from marriage, have not affected all fami- changes is dramatic in some countries ance of the social security system would lies similarly. Isen and Stevenson exam- and will deeply affect future labor, finan- be $2.6 trillion higher in present value if ine how marital and fertility patterns cial, and goods markets. The expected fertility were high (2.3 children/woman) have changed along racial and educational strain on public budgets and especially rather than low (1.7). Partly because of lines for men and women. Marriage and social security already has received prom- their long-run effects on population age remarriage rates have risen for women inent attention, but aging poses many structure, national fertility levels are con- with a college degree relative to those other economic challenges that threaten sidered “too low” by a majority of govern- with fewer years of education, eroding a productivity and growth if they are not ments in developed countries. Preston long-standing gap caused by greater mar- addressed. There is no shortage of policy and Sten review the major factors that riage propensities among less educated proposals to address population aging. appear to be affecting fertility levels in the women. In contrast, there has been lit- However, little is known about behavioral United States, with an eye towards mak- tle change in marital patterns by educa- reactions, for example, to pension and ing defensible statements about future tion among men. Divorce has been falling labor market reform. Börsch-Supan and directions of fertility. There is no single for all groups, but fell earlier and more Ludwig shed light on such reactions in widely accepted framework for analyzing sharply among college graduates. Further, three large Continental European coun- the determinants of fertility at the level of the authors find that total fertility relative tries. France, Germany, and Italy have a population. In its place, the authors pur- to women’s educational attainment has large pay-as-you-go pension systems and sue an eclectic, inductive approach, sur- changed little, even though women’s edu- vulnerable labor markets. At the same veying the landscape of fertility variation cational attainment has been increasing time, they show remarkable resistance in search of clues about its principal driv- and women have fewer children at higher against pension and labor market reform. ers. Their search considers variation over levels of educational attainment. There The key issues discussed in this paper time and space and across individuals. has been a widely noted rise in the age of are: interactions between pension and Jones and his co-authors revisit the mothers, and the authors show that most labor market policies, and the behavioral relationship between income and fertil- of it comes from an increase in the age of reactions to reform. Which behavioral ity. There is overwhelming empirical evi- childbirth for women with more educa- reactions will strengthen, and which will dence that fertility is negatively related to tion, with little change in fertility timing weaken reform policies? Can Old Europe income in most countries at most times. for those with less education. prosper even if behavioral reactions coun- Several theories have been proposed to Government policies based on age ter current reform efforts? explain this somewhat puzzling fact. The do not adjust to the fact that, because of Tuljapurkar attempts to show that most common is based on the opportu- improvements in mortality, a given age the variance in age at death for adults is a nity cost of time being higher for individ- is associated with higher remaining life useful and important dimension of mor- uals with higher earnings. Alternatively, expectancy and lower mortality risk rel- tality change. Trends in this variance are people might differ in their desire to pro- ative to earlier time periods. Goda and informative about the speed and the age- create and, accordingly, some people will Shoven examine four possible meth- pattern of mortality change. The decom- invest more in children and less in mar- ods for adjusting the eligibility ages for position of this variance with respect to ket-specific human capital and thus have Social Security, Medicare, and Individual risk factors provides useful insights into lower earnings. The authors revisit these Retirement Accounts to determine what the explanatory power of different fac- and other possible explanations. They find eligibility ages would be today, and in tors that are correlated with mortality. that these theories are not as robust as is 2050, if adjustments for mortality Historical and economic analyses can ben- commonly believed, and that they often improvement were taken into account. efit from an examination of variance in depend on assumptions that are open to They find that historical adjustment of eli- age at death in addition to the tradition- question. gibility ages for age inflation would have ally important studies of life expectancy. The production efficiencies of house- increased ages of eligibility by approx- Poterba, Venti, and Wise have two hold specialization have declined with imately 0.15 years annually. Failure to key goals in this paper: 1) to understand the development of technologies that adjust for mortality improvement implies the extent of uncertainty about home simplify household production and the that the percent of the population eligible equity at older ages; and 2) to explore increase in market substitutes for goods to receive full Social Security benefits and how one might project the trend in the previously provided by homemakers. Medicare will increase substantially rela- home equity of younger cohorts as they Additionally, the opportunity cost of hav- tive to the share eligible under a policy of approach retirement. They begin by ing a household specialist has risen as the age adjustment. describing the change over time in the barriers to women in the workplace have Population aging will be a major relationships between age, home owner- eroded. These developments, which have determinant of long-run economic devel- ship, and home values. They find that the made way for an increase in the relative opment in industrial and developing age profile of home ownership rates has importance of the consumption benefits countries. The extent of the demographic changed little over the past two de–cades.

16 NBER Reporter • 2008 Number 2 On the other hand, there have been very you-go pension programs toward greater welfare implications of alternative fund- large increases in the value of owner-occu- funding. From a microeconomic perspec- ing schemes for Medicare. In the base- pied homes, and in home equity, over the tive, many workers may not discern any line closed-economy model, they find, past two decades. Using cohorts attaining practical effect from the restructuring of the labor income tax will have to increase retirement age in 1990 and in 2010, the the approach to financing their from 23 percent in 2005 to 36 percent authors simulate the evolution of home pensions. Even from a macroeconomic in 2080 to finance the rising costs of values over the course of a typical retire- perspective, there are questions over Medicare. However, under an open-econ- ment to explore the relationship between whether some of the move toward pen- omy scenario, the tax would have to rise home equity at retirement and home sion funding that has arisen in recent by much less. Limiting the increase in equity at older ages, when it tends to be years is more cosmetic than real. Whether the wage tax through either a rise in the drawn down. Because real home prices pension plans are funded or not, popula- Medicare premium or a delay in the age of rose during the sample period they use to tion aging will likely result in slower labor retirement is welfare improving. forecast future price patterns, for both forces growth in most cases. This slower Italians start adult activities, such as cohorts, their projections suggest that labor force growth has two important leaving the parental home, at a much home equity at older ages is likely to be implications. First, labor force growth later age than is common in other coun- much greater than equity at retirement. rates are one of the primary drivers that tries at comparable levels of development. Even when they truncate their sample of underlie economic growth. Second, a Billari and Tabellini ask whether this house price changes before the most growing aged population in the face of a late transition into adulthood influences recent market declines, their projections stable or diminished work force implies the lifetime economic opportunities of suggest a probability of between 10 and significant increases in aged dependency. individuals. A priori, it is not clear if such 14 percent that real home equity will The combination of these forces will limit influence exists, and what its direction decline between the ages of 59 and 79. future growth in living standards in the is. After providing stylized cross-coun- That probability rises substantially when developed economies of the world. If our try evidence on the potential negative they expand their sample of housing economies cannot meet public expecta- impact of the late transition to adult- returns to include the experience of the tions about economic performance, pen- hood, the authors consider a longitudinal most recent two years. Cross-section and sion policy will play a major role in allo- survey of Italian men in their mid-30s. As cohort data also show that over a 20-year cating the economic disappointment. a measure of transition into adulthood, period marked by very large fluctuations Schieber suggests that some countries they focus on leaving the parental home in the growth rate of home value, very face such significant demographic shifts for the first time; they find that individ- large increases in household wealth, and a toward older populations that pension uals who leave the parental home earlier large decline in mortgage rates, the ratio funding will offer little practical relief. In in life earn a higher income in their mid- of home equity to total non-pension these cases, the whole concept of retire- 30s. Estimation by instrumental variables wealth remained remarkably stable. This ment that has persisted over much of the suggests that the authors capture a causal empirical regularity leads them to con- past century will have to be revisited. effect, from the age of leaving the paren- sider whether projections of the home Attanasio and his co-authors develop tal home to subsequent economic out- equity of future retirees might be based a general equilibrium, overlapping-gener- comes, and the result is robust to various on forecasts of the wealth of future house- ations model of the U.S. economy where specification of the model. The authors holds. The recent turmoil in the housing households face uncertain health sta- speculate on the potential mechanisms market adds interest to such projections tus that in turn determines their mortal- through which this effect might operate. but also, by drawing attention to the large ity rate and their medical expenditures. Given these results, policies that aim at changes in home value and home equity Households make consumption and speeding up the transition to adulthood that can occur over a short period of time, labor supply decisions, and can imper- of young Italians might have positive eco- raises speculation about whether the past fectly insure health shocks through mar- nomic effects. empirical regularity will continue in the kets. The government provides partial These papers will appear in an NBER future. insurance through Medicare and “social Conference Volume published by the Some analysts conclude that aging assistance,” and runs a Pay-As-You-Go University of Chicago Press. Its availabil- societies must radically modify their social security system. The authors cali- ity will be announced in a future issue of retirement systems in order to deal with brate the model based on the projected the NBER Reporter. They will also be new economic realities. There are several demographic and medical expenditures available at “Books in Progress” on the examples where the response to popula- trends for the next 75 years. The model NBER’s website. tion aging has been to restructure pay-as- is used to study the macroeconomic and

NBER Reporter • 2008 Number 2 7 Innovation Policy and the Economy

The NBER’s ninth annual Richard Freeman, Harvard University Douglas Lichtman, University of Conference on Innovation Policy and and NBER, and John Van Reenen, California, Los Angeles, “Pricing the Economy took place in Washington London School of Economics and Patents” on April 15. The conference was orga- NBER, “What if Congress Doubled nized by NBER Research Associates R and D Spending on the Physical Alvin Roth, Harvard University and Adam B. Jaffe of Brandeis University, Sciences?” NBER, “What Have We Learned from Joshua Lerner of Harvard University, Market Design?” and Scott Stern of Northwestern Joseph Farrell, University of California, University. The following papers were Berkeley, “Intellectual Property as a Peter Cramton, University of Maryland, discussed: Bargaining Environment” “Innovation and Market Design”

Many business, academic, and scien- extensively. But there has been much In theory, markets would be a preferred tific groups have recommended that the less study of private agreements that go mechanism for setting the terms of pat- Congress substantially increase R and D deeper than that and agree in advance ent deals. In a host of common infringe- spending in the near future. President on how future patents, or patents not yet ment situations, however, market-based Bush’s American Competitiveness identified, will be treated. One can think negotiation is either not an option or not Initiative calls for a doubling of spend- of such agreements as setting up a zone a trustworthy measure of patent value. ing over the next decade in selected of “private law”. This can be either good Roth discusses some things we have agencies that deal with the physical sci- or bad, and a traditional antitrust-based learned about markets, in the process ences, including the National Science approach tries to ban the bad ones and of designing marketplaces to fix market Foundation. Freeman and Van Reenen allow the good ones. But because private failures. To work well, marketplaces have consider the rationale for increased R negotiation is often imperfect and diffi- to provide thickness, that is they need and D spending in the context of the glo- cult, it makes sense to consider whether to attract a large enough proportion of balization of economic activity. To assess and how public policy can facilitate ben- the potential participants in the market; the likely consequences of a large increase eficial private agreements and learn from they have to overcome the congestion in R and D spending, they examine how them, not merely allow them. Farrell that thickness can bring, by making it the 1998–2003 doubling of the NIH explores some of these issues. possible to consider enough alternative budget affected the bio-medical sciences. Lichtman notes that over the last transactions to arrive at good ones; and They find that the rapid increase and many years, patent reform efforts have they need to make it safe and sufficiently ensuing deceleration in NIH spending focused primarily and understandably simple to participate in the market, as created substantial adjustment problems on the problem of incorrectly issued opposed to transacting outside of the in the market for research, particularly patents. These efforts have pointed out market, or having to engage in costly and for younger investigators, and failed to fundamental flaws in the system for risky strategic behavior. Roth draws on address long-standing problems with sci- evaluating patent applications, rightly recent examples of market design rang- entific careers that are likely to deter exhibiting concern that many patents are ing from labor markets for doctors and many young persons from choosing a issued today despite the fact that the pur- new economists, to kidney exchange, scientific career. They argue that fund- ported inventions are not meaningfully and school choice in New York City and ing agencies should tilt their awards to different from what was known before. Boston. younger researchers on the grounds that There is a second substantial problem Market design plays an essential any research project does two things: undermining the patent system, how- role in promoting innovation. Cramton it produces knowledge and adds to the ever, and that is the mistaken approach examines emission allowance auc- human capital of researchers, which has to setting damages and influencing the tions, air slot auctions, spectrum auc- greater value for younger persons because terms of licensing arrangements. In this tions, and electricity markets, and dem- of their longer future career life-spans. regard, the patent system is built on onstrates how the market design can Patent (and other intellectual prop- an intuition that prices and other deal encourage innovation. Improved pricing erty) policy anticipates and relies on terms ought to be set in the market, and information is one source of innovation. firms negotiating “in the shadow of the thus that patent remedies ought to be Enhancing competition is another driver law” when it comes to licensing and designed with an eye toward driving par- of innovation seen in all of the applica- using existing patents, and the econom- ties to engage in market-based negotia- tions. Market design fosters innovation ics of that process have been studied tion. But that is quite simply a mistake. in other ways as well by addressing other

18 NBER Reporter • 2008 Number 2 potential market failures. of Chicago Press. Its availability will “Books in Progress” on the NBER’s These papers will appear in an annual be announced in a future issue of the website. volume published by the University Reporter. They can also be found at

Risks of Financial Institutions

The NBER held a conference on of Chicago, “The Consequences of Contagion from Counterparty Risk” “Risks of Financial Institutions” at the Mortgage Credit Expansion: Evidence Discussant: Richard Cantor, Moody’s University of Chicago’s Gleacher Center from the 2007 Mortgage Default Crisis” Investors Service on April 24. Conference organizers (NBER Working Paper No. 13936) Mark Carey, Federal Reserve Board of Discussant: Stuart Gabriel, University of Amir Khandani, MIT, and Andrew Governors, and Rene M. Stulz, NBER California, Los Angeles Lo, MIT and NBER, “What Happened and Ohio State University, chose these to the Quants in August 2007?” papers for discussion: Robert H. Litzenberger, Azimuth Discussant: Kent Daniel, Northwestern Asset Management, and David M. University Markus K. Brunnermeier, Princeton Modest, J.P. Morgan Chase, “Crisis University and NBER, “Deciphering and Non-Crisis Risk in Financial Monica Billio and Loriana the 2007/8 Liquidity and Credit Markets: A Unified Approach to Risk Pelizzon, University of Venice, and Crunch” Management” Mila Getmansky, University of Discussant: Raghuram Rajan, University Discussant: Philippe Jorion, University Massachusetts, “Crises and Hedge Fund of Chicago and NBER of California, Irvine Risk” Discussant: Tobias Adrian, Federal Atif Mian, University of Chicago and Philippe Jorion, and Gaiyan Zhang, Reserve Bank of New York NBER, and Amir Sufi, University University of Missouri, “Credit

Brunnermeier studies the underly- margins force financial institutions to cut Mian and Sufi offer evidence that a ing forces that drove adverse events in back on leverage, exacerbating the ini- rapid expansion in the supply of mort- the financial markets in 2007 and 2008. tial price decline. A key problem is the gages driven by disintermediation explains He explains how new structured financial maturity mismatch caused by leveraged a large fraction of recent U.S. house price products, off-balance sheet vehicles, and financing. The second set of amplifica- appreciation and subsequent mort- the transformation from a classical bank- tion mechanisms works through the lend- gage defaults. They identify the effect of ing model to an “originate and distribute ing channel. Uncertainty about future shifts in the supply of mortgage credit by model” led to a deterioration of lending funding needs, combined with poten- exploiting within-county variation across standards and to a boom in house prices. tially limited access to the lending mar- zip codes that differed in latent demand He provides an event-log book about ket at that time, can lead to hoarding and for mortgages in the mid-1990s. From the current market turmoil and identifies interest rate surges in the interbank mar- 2001 to 2005, high latent-demand zip various amplification mechanisms that ket. Brunnermeier also discusses runs on codes experienced large relative decreases explain how shocks to the financial system financial institutions. Runs occur when in denial rates, increases in mortgages could cause such large dislocations. The each individual financier has an incentive originated, and increases in house price first mechanism is attributable to borrow- to curtail funding before others. Finally, appreciation, despite the fact that these ers’ balance sheet effects involving liquid- when financial institutions are lenders zip codes experienced significantly neg- ity spirals. When asset prices and mar- and borrowers at the same time, network ative relative income and employment ket liquidity drop during times of crisis, and gridlock risk might emerge. In a grid- growth over this time period. These pat- funding requirements for financial insti- lock situation, each individual institu- terns for high latent-demand zip codes tutions increase. This happens because tion is not able to pay its obligation only were driven by a sharp relative increase in the collateral value of the assets on bor- because the others are not paying theirs. the fraction of loans sold by originators rowers’ balance sheets erodes, and mar- While coordination might resolve this, it shortly after origination, a process that the gins rise, or investors are unable to roll- is difficult to achieve in today’s complex authors refer to as “disintermediation.” over their short-term liabilities. Higher and interweaved financial system. The increase in disintermediation-driven

NBER Reporter • 2008 Number 2 9 mortgage supply to high latent-demand severe negative abnormal equity returns side the long/short equity sector — in a zip codes from 2001 to 2005 led to subse- and increases in CDS spreads. In addi- completely unrelated set of markets and quent large increases in mortgage defaults tion, creditors are more likely to suffer instruments — suggesting that systemic from 2005 to 2007. The results suggest from financial distress later. These effects risk in the hedge-fund industry may have that moral hazard on behalf of originators are stronger for industrial creditors than increased in recent years. selling mortgages is a main culprit for the financials. Simulations calibrated to these Billio, Pelizzon, and Getmansky U.S. mortgage default crisis. results indicate that counterparty risk can study the effect of financial crises on Litzenberger and Modest develop potentially explain the observed excess hedge fund risk. Using a regime-switching an analytically tractable risk management clustering of defaults. This suggests that beta model, they separate systematic and metric that measures potential exposures counterparty risk is an important addi- idiosyncratic components of hedge fund to financial crises and also captures vola- tional channel of credit contagion and exposure. The systematic exposure to vari- tility during non-crisis times. Their mul- that current portfolio credit risk models ous risk factors is conditional on market tiple-regime stress-loss risk framework understate the likelihood of large losses. volatility conditions. They find that in the assumes markets are characterized by qui- During the week of August 6, 2007 a high-volatility regime (when the market is escent (non-crisis) periods most of the number of quantitative long/short equity rolling-down and is likely to be in a crisis time, interspersed with infrequent cri- hedge funds experienced unprecedented state) most strategies are negatively and sis periods where 4-5 sigma events can losses. Based on TASS hedge-fund data, significantly exposed to the Large-Small occur with non-negligible probabilities. TAQ data, and simulations of a specific and Credit Spread risk factors. This sug- The framework is flexible and can incor- long/short equity strategy, Khandani and gests that liquidity risk and credit risk are porate arbitrary numbers and types of cri- Lo hypothesize that the losses were initi- potentially common factors for different ses. One of the primary lessons of 1998 ated by the rapid “unwind” of one or more hedge fund strategies in the down-state and 2007 is that returns can be correlated sizable quantitative equity market-neu- of the market, when volatility is high and because collections of trades (or strate- tral portfolios. Given the speed and price returns are very low. They further explore gies) are sometimes financed from a com- impact with which this occurred, it was the possibility that all hedge fund strate- mon pool of capital, even though returns likely the result of a forced liquidation gies exhibit a high volatility regime of the to the strategies have low correlations dur- by a multi-strategy fund or proprietary- idiosyncratic risk, which could be attrib- ing quiescent periods. trading desk, possibly because of a mar- uted to contagion among hedge fund Standard credit risk models cannot gin call or a risk reduction. These initial strategies. In their sample, this event hap- explain the observed clustering of default, losses then put pressure on a broader set pened only during the Long-Term Capital sometimes described as “credit conta- of long/short and long-only equity port- Management crisis of 1998. Other crises, gion.” Jorion and Zhang provide the first folios, causing further losses by trigger- including the recent subprime mortgage empirical analysis of credit contagion via ing stop/loss and de-leveraging policies. crisis, affected hedge funds only through direct counterparty effects. Using a unique A significant rebound of these strategies systematic risk factors and did not cause database, they examine the wealth effects occurred on August 10, which is also con- contagion among hedge funds. of bankruptcy announcements on cred- sistent with the unwind hypothesis. The itors. On average, creditors experience dislocation was triggered by events out-

20 NBER Reporter • 2008 Number 2 Climate Change: Past and Present

An NBER/Universities Research Hybrid Corn: American Corn Yields, Karen Clay, Carnegie Mellon and Conference on “Climate Change: 1866–2002” NBER, and Werner Troesken, George Past and Present,” organized by Gary Mason University and NBER, “On the D. Libecap, University of California, Raghav Gaiha, Vani S. Kulkarni, and Seasonality of Disease: Implications Santa Barbara and NBER, and Richard Kenneth Hill, Harvard University; and for the Effects of Climate Change on H. Steckel, Ohio State University and Shantanu Mathur, International Fund Health” NBER, took place in Cambridge on May for Agricultural Development, “On 30 and 31. These papers were discussed: Devastating Droughts” Haggay Etkes, Hebrew University, “The Impact of Short-Term Climate Richard H. Steckel, “Tree Rings Alan L. Olmstead, University of Fluctuations on Rural Population in Climate Change and Impacts: Sketching California, Davis and NBER, and the Desert Frontier Nahiye of Gaza (ca. a Research Agenda” Paul W. Rhode, University of Arizona 1519–57)” and NBER, “Adjusting to Climatic Melissa Dell, MIT; Benjamin F. Jones, Variation: Historical Perspectives Price Fishback, University of Arizona Northwestern University and NBER; from North American Agricultural and NBER; Trevor Kollman, and Benjamin A. Olken, Harvard Development” University of Arizona; Michael Haines, University and NBER, “Climate Colgate College and NBER; Paul W. Change and Economic Growth: Wolfram Schlenker, Columbia Rhode; Melissa Thomasson, Miami Evidence from the Last Half Century” University and NBER, and Michael University and NBER; and Shawn Roberts, USDA, “Estimating the Kantor, University of California, Zeynep K. Hansen, Boise State Impact of Climate Change on Crop Merced and NBER, “The Health University and NBER; Gary D. Yields: The Importance of Nonlinear Consequences of Natural and Economic Libecap; and Scott E. Lowe, Boise Temperature Effects”(NBER Working Disasters: 1930–1940” State University, “Climate Variability Paper No. 13799) and Water Infrastructure: Historical Valerie A. Mueller and Daniel Experience in the Western United Hoyt Bleakley, University of Chicago E. Osgood, Columbia University, States” and NBER, and Sok Chul Hong, “Long‑term Consequences of University of Chicago, “The Impact of Short‑term Precipitation Shocks: Richard Sutch, University of Weather on U.S. Farm Productivity: Evidence from Brazilian Migrant California, Riverside and NBER, Historical Patterns and Relation to the Households” “Henry Agard Wallace, the Iowa Corn Changing Disease Environment” Yield Tests, and the Adoption of

Economists and other social scien- North America based on tree rings, dis- temperatures appear to reduce growth tists who seek to measure the relationship cusses various databases that are available, rates, rather than just the level of output, in between climate and socioeconomic activ- and considers some applications of these poor countries. Third, higher temperatures ity are limited to somewhat more than one data for studying settlement of the conti- have wide effects in poor nations, reduc- century of data from instrument records. nent; demographic patterns of health and ing agricultural output, industrial output, Fortunately for this research agenda, den- migration; agricultural prices; financial and aggregate investment, and increasing dochronologists have been rapidly expand- stress; and government responses to cli- political instability. Should future impacts ing our knowledge of past precipitation mate change. of climate change mirror these historical and temperature. Tree rings reliably recon- Dell and her co-authors use annual effects, the negative effect on poor coun- struct the Palmer Drought Severity Index variation in temperature and precipita- tries might be substantial. (PDSI), which measures effective or net tion over the past 50 years to examine the Hansen and his co-authors have con- precipitation. Dendochronologists have impact of climatic changes on economic structed (and are still expanding) an inte- estimated the PDSI in North America activity throughout the world. They find grated dataset on water supply and water for a grid of 286 locations based on 835 three primary results. First, higher tem- infrastructure in the states west of the one tree-ring chronologies that go back hun- peratures substantially reduce economic hundredth meridian. This county-level dreds of years. Steckel briefly explains the growth in poor countries but have little dataset includes details on all the con- methodology of climate reconstruction in effect in rich countries. Second, higher structed dams and canals as well as aqui-

NBER Reporter • 2008 Number 2 21 fers and streams. The dam data starts in in flood -prone counties would lessen after moderate learning has the potential to 1850 and goes through 2001. This exten- the construction of dams. Since they have avert a large fraction of deaths. But capac- sive dataset will account for the entire all the major dams and canals constructed ity building— synonymous with availabil- water supply and water distribution infra- in last 150 years in their dataset, they ity of more resources for disaster preven- structure in the western United States. will examine the impact of the growth in tion — also has considerable potential in The dataset is spatially linked to topo- water infrastructure construction on agri- averting deaths. In fact, these findings graphic characteristics, historical climate cultural composition and production, and are broadly consistent with the view that data, historical agricultural data, and his- on flood control over time. fatalities are greater in countries with weak torical population data at a county level Sutch [2007] makes the following governments and pervasive poverty. using GIS. The authors use this dataset claims: 1) There was not an unambigu- Providing greater historical perspec- to seek answers to several questions. First, ous yield advantage of hybrid corn over tive would enlighten current discussions they are interested in understanding the the open-pollinated varieties before 1936. about future human responses to climatic historical pattern of these construction 2) The early adoption of hybrid corn can variation. During the nineteenth and projects. Specifically, they examine the fac- be explained better by a sustained propa- twentieth centuries, new biological tech- tors and influences that explain the timing ganda campaign conducted by the U.S. nologies allowed North American farm- and the location decisions of these dams Department of Agriculture at the direc- ers to push cropping into environments and canals. They are most interested in tion of the Secretary of Agriculture, Henry previously thought too arid, too variable, investigating and establishing the poten- Agard Wallace. The Department’s cam- and too harsh to cultivate. Olmstead and tial link between the climate variability paign echoed that of the commercial seed Rhode document these changes for three (changes in annual precipitation totals companies. 3) The early adopters of hybrid major staple crops, noting that the cli- and mean monthly temperatures) and the seed were followed by later adopters as matic challenges that previous generations construction of water infrastructure in a consequence of the droughts of 1934 of farmers overcame often rivaled the cli- the western states. In addition, they ana- and 1936. The eventual improvement of matic changes predicted for the next hun- lyze the impact of the water infrastruc- yields as newer varieties were introduced dred years in North America. Further ture construction on agricultural produc- explains the continuation and acceleration analysis is needed to understand the tim- tion, and on flood control. They also are of the process. ing, causes, and costs of these adjustments, interested in analyzing other determinants Droughts often turn into famines. Loss they conclude. of investment in water storage and dis- of agricultural output and food shortage The United States produces 1 4 per- tribution channels, including population are, however, not the only consequences. cent of the world’s corn and 38 percent growth and the increased electrification of There are often large second round effects, of the world’s soybeans, so any impact urban residences. They have detailed infor- some of which persist over time. By the on U.S. crop yields will have implications mation on the purpose of dam construc- time these effects play out, the overall eco- for world food supply. Schlenker and tion, ranked in terms of priority. Thus, nomic loss is substantially greater than Roberts pair a panel of county-level crop they can separately focus on dams con- the first round loss of income. Hardships yields in the United States with a fine- structed mostly for flood control, for irri- manifest in malnutrition, poverty, disin- scale weather dataset that incorporates gated agriculture, and water/power sup- vestment in human capital (for example, the whole distribution of temperatures ply for the areas with larger populations. withdrawal of children from school), liq- between the minimum and maximum They hypothesize that counties, which are uidation of assets (for example, sale of within each day and across all days in the included in the water supply and distri- livestock) with impairment of future eco- growing season. Yields increase in temper- bution infrastructure, were better able to nomic prospects, and, in extreme cases, ature until about 29C for corn, 30C for deal with the problems of short-term cli- mortality, given the incompleteness of soybeans, and 32C for cotton, but temper- matic variability (either because of natural credit and insurance markets. Gaiha and atures above these thresholds become very variability in the hydrologic cycle or dis- his co-authors build on the extant litera- harmful. The slope of the decline above ruptions of the cycle) in terms of smooth- ture. While the frequency of droughts the optimum is significantly steeper than ing out agricultural production over time has risen, their deadliness has declined. the incline below it. The same nonlin- relative to similar counties without such This paper sheds light on the underlying ear and asymmetric relationship is found infrastructure. Having identified major geographical, institutional, development whether the authors consider time-series drought periods over time, they can look indicators in explaining inter-country dif- or cross-sectional variation in weather and at the variation in agricultural produc- ferences in mortality. The analysis also yields. This suggests limited potential for tion during these periods before and after confirms the favorable effects of openness adaptation within crop species because dam/canal construction as well as during in saving human lives, and that much of the latter includes farmers’ adaptations normal climatic times in a difference-in- this devastation appears to be avoidable to warmer climates and the former does difference setting. Similarly, they hypoth- through a timely and speedy entitlement not. Area-weighted average yields given esize that the problems related to flooding protection strategy. Furthermore, even current growing regions are predicted to

22 NBER Reporter • 2008 Number 2 decrease by 31–43 percent under the slow- would have died anyway even without the severe droughts. Meanwhile, excessive est warming scenario and 67–79 percent extremes in temperature. Part II also asks flooding and severe rainfall struck other under the most rapid warming scenario by a related question: whether poor socio- sections. Recent corrections to the cli- the end of the century. economic groups were more vulnerable mate data show that the 1930s included Bleakley and Hong consider the to changes in temperature than wealth- several of the hottest years on record in impact of weather on farm values. Using ier groups. Part III explores how public American climate history. Fishback and both cross-sectional and panel methods, health interventions altered seasonal dis- his co-authors examine the impact of nat- they document economically and statisti- ease patterns and helped protect popula- ural disasters and economic disasters on cally significant differences in the response tions against the diseases related to tem- infant mortality rates and adult mortality of farm productivity to weather over time. perature change. rates during the course of the 1930s. They In recent decades, farm value has been Etkes explores the impact of short- have constructed an annual panel dataset (weakly) increasing in temperature and term climate fluctuations on rural pop- for over 3000 counties for the years 1930 rainfall. On the other hand, high levels of ulation by examining the differences in through 1940. The dataset includes annual temperature or rain depressed farm value the latter’s growth rates within the Gaza data on infant mortality rates in both in the nineteenth century. This suggests an nahiye (ca. 1519-57). This small adminis- rural and city areas within the counties. important role for technological adapta- trative unit (a trapezoid of about 60 x 35 When they examine the pure relationships tion in reducing the impact on farm pro- Km) borders the desert and included between climate and natural disasters and ductivity of hotter and wetter weather both southern semi-arid and northern infant mortality, they find that high tem- (a possible outcome of climate change in Mediterranean climate zones. The demo- peratures and severe wetness and flooding some places). One particular adaptation— graphic data — the number of rural tax- were associated with higher infant mortal- the eradication of malaria—accounts for payers — were deciphered from Ottoman ity. After controlling for income and many a substantial fraction of this difference tax surveys, and the indications for his- government programs during the 1930s, over time. These results also suggest that torical climate fluctuations include tree they find that climate has very little impact malaria reduced farm productivity by a rings, grain prices, and qualitative evi- on infant mortality rates. These prelimi- factor of 2–3 in the most malarious parts dence derived from historical documents. nary results suggest that climate variation of the South relative to the Plains in the The study shows that the population in and natural disasters influence infant mor- late 1800s. southern semi-arid villages was more sensi- tality by affecting income and other eco- Clay and Troesken hope to establish tive to climate fluctuations than the popu- nomic features, while having only a small the robustness of the seasonality of disease lation in northern Mediterranean villages. direct effect after controlling for those and to better understand the mechanisms In other words, dry years were followed factors. that underlie the connection between cli- by a relatively slow demographic growth Mueller and Osgood find that large matic variation and disease. Their paper in southern villages, while bountiful years short-term precipitation shocks damage has three parts. Part I explores the sea- were followed by accelerated southern the long-term income of households that sonal patterns in deaths from six causes: demographic growth. The study inter- have permanently migrated from rural to pneumonia, typhoid fever, scarlet fever, prets this pattern as south-north migra- urban areas. This outcome is consistent diphtheria, infant mortality, and small- tion aimed at smoothing consumption with the behavior of credit-constrained pox. They find that these infectious dis- across wet and dry years. Similar circular rural households who are willing to accept eases are highly sensitive to seasonal vari- migration are familiar as survival strategy lower long-term income in urban areas fol- ation in temperature or rainfall but that in other regions such as the Sahel in Sub- lowing the depletion of their productive the effects of temperature are not uniform Saharan Africa. assets during an adverse shock. The accep- or linear. Part I also asks why seasonal During the 1930s, the economy expe- tance of lower income persists into the long variation in temperature affects disease rienced the worst Depression in U.S. his- term, as long as mechanisms for rebuilding rates. The central finding here is that expo- tory. In addition, the American popula- capital remain unavailable. Their empirical sure to organic and inorganic pathogens tion faced the trials of Job as it was struck evidence suggests that there may be a link depends heavily on temperature. Part II by a series of natural disasters. Everybody between large precipitation shocks in rural looks for evidence of what demographers remembers the Dust Bowl as it kicked areas and urban poverty. Further explora- and epidemiologists refer to as harvest- up huge amounts of particulates into the tion is warranted on the mechanisms by ing, such that unusually high or low tem- air, but this was only the most famous which natural disasters cause these long- peratures are killing only the weakest and of many other disasters. Other parts of term losses. most unhealthy individuals, people who the country were struck by extreme and

NBER Reporter • 2008 Number 2 23 NBER News

2007–8 Awards and Honors

A number of NBER researchers received honors, prizes, awards, and professional kudos during 2007 and early 2008. In alpha- betical order, those so honored (excluding honors by the individual’s own university) are:

Alberto Abadie was awarded the become Editor-in-Chief of the Journal of “The Financial Accelerator: Evidence 2007 Gosnell Prize by the Society of Economic Perspectives. From International Housing Markets,” Political Methodology for “the best work Laurence Baker will receive the co-authored with Heitor Almeida and in methods presented at any political American Society of Health Economists Crocker Liu. science conference during the preceding (ASHE) Medal at their meeting in June David Card won the Econometric year” for his paper, “Synthetic Control 2008. The ASHE Medal is awarded bien- Society’s Ragnar Frisch Prize. Methods for Comparative Case Studies: nially “to the economist age 40 or under Amitabh Chandra and Douglas Estimating the Effect of California’s who has made the most significant contri- Staiger won the Arrow award for the best Tobacco Control Program,” with Alexis butions to the field of health economics.” published paper in Health Economics. Diamond and Jens Hainmueller. Katherine Baicker was named to the The paper was “Productivity Spillovers in John M. Abowd was elected a Fellow Editorial Board of Health Affairs; she also Healthcare: Evidence from Heart-Attack of the Society of Labor Economists in was named a Commissioner of the Robert Treatments” appearing in the February 2007. Wood Johnson Commission to Build a 2007 Journal of Political Economy. Katharine Abraham became a Fellow Healthier America, and was appointed to Lauren Cohen won the 2007 of the Society of Labor economists. the Institute of Medicine’s Committee on Smith Breeden Prize (distinguished Daron Acemoglu received an hon- the Effects of Lacking Health Insurance paper) in the Journal of Finance for her orary doctorate from the University of Coverage. paper, “Supply and Demand Shifts in Utrecht, Netherlands. Lucian Bebchuk was elected the Shorting Market” (with Karl Diether, Yacine Ait-Sahalia was awarded a President of the American Law and Ohio State and Christopher Malloy, of Guggenheim Fellowship for 2008–9 for Economics Association. HBS); she also won the 2007 Barclays his work on jump and volatility. He also Roland Benabou delivered the Global Investors Prize from the European was elected a Fellow of the American Joseph Schumpeter lecture at the meetings Finance Association for Best Paper on Statistical Association. of the European Economic Association in Asset Pricing for, “The Small World Heitor Almeida, Murillo Campello, Budapest, August 2007. of Investing: Board Connections and and Crocker Liu won the Goldman Sachs Jonathan Berk’s paper, “A Critique Mutual Fund Returns “ (with Andrea Prize for the best paper published in of Size Related Anomalies,” was selected Frazzini of Chicago GSB and Christopher the Review of Finance in 2007 for “The as one of 100 seminal papers published Malloy, of HBS) and the 2007 Society Financial Accelerator: Evidence from in Oxford University Press’s 100-year of Quantitative Analysts Award from International Housing Markets.” history. the Western Finance Association for Lee J. Alston gave his Presidential David G. Blanchflower was awarded Best Paper in Quantitative Investments Address to the International Society an honorory Doctor of Letters from his for, “Attracting Flows by Attracting Big for the New Institutional Economics in alma mater, the University of Leicester, Clients: Conflicts of Interest and Mutual Reykjavik, Iceland in June 2007. in the UK. Fund Portfolio Choice” (with Breno Orley Ashenfelter is a Distinguished Nick Bloom won an Alfred P. Sloan Schmidt from USC). Fellow of the American Economic Research Fellowship. Maureen Cropper was elected to Association. Howard Bodenhorn received a John The National Academy of Sciences in Susan Athey received the American Simon Guggenheim Fellowship for a April 2008. The National Academy of Economic Association’s John Bates Clark projecttitled “The Political Economy of Sciences is “a private organization of sci- Medal in 2007. Reform in Jacksonian New York.” entists and engineers dedicated to the David Autor was awarded the 2008 John Y. Campbell delivered the lun- furtherance of science and its use for the Sherwin Rosen Prize by the Society of cheon address to the annual meeting of the general welfare. It was established in 1863 Labor Economists “In recognition of Western Finance Association in 2007. by a congressional act of incorporation outstanding contributions to the field of Murillo Campello won the 2007 signed by Abraham Lincoln that calls on labor economics.” He also was selected by Goldman Sachs Best Paper Award from the Academy to act as an official adviser to the American Economic Association to the Review of Finance for the article the federal government, upon request, in

24 NBER Reporter • 2008 Number 2 any matter of science or technology.” of the state of New York, empowering the Shapiro received the 2007 Robert H. David M. Cutler was elected into the Comptroller to offer an independent and Durr Award from the Midwest Political American Academy of Arts and Sciences. binding forecast in case the Executive and Science Association. Angus Deaton was elected President- Legislative branches fail to reach a con- Mark Gertler was awarded a Elect of the American Economic sensus forecast by March 1 of each year. Guggenheim Fellowship for 2007. Association. His term as President will The Center’s forecast has had an impor- Robert Gibbons became a Fellow of begin in 2009. He received a Laurea tant impact this year. the Society of Labor Economists in spring Honoris Causa from the University of Robert C. Feenstra presented the 2008. Rome, Tor Vergata, in June 2007, and an Zeuthen Lectures at the University of Roger Gordon became a Fellow Honorary D.Sc (Econ.) from University Copenhagen in April 2007. Those lec- of the American Academy of Arts and College, London, in September 2007. tures will be published as an MIT Press Sciences. Stefano DellaVigna was awarded a book. Pierre-Olivier Gourinchas was run- Sloan Research Fellowship in February Amy Finkelstein received a 2007–9 ner up for the 2007 prize for the best 2008. Alfred P. Sloan Research Fellowship. French young economist (under 40) Darrell Duffie was elected a Fellow Kristin Forbes was selected as a awarded by Cercle des Economistes and of the American Academy of Arts and member of the Trilateral Commission in Le Monde. He won the 2008 Bernacer Sciences in 2007. 2007. Prize for the best European econo- Esther Duflo is the Inaugural Barbara Fraumeni is the incoming mist under 40 working on finance and holder of the Chair “Knowledge against Chair of the Committee on the Status macroeconomics. Poverty” in the College de France, Paris, of Women in the Economics Profession Jeffrey Grogger won the 2007 for 2008/9. (CSWEP), a committee of the American Outstanding Statistical Application Esther Duflo, Jeffrey Liebman, Economic Association. Award from the American Statistical and Emmanuel Saez (along with Peter Victor Fuchs gave the 2007 John Association for his paper with Greg Orszag and William Gale) received the Eisenberg Legacy Lecture on December Ridgeway entitled “Testing for Racial 2007 TIAA-CREF Paul A. Samuelson 13, 2007. That Lecture honors Dr. John Profiling in Traffic Stops from Behind a Certificate of Excellence for their paper Eisenberg, a renowned internist and Veil of Darkness” (Journal of the American “Saving Incentives for Low- and Middle- health services researcher who directed Statistical Association, 101 (475), Septem­ Income Families: Evidence from a Field the Agency for Healthcare Research and ber 2006, pp. 878–87.) Experiment with H&R Block,” published Quality (AHRQ) from 1997 to 2002. Reuben Gronau received the Jacob in the Q uarterly Journal of Economics in Also, the American Society of Health Mincer Award for lifetime contribution November 2006. Economists (part of the International to Labor Economics from the Society of William Easterly delivered the D. Health Economics Association) has Labor Economics. Gale Johnson Lecture at the University of named a Lifetime Achievement Award Gene M. Grossman delivered the Chicago and the Zale Lecture at Stanford after him: the Victor R. Fuchs ASHE Condliffe Memorial Lecture at the University. Career Award for Lifetime Contributions University of Canterbury in November Sebastian Edwards delivered the to the Field of Health Economics. It is to 2006 and the Walras-Pareto Lectures at Figuerola Lecture at the Universidad be awarded biennially to an economist the University of Lausanne in May 2007. Carlos III in Madrid in September 2007; who has made significant lifetime contri- Michael Grossman received the he also delivered the Corden Lecture at butions to the field of health economics. 2008 Victor R. Fuchs Award. This Award the University of Melbourne in October Martin Gaynor, Jian Li, and William for Lifetime Contributions to the Field 2007. B. Vogt received the first annual Victor of Health Economics is presented bienni- Ronald G. Ehrenberg was awarded R. Fuchs Research Award from RAND ally “to an economist who has made sig- an Honorary Doctor of Science degree and the Forum for Health Economics nificant lifetime contributions to the field by the State University of New York in and Policy for their paper, “Substitution, of health economics.” He also became May 2008. Spending Offsets, and Prescription Drug President-elect of the American Society Isaac Ehrlich has been selected by Benefit Design.” This $10,000 prize, spon- of Health Economists in the fall of 2006 the Comptroller of the State of New sored by RAND, goes to the authors of and will be President of that organiza- York to be in charge of developing an the best paper with the potential to spawn tion as of June 26, 2008. Grossman served independent annual forecast of the state’s new research in an underdeveloped area as President of the Eastern Economic tax revenue (a job similar to that of the of health economics or health policy. The Association from February 2007 through CBO), heading a team of professors from winning paper was published in Forum for March 2008 and delivered the presidential the Center of Human Capital at SUNY Health Economics & Policy: Vol. 10: Iss. 2, address at the annual conference of that Buffalo. The assignment follows - abud Article 4. association in Boston in March 2008. get reform act enacted by the legislature Matthew Gentzkow and Jesse Michael R. Haines continues as

NBER Reporter • 2008 Number 2 25 Treasurer of the Social Science History Kevin Lang won Honorable Mention Olivia S. Mitchell and Annamaria Association (since 2005). Also, Historical (runner-up) for the best new Professional/ Lusardi won the 2007 Fidelity Pyramid Statistics Of The United States, Millennial Academic Book in Sociology and Social Prize for their paper “Baby Boomer edition (NY: Cambridge University Press, Work from the Association of American retirement security: The roles of plan- 2006), in 5 volumes and an electronic edi- Publishers. ning, financial literacy, and housing tion, won the Thomas Jefferson Prize from David S. Lee and Adriana Kugler wealth” (which appeared in the Journal the Society for History in the Federal both won the Labor and Employment of Monetary Economics, Volume 54, Issue Government. The award was presented in Relations Association’s 2007 John T. 1, Pages 205–24 in January 2007.) The Washington, D.C. in March 2008. Among Dunlop Scholar Award for “outstand- prize was given for their work on advanc- the editors-in-chief of this publications ing academic contributions to research by ing an understanding of the importance are several NBER Research Associates: recent entrants to the field.” of financial literacy and planning in help- Richard Sutch, Gavin Wright, Alan Ronald Demos Lee became a ing Americans to reach their financial Olmstead, Michael Haines. member of the American Philosophical goals. Mitchell also received the 2008 Oliver Hart was the Sukhamoy Society. Roger F. Murray Prize from the Institute Chakravarty Memorial Lecturer, Delhi Eric M. Leeper has been offered a for Quantitative Research in Finance (1st School of Economics, and the (Inaugural) Professorial Fellowship in Monetary and Prize); and the 2008 Carolyn Shaw Bell Coase Lecturer at the London School of Financial Economics by the Reserve Bank Award of the Committee on the Status of Economics in 2007. of New Zealand and Victoria University Women in the Economics Profession. Igal Hendel and Aviv Nevo were and will spend fall 2008 in Wellington, Randall Morck received an “All awarded the Compass Prize 2007 (prize NZ. Star Paper Award” from The Journal for the paper making the most significant Jeffrey Liebman, Esther Duflo, of Financial Economics for cites to his contribution to the understanding and Emmanuel Saez (along with Peter 2000 article with Bernard Yeung and implementation of competition policy) Orszag and William Gale) received the Wayne Yu: “The Information Content for “Measuring the Implications of Sales 2007 TIAA-CREF Paul A. Samuelson of Stock Markets: Why Do Emerging and Consumer Inventory Behavior.” Certificate of Excellence for their paper Markets Have Synchronous Stock Price Robert Inman was appointed a “Saving Incentives for Low- and Middle- Movements?”It appeared in the Journal of Fellow of the Rockefeller Foundation, Income Families: Evidence from a Field Financial Economics, 58(1), pp. 215-60. Bellagio, Italy for the Fall, 2007. Experiment with H&R Block,” published Dale Mortensen and Finis Welch Panle Jia received the Zellner Thesis in the Q uarterly Journal of Economics in shared the 2007 Society of Labor Award for Best Dissertation in Business November 2006. Economists’ Jacob Mincer prize honor- and Economics Statistics from the Andrew W. Lo won the 2007 Smith- ing lifetime achievements in the field of American Statistical Association in 2007. Breeden Distinguished Paper award (with labor. Sebnem Kalemli-Ozcan received Jiang Wang); the Boston Security Analysts Stefan Nagel and Jonathan Lewellen two European awards: the 2008 Wim Society Honorary Member award; and the won the Fama/DFA prize for the best Duisenberg Research Fellow (of the Market Technicians Association Annual paper in asset pricing in the Journal of European Central Bank, or ECB) for lead- Award. Financial Economics for “The conditional ing economists in academia, central banks, Annamaria Lusardi and Olivia CAPM does not explain asset pricing and top research institutions who are rec- Mitchell won the Fidelity Pyramid Prize, anomalies.” ognized international experts in their field; a $50,000 award given to authors of pub- Derek Neal is President-Elect of the and the Marie Curie Reintegration Award lished applied research that best helps Midwest Economics Association and (of the European Commission, or EC) address the goal of improving lifelong a new Fellow of the Society of Labor designed to attract top-class researchers financial well-being for Americans. Economists. from third countries to work and under- Lisa M. Lynch received the 2007 Joseph Newhouse was named to the take research in Europe. Susan C. Eaton Scholar-Practioner Congressional Budget Office’s Board of Edward J. Kane was named the award from the Labor and Employment Health Advisers and to the Comptroller 2008 Midwest Financial Association Relations Assocation. General’s Advisory Board. Distinguished Scholar. N. Gregory Mankiw was inducted as Ariel Pakes was selected as the Timothy J. Kehoe was made Doctor a Fellow in the American Academy of Arts Distinguished Fellow of the Industrial Honoris Causa by the Universidad de and Sciences in 2007. OrganizationSociety for 2007. Vigo in February 2008. Robert C. Merton was awarded Stavros Panageas won the 2007 Per Krusell received the Sööderberg the honorary degree, Doctor of Science Geewax-Terker Prize in Investment Prize in economic sciences. It is given Honoris Causa, by Claremont Graduate Research for the best paper published every four years to an economist with a University, Claremont California at a cer- in the Rodney White Working Paper connection to Sweden. emony on May 17, 2008. Series at Wharton that year. The paper

26 NBER Reporter • 2008 Number 2 is titled “Optimal Retirement Benefit Fellowship in 2007. In April 2008 he the Intergovernmental Panel on Climate Guarantees.” was elected a member of the National Change, which was the co-recipient (with Mark Pauly received the National Academy of Sciences. Albert Gore Jr.) of the 2007 Nobel Peace Institute of Health Care Management Anna J. Schwartz was inducted Prize. Stavins also became the founding Foundation’s Research Award for “Is into the American Academy of Arts and Editor of the Review of Environmental Health Insurance Affordable for the Sciences on October 6, 2007. Economics and Policy. Uninsured?” with M. Kate Bundorf Fiona M. Scott Morton and co- Jeremy C. Stein was named a Fellow (Journal of Health Economics, July 2006), authors Florian Zettelmeyer and Jorge of the American Academy of Arts and in May 2007; the John M. Eisenberg Silva-Risso received the 2007 Green Sciences in 2008, and was also elected Excellence in Mentorship Award, Agency Award for their article “How the Internet president of the American Finance for Health Care Research and Quality Lowers Prices: Evidence from Matched Association. in June 2007; and the Distinguished Survey and Automobile Transaction Andrew Sweeting won the Austin Investigator Award, AcademyHealth, also Data,” published in the May 2006 issue of Robinson Memorial Prize for the best in June 2007. the Journal of Marketing Research. paper by an economist within five years of Lubos Pastor (and co-author Robert Jesse Shapiro and Matthew Ph.D. published in the Economic Journal Stambaugh) received the Goldman Sachs Gentzkow received the 2007 Robert H. in 2007. The paper was “Market Power Asset Management Award for the best Durr Award from the Midwest Political in the England and Wales Wholesale paper in empirical investments presented Science Association for their paper “What Electricity Market 1995–2000.” at the 2007 meetings of the Western Drives Media Slant?” The award is given Richard Sylla was elected vice chair- Finance Association. “for the best paper applying quantitative man of the board of trustees of the Museum John Pencavel won the Society of methods to a substantive problem.” of American Finance, a Smithsonian affil- Labor Economists’ Mincer Award for Robert J. Shiller was named iate located at 48 Wall Street, New York Lifetime Achievement. “Global Risk Manager of the Year” by the City. The election occurred in October Tomas Philipson received the GlobalAssociation of Risk Managers in 2007. Garfield Award from Research America 2007. He is the first academic to receive Michele Tertilt was awarded the NSF for best health economics paper in a given this honor. CAREER grant (the NSF’s major award year. Hans-Werner Sinn delivered in support of early-career scientists). She Joshua D. Rauh received the Brattle the Thuenen Lecture to the German is also the W. Glenn Campbell and Rita Prize, First Prize Paper, for the best cor- Association of Economists (Verein für Ricardo-Campbell National Fellow at the porate finance paper published in the Socialpolitik) in Munich in October Hoover Institution during academic year Journal of Finance in 2006. The award was 2007. 2007/8. announced on January 6, 2007. James M. Snyder was inducted Laura Veldkamp, Patrick Bolton, Jonah Rockoff received a Smith into the American Academy of Arts and and Markus Brunnermeier’s paper Richardson Public Policy Fellowship, Sciences in 2007. “Leadership, Coordination and Mission- given annually to three young economists Gary R. Solon was elected a Fellow Driven Management” won the J.P. working in public economics and public of the Society of Labor Economists. Morgan Prize for the best paper at the policy in the United States, in 2007. Douglas O. Staiger and Amitabh Utah Winter Finance Conference. Dani Rodrik won the inaugural Chandra received the Kenneth J. Arrow Pietro Veronesi received the Fama/ Albert O. Hirschman award of the Social Award in Health Economics from DFA Prize for the best paper in the Journal Science Research Council in 2007, pre- the International Health Economics of Financial Economics in the Areas of sented for “outstanding contributions to Association for the best paper in health Capital Markets and Asset Pricing (second international, interdisciplinary social sci- economics in 2007: “Productivity prize) for “Was There a NASDAQ Bubble ence research, theory, and public com- Spillovers in Health Care: Evidence from in the Late 1990s?” (with Lubos Pastor), munication, in the tradition of Albert the Treatment of Heart Attacks,” pub- 2006, Journal of Financial Economics 81, Hirschman.” lished in the Journal of Political Economy, 61–100. (The prize is awarded in 2007 for Harvey S. Rosen won the National February 2007. a publication in 2006.) Tax Association’s Daniel M. Holland Robert Stambaugh (and co-author W. Kip Viscusi was appointed as Medal for distinguished lifetime contri- Lubos Pastor) received the Goldman an Honorary Member, Academy of butions to the study and practice of public Sachs Asset Management Award for Economics and Finance. finance in 2007. the best paper in empirical investments Joel Waldfogel and his co-author Steve Ross won the Jean-Jacques presented at the 2007 meetings of the Lu Chen won the Journal of Industrial Laffont Prize, 2007. Western Finance Association. Economics’s “Best Article of the Year” Jose A. Scheinkman was awarded Robert Stavins, Charles Kolstad, and Prize for 2006 for “Does Information a John Simon Guggenheim Memorial Joseph Stiglitz served as Lead Authors for Undermine Brand? Information Inter­

NBER Reporter • 2008 Number 2 27 mediary Use and Preference for Branded Association Meeting, for “Liquidity and Michael Woodford was awarded the Web Retailers”, which appeared in the Market Crashes”’ (with J. Huang), 2007. 2007 Deutsche Bank Prize in Financial December 2006 issue (Volume LIV No. Michael Whinston was elected a Economics. 4). Fellow of the American Academy of Arts Stephen R. Yeaple was awarded the Jiang Wang received the 2007 and Sciences. 2008 Bhagwati Award for his paper, “A Smith Breeden Prize for distinguished Michelle J. White is President-elect Simple Model of Firm Heterogeneity, paper in the Journal of Finance for of the American Law and Economics International Trade, and Wages.” This “Trading Volume: Implications of An Association. (She will succeed Lucian award is given every other year for the Intertemporal Asset Pricing Model”(with Bebchuk, who is now president.) best paper in the Journal of International A.W. Lo), 2007 and the New York Stock Barbara (Bobbi) Wolfe has been Economics. Exchange Award for the best paper awarded a Guggenheim Fellowship for on equity trading, Western Finance 2008.

Program and Working Group Meetings

Science of Science Day

The NBER’s Program on Productiv­ Pierre Azoulay; Joshua Graff Zivin, Jacques Mairesse, CREST‑ENSAE ity held a “Science of Science Day” when Columbia University and NBER; and NBER, “Individual Productivity it met on March 14. Program Director and Jialan Wang, MIT, “Superstar Differences in Scientific Research: Ernst R. Berndt and NBER Research Extinction” An Econometric Exploration of Associate Pierre Azoulay, both of MIT, Discussant: Lynne G. Zucker, University Publications by French Physicists” organized the meeting. These papers of California, Los Angeles and NBER Discussant: Iain Cockburn, Boston were discussed: University and NBER Liam Brunt, University of Lausanne; Shulamit Kahn, Boston University, Josh Lerner, Harvard University and David Popp, Syracuse University and Megan J. MacGarvie, Boston NBER; and Tom Nicholas, Harvard and NBER, and Richard Newell, University and NBER, “How Important University, “Inducement Prize and Duke University, “Where Does is Location for Research Productivity in Innovation” Energy R and D Come From? A Science?” Discussant: Michael Kremer, Harvard First Look at Crowding Out from Discussant: Richard B. Freeman, University and NBER Environmentally‑Friendly R and D” Harvard University and NBER Discussant: Rebecca Henderson, MIT Laure Turner, CREST‑ENSAE; and and NBER

Kahn and MacGarvie ask whether uses a new and unique dataset of foreign- dents who left the United States have scientists located outside the United born U.S.-educated scientists that allows more or fewer publications, patents, cita- States are at a disadvantage when it comes its authors to exploit exogenous variation tions, and collaborators when compared to research productivity, collaboration, in post-Ph.D. location induced by visa with a control student with the same and knowledge diffusion. The principal status. They thus are able to compare stu- advisor. They also ask whether these stu- difficulties of comparing scientists inside dents who were required by law to leave dents have more or fewer international the United States with those outside the the United States upon the completion collaborations. United States arise from unobserved het- of their studies with similar students who Azoulay and his co-authors estimate erogeneity among scientists and the endo- were allowed to remain in the United the magnitude of spillovers generated by geneity of location choices. This paper States. The researchers assess whether stu- 137 academic “superstars” in the life sci-

28 NBER Reporter • 2008 Number 2 ences onto their coauthors’ research pro- a prolific minority of scientists produces to the potential magnitude of such crowd- ductivity. These researchers died while most of the publications and accounts for ing out effects. Popp and Newell attempt still being actively engaged in science, thus most of citations in almost all research to address this question. They consider providing an exogenous source of varia- fields. Turner and Mairesse investigate the private opportunity costs of climate tion in the structure of their collabora- to what extent such dispersion and per- R and D, asking whether an increase in tors’ coauthorship networks. Following sistence can be accounted for by three climate R and D represents new R and the death of a superstar, coauthors suf- types of factors in so far as they can mea- D spending, or whether some (or all) of fer a lasting 8 to 18 percent decline in sure them: individual variables, mainly the additional climate R and D comes at their quality-adjusted publication output. age and gender, career stage variables, and the expense of other R and D. They begin These findings are surprisingly homoge- laboratory variables. Does individual sci- at the industry level, using sectoral data nous across a wide range of coauthor and entific productivity significantly drop off on R and D expenditures to ask whether coauthor/superstar dyad characteristics. as scientists become older and more or less increases in energy R and D spending Together, they suggest that part of the advanced in their careers? Is it strongly crowd out other R and D spending in var- scientific field embodied in the “invisible related or not to career promotions and ious industries. Preliminary results show college” of coauthors working in that area to the productivity and quality of the lab- some evidence of crowding out in sectors dies along with the star — a genuine and oratories in which scientists work? Even if active in energy R and D, but not in sec- irreplaceable loss of human capital. these factors prove quite significant, is it tors that do not perform energy R and Brunt and his co-authors examine nonetheless the case that individual pro- D. This suggests that funds for energy R prizes as an inducement for innovation, ductivity differences remain largely unac- and D do not come from other sectors, using a novel dataset of awards for inven- counted by them, and that they have to but may come from a redistribution of tiveness offered by the Royal Agricultural be mostly imputed to unobserved indi- research funds in sectors that are likely to Society of England from 1839 to 1939. vidual circumstances and characteris- perform energy R and D. Given this, the At annual shows, the RASE held com- tics, or so called “individual effects”? To authors proceed with a detailed look at petitive trials and awarded medals and answer such questions, the authors have climate R and D in two sectors: automo- monetary prizes (exceeding one million put together a 12- year panel database for tive manufacturing and alternative energy. pounds in current prices) to spur techno- 465 condensed matter physicists working Linking patent data and financial data logical development. This paper uncov- in the French public research organiza- by firm, they examine patenting behav- ers large effects of the prizes on contest tion CNRS, and have specified and esti- ior of large firms. For these firms, they entries, especially for the Society’s gold mated a simple econometric model for are able to separately identify patents per- medal. Matching award and patent data, both “productivity” and “quality,” mea- taining to innovative energy technology the authors also detect large effects of the sured respectively by the number of pub- (for example, wind energy and fuel cells) prizes on the quality of contemporaneous lications per scientist per year and by the from all other patents. They ask whether inventions. These results hold even dur- corresponding average citation impact of an increase in these energy patents leads ing the period when prize categories were these publications. to a decrease in other types of patenting determined by a strict rotation scheme, Recent efforts to endogenize tech- activity, which would suggest R and D thus overcoming the potential confound nological change in climate policy models crowding out. Preliminary results show that awards were offered in “hot” technol- demonstrate the importance of account- no evidence of crowding out within these ogy sectors. The evidence suggests that ing for the opportunity cost of climate firms. Finally, the authors use patent cita- prize awards can be a powerful mecha- R and D investments. Because the social tion data to ask where the most valuable nism for encouraging competition and returns to R and D investments are typi- patents in these sectors come from. They that prestigious non-pecuniary prizes can cally higher than the social returns to other find that firms specializing in alternative be a particularly effective inducement for types of investment, any new climate miti- energy research provide more valuable innovation. gation R and D that comes at the expense patents than larger firms for which alter- An empirical regularity has often of other R and D investment may dampen native energy R and D is just part of a been observed in economics of science: the overall gains from induced technolog- broader portfolio. productivity differences among research- ical change. Unfortunately, there has been ers are extremely large and persistent, and little empirical work to guide modelers as

NBER Reporter • 2008 Number 2 29 Asset Pricing Program Meeting

The NBER’s Asset Pricing Program Miao, Boston University, “Advance Pedro Santa‑Clara, University of met in Chicago on April 25. Hanno Information and Asset Prices” California, Los Angeles and NBER, and Lustig, University of California, Los Discussant: Snehal Banerjee, Shu Yan, University of South Carolina, Angeles and NBER, and Monika Northwestern University “Crashes, Volatility, and the Equity Piazzesi, University of Chicago, orga- Premium: Lessons from the S&P 500 nized the meeting. These papers were Zhiguo He and Arvind Options” discussed: Krishnamurthy, Northwestern Discussant: Ravi Bansal, Duke University, “Intermediary Asset Pricing” University and NBER Zhi Da and Pengjie Gao, University of Discussant: Stavros Panageas, University Notre Dame, and Ravi Jagannathan, of Pennsylvania John H. Cochrane University of Northwestern University and NBER, Chicago and NBER, “A Mean‑Variance “When Does A Mutual Fund’s Trade Christine A. Parlour and Johan Benchmark for Intertemporal Portfolio Reveal its Skill?” Walden, University of California, Theory” Discussant: Jonathan Berk, University Berkeley, “Capital, Contracts, and the Discussant: Stanley Zin, Carnegie of California, Berkeley and NBER Cross Section of Stock Returns” Mellon University and NBER Discussant: Adriano Rampini, Duke Rui Albuquerque and Jianjun University

Da, Gao, and Jagannathan conjec- tations equilibrium framework in which replicate observed crisis recovery times. ture that a mutual fund manager with investors are heterogeneous in their infor- They also calibrate their model to a broad superior stock selection ability is more mation and investment opportunities. intermediation scenario where parame- likely to benefit from trading in stocks The authors assume that informed inves- ters are chosen so that the marginal inves- affected by information-events. Taking tors privately receive advance information tor is an amalgam of the intermediaries the probability of informed trading (PIN, about company earnings that materializes we observe in practice. They show that Easley, Kiefer, O’Hara, and Paperman, into the future. This information is imme- the intermediation effects help to gener- 1996) to measure the amount of informed diately incorporated into prices, and thus ate a volatile pricing kernel and a market trading in a stock, and inferring mutual stock prices may move in ways unrelated risk premium matching the empirically fund trades from a large sample of mutual to current fundamentals. Investors’ specu- observed equity premium. fund holdings, the authors provide empir- lative and rebalancing trades in response Parlour and Walden present a trac- ical support for the conjecture. Funds to advance information generate short- table, static, general equilibrium model trading high-PIN stocks exhibit supe- run momentum, mimicking an underre- with multiple sectors in which firms offer rior performance on average, and superior action pattern. When this information workers incentive contracts and simul- performance that is more likely to persist. materializes, the stock price reverts back taneously raise capital in stock markets. The findings are not attributable to price to its long-run mean, mimicking an over- Workers optimally invest in the stock momentum or the higher returns earned reaction pattern. market and at the same time hedge labor by high-PIN stocks on average. The con- He and Krishnamurthy present an income risk. Firms rationally take agents’ clusions remain the same after testing for equilibrium asset pricing model in which portfolio decisions into account. In equi- alternative measures for the amount of intermediaries are marginal in setting librium, the cost of capital of each sec- informed trading. Decomposing a fund’s prices. The intermediaries modeled are tor is endogenous. The authors compare stock selection ability into “informed hedge funds, mutual funds, banks, or the first-best, in which workers’ effort trading” and “liquidity provision” adds insurance companies. The authors cali- levels are observable, to an economy in further insight into fund’s underlying brate the model to a hedge fund crisis which workers’ effort is observed with strengths. Impatient informed trading is a episode where parameters are chosen so noise. In the presence of moral hazard, significant source of alpha for funds trad- that the marginal investor resembles a the CAPM fails because firms, by choos- ing high-PIN stocks, while liquidity pro- hedge fund with leverage. The research- ing optimal incentive contracts, transfer vision is more important as a source of ers are able to qualitatively and quanti- risk both through wages and through the alpha for funds trading low-PIN stocks. tatively match the behavior of risk pre- stock market. This leads to several cross- Albuquerque and Miao provide an mia and interest rates in a financial crisis. sectional asset pricing “anomalies,” such as explanation for momentum and reversal Moreover, their model captures the slow size and value effects. As the researchers in stock returns within a rational expec- mobility of capital during a crisis and can characterize optimal contracts, they pres-

30 NBER Reporter • 2008 Number 2 ent empirical predictions relating workers’ late the implied measures of ex-ante risk intertemporal problems in incomplete compensation, firm productivity, firm size into an ex-ante risk premium. They find markets, with non-marketed income. and financial market abnormal returns. that the average premium that compen- Investors first hedge non-traded income They also demonstrate some general equi- sates the investor for the ex-ante risks and preference shocks. Then, their opti- librium implications of endogenous con- implicit in option prices, 11.8 percent, is mal payoffs are split between an indexed tracts; for example the ex-ante value of more than 70 percent higher than the pre- perpetuity and a “long-run mean-vari- human capital can be higher in an econ- mium required to compensate the same ance efficient” payoff, which avoids varia- omy with moral hazard. investor for the realized volatility, 6.8 per- tion over time as well as variation across Santa‑Clara and Yan use a novel cent. Moreover, the ex-ante equity pre- states of nature. In equilibrium, the mar- pricing model to imply times series of dif- mium that they uncover is highly volatile, ket payoff and the average outside-income fusive volatility and jump intensity from with values between 0.3 and 54.9 percent. hedge payoff span the long-run mean- S&P 500 index options. These two mea- The component of the premium that cor- variance frontier, and long-run expected sures capture the ex-ante risk assessed by responds to jump risk varies between zero returns are linear functions of long-run investors. The researchers find that both and 45.4 percent. The equity premium market and outside income-hedge betas. components of risk vary substantially over implied from option prices significantly State variables for investment opportuni- time, are quite persistent, and correlate predicts subsequent stock market returns. ties and outside income are conveniently with each other and with the stock index. Cochrane notes that by reinterpret- absent in these characterizations. Using a simple general equilibrium model ing the symbols, one-period mean-vari- with a representative investor, they trans- ance portfolio theory can apply to dynamic

Corporate Finance

NBER’s Program on Corporate Lucian A. Bebchuk, Harvard Benjamin Keys, University of Finance met in Chicago on April 25. University and NBER; Martijn Michigan; Tanmoy Mukherjee, Sorin Atif Mian, NBER and University of Cremers, Yale University; and Urs Capital Management; Amit Seru, Chicago, and Manju Puri, NBER and Peyer, INSEAD, “CEO Centrality” University of Chicago; and Vikrant Duke University, organized this agenda: Discussant: Rajesh Aggarwal, University Vig, London Business School, “Did of Minnesota Securitization Lead to Lax Screening? Raghuram G. Rajan, University Evidence From Subprime Loans” of Chicago and NBER, and Rodney Robin Greenwood and Samuel Discussant: Daniel Bergstresser, Ramcharan, International Monetary Hanson, Harvard University, and Harvard University Fund, “Landed Interests and Financial Jeremy C. Stein, Harvard University Underdevelopment in the United and NBER, “A Gap‑Filling Theory of Atif Mian, and Amir Sufi, University States” Corporate Debt Maturity Choice” of Chicago, “The Consequences of Discussant: Shawn Cole, Harvard Discussant: Adriano Rampini, Duke Mortgage Credit Expansion: Evidence University University from the 2007 Mortgage Default Crisis” (NBER Working Paper No. Vojislav Maksimovic and N.R. Christopher Mayer, Columbia 13936 — See “Risks of Financial Phabhala, University of Maryland, University and NBER; Tomasz Institutions” earlier in this issue for a and Gordon Phillips, University of Piskorski, Columbia University; and description of this paper.) Maryland and NBER, “ Post‑Merger, Alexei Tchistyi, New York University, Discussant: Nicholas Souleles, Restructuring and the Boundaries of the “The Inefficiency of Refinancing: University of Pennsylvania and NBER Firm” Why Prepayment Penalties Are Good Discussant: Murillo Campello, for Risky Borrowers” University of Illinois Discussant: Bruce Carlin, UC, Los Angeles

It is usually thought that the wealthy nomic institutions only in poor, undemo- find that even in the United States in the have an ability to limit wider access to eco- cratic countries. Rajan and Ramcharan early decades of the twentieth century,

NBER Reporter • 2008 Number 2 31 landed interests seem to play a signifi- lic firms. Their proxy for CEO centrality show that prepayment penalties are wel- cant role in the spread of financial insti- is the fraction of the top-five compensa- fare improving, and are more beneficial tutions. Counties with very concentrated tion captured by the CEO. They find that to borrowers with higher risk of default. land holdings tended to have dispropor- CEO centrality is negatively associated Mobility mitigates the benefits of prepay- tionately fewer banks per capita and fewer with firm value (as measured by industry- ment penalties only for the safest borrow- national banks. Moreover, aggregating adjusted Tobin’s Q). This result is robust ers. Empirical evidence from more than land distribution up to the state level, to controlling for all standard controls in 43,000 securitized fixed rate mortgages states that had higher land concentration Q regressions as well as additional controls (FRMs) supports the predictions of the passed more restrictive banking legisla- such as: CEO tenure, whether the CEO is model. First, borrowers who receive posi- tion. Finally, financial underdevelopment, a founder or a large owner, and whether tive credit shocks are much more likely as determined historically by land con- the company’s top-five aggregate compen- to prepay their mortgage, the riskier they centration, was negatively correlated with sation is high or low relative to peer com- are. Second, less creditworthy borrowers subsequent manufacturing growth, right panies, and in stronger companies with are most likely to have prepayment penal- up to the 1970s. Since these effects are high entrenchment levels. CEO central- ties, but also receive appreciable benefits. observed across counties possessing similar ity also is richly related to firms’ behavior Subprime borrowers with FICO scores political and legal institutions at the state and performance. In particular, CEO cen- below 620 obtain rates as much as 0.7 per- level, the evidence is suggestive that the trality is correlated with: lower (industry- cent lower than similar borrowers with origins of underdevelopment lie, in part, adjusted) accounting profitability; lower fully prepayable mortgages and default at in the historical pattern of constituencies stock returns accompanying acquisitions a lower rate (13 percent default rate) than or interests. announced by the firm and higher likeli- comparable borrowers with no prepay- Mergers and acquisitions are a fast way hood of a negative stock return accompa- ment penalties (18 percent default rate). for a firm to acquire assets. Using plant- nying such announcements; higher odds These findings suggest that regulations level data, Maksimovic and his co-authors of the CEO’s receiving a “lucky” option banning refinancing penalties might have examine how firms redraw their boundar- grant at the lowest price of the month; the unintended consequence of raising ies after acquisitions. They find that there greater tendency to reward the CEO for interest rates, increasing mortgage default, is a surprisingly substantial amount of luck in the form of positive industry-wide and limiting available credit for the riski- restructuring in a short period after merg- shocks; lower likelihood of CEO turnover est borrowers. ers are consummated. Acquirers sell 27 controlling for performance; and lower Theories of financial intermediation percent and close 19 percent of acquired firm-specific variability of stock returns suggest that securitization, the act of con- plants within three years after completing over time. verting illiquid loans into liquid securities, an acquisition. Plants that belong to the Greenwood and his co-authors argue could reduce the incentives of financial target’s peripheral divisions, especially in that time-series variation in the maturity intermediaries to screen borrowers. Keys industries in which asset values are increas- of aggregate corporate debt issues arises and his co-authors empirically examine ing and in industries in which the acquirer because firms behave as macro liquid- this question using a unique dataset on does not have a comparative advantage, are ity providers, absorbing the large sup- securitized subprime mortgage loan con- more likely to be sold by the purchasing ply shocks associated with changes in the tracts in the United States. They exploit firm. Acquirers who exhibit skill in run- maturity structure of government debt. a specific rule of thumb in the lending ning their peripheral businesses tend to The researchers document that when the market to generate an exogenous varia- retain acquired plants. Plants retained by government funds itself with relatively tion in ease of securitization and com- acquirers increase in productivity whereas more short-term debt, firms fill the result- pare the composition and performance sold plants do not. The extent of post- ing gap by issuing more long-term debt, of lenders’ portfolios around the ad-hoc merger restructuring activities and their and vice-versa. This type of liquidity pro- threshold. Conditional on being securi- cross-sectional variation do not support an vision is undertaken more aggressively:1) tized, the portfolio that is more likely to empire building explanation for mergers. in periods when the ratio of government be securitized defaults by around 20 per- Acquirers readjust their firm boundaries in debt to total debt is higher; and 2) by firms cent more than a similar risk-profile group ways that are consistent with the exploita- with stronger balance sheets. Their theory with a lower probability of securitization. tion of their comparative advantage across provides a new perspective on the appar- Crucially, these two portfolios have simi- industries. ent ability of firms to exploit bond-market lar observable risk characteristics and loan Bebchuk and his co-authors investi- return predictability with their financing terms. Since these findings are conditional gate the relationship between CEO central- choices. on securitization, the researchers conduct ity — the relative importance of the CEO Mayer and his co-authors explore additional analyses to address selection on within the top executive team in terms of the practice of mortgage refinancing in a the part of borrowers, lenders, or investors ability, contribution, or power — and the dynamic competitive lending model with as explanations value, performance, and behavior of pub- risky borrowers and costly default. They

32 NBER Reporter • 2008 Number 2 International Trade and Organizations

NBER’s Working Group on Trade, Multinational Production, and of Economics, and Peter K. Schott, International Trade and Organizations Diffusion” Yale University and NBER, “Intra‑Firm met in Cambridge on April 26. Pol Discussant: Costas Arkolakis, Yale Trade and Product Contractibility” Antràs, NBER and Harvard University, University and NBER Discussant: Nathan Nunn, Harvard organized the meeting and chose these University and NBER papers for discussion: Paola Conconi and Patrick Legros, ECARES, Université Libre de Bruxelles, Yongmin Chen, University of Arnaud Costinot, University of and Andrew F. Newman, Boston Colorado, Boulder; Ignatius California, San Diego and NBER, University, “Trade Liberalization and Horstmann, University of Toronto; “Heterogeneity and Trade” Organizational Choice” and James Markusen, University Discussant: Oleg Itskhoki, Harvard Discussant: Emanuel Ornelas, London of Colorado, Boulder and NBER, University School of Economics “Physical Capital, Knowledge Capital and the Choice Between FDI and Natalia Ramondo, University of Andrew B. Bernard, Dartmouth Outsourcing” Texas, and Andrés Rodriguez‑Clare, College and NBER; J. Bradford Jensen, Discussant: Kalina Manova, Stanford Pennsylvania State University and Georgetown University and NBER; University NBER, “The Gains from Openness: Stephen J. Redding, London School

Aggregate production functions are Conconi, Legros, and Newman affecting consumers, which suggests a a standard feature of the trade theorist’s embed a simple incomplete-contracts potential complementarity between trade toolbox. While this modeling device has model of organization design in a stan- policy and corporate governance policy. generated some fundamental insights, it dard two-country, perfectly-competitive Bernard and his co-authors exam- presents one obvious shortcoming: it nec- trade model to examine how the liber- ine the determinants of intra-firm trade essarily ignores any effect that the distribu- alization of product and factor markets in U.S. imports using detailed country- tion of factor endowments across agents affects the ownership structure of firms. product data. They create a new measure may have on international trade flows. In their model, managers decide whether of product contractibility based on the Costinot develops a general framework to integrate their firms, trading off the degree of intermediation in international that can shed light on these effects and dis- pecuniary benefits of coordinating pro- trade for the product. They find impor- cusses several applications. duction decisions with the private ben- tant roles for the interaction of coun- Ramondo and Rodriguez‑Clare efits of operating in preferred ways. The try and product characteristics in deter- quantify the role played by trade, multi- price of output is a crucial determinant mining intra-firm trade shares. Intrafirm national production (MP), and diffusion of this choice, because it affects the size trade is high for products with low levels of ideas in generating gains from “open- of the pecuniary benefits. In particular, of contractability sourced from countries ness”. They extend the Eaton and Kortum non-integration is chosen at “low” and with weak governance, for skill-intensive (2002) model of trade by introducing MP “high” prices, while integration occurs products from skill-scarce countries, and and diffusion of ideas. A key contribution only at moderate prices. Organizational for capital-intensive products from capi- is to model the simultaneous role of trade, choices also depend on the terms of trade tal-abundant countries. MP, and diffusion, and explore some of the in supplier markets, which affect the divi- A somewhat older literature, sup- interactions among these different chan- sion of surplus between managers. The ported by considerable empirical evidence, nels. Both trade and MP are substitutes authors obtain three main results. First, considered the multinational firm’s mode with diffusion, but the relationship among joint product and factor market integra- choice for foreign production between trade and MP is more complex. Trade and tion leads to the convergence of organiza- an owned subsidiary and a licensing con- MP are alternative ways to serve a foreign tion design across countries. Second, even tract in an environment where the firm is market, which makes them substitutes, but in the absence of factor movements, the transferring primarily knowledge-based the authors also allow for complementar- price changes triggered by liberalization assets. An important assumption in this ity by having MP rely on imports of inter- of product markets can lead to significant literature is that the relevant knowledge is mediate goods from the home country. organizational restructuring within coun- absorbed by the local manager or licensee They use trade and MP data to estimate tries. Third, the removal of barriers to over the course of time: knowledge is the model and quantify the gains from factor mobility can induce further orga- non-excludable. More recently, a num- openness, trade, MP, and diffusion. nizational changes, sometimes adversely ber of papers have adopted a property-

NBER Reporter • 2008 Number 2 33 right view of the firm, and assume the combine both forms of capital assets in a is high relative to the book value of cap- application abroad of physical capital and single model. Their model predicts that ital (high Tobin’s-Q). They believe that that ownership rights guarantee that the foreign direct investment (owned subsid- this prediction is consistent with existing owner retains full and exclusive rights to iaries) is more likely than licensing when empirical evidence. the capital should a relationship break the ratio of knowledge capital to physi- down. Chen, Horstmann, and Markusen cal capital is high, or when market value

Education Program Meets

NBER’s Program on Education, Pascaline Dupas, Dartmouth College; and Sophisticated Players in the Boston directed by Caroline M. Hoxby of and Michael Kremer, Harvard Mechanism” Stanford University, met in Cambridge University and NBER, “Peer Effects and on May 1. These papers were discusssed: the Impact of Tracking: Evidence from a C. Kirabo Jackson, Cornell University, Randomized Evaluation in Kenya” “A Little Now for a Lot Later: A John P. Papay and John B. Willet, Look at a Texas Advanced Placement Harvard University, and Richard J. Ofer Malamud, University of Chicago, Incentive Program” Murnane, Harvard University and and Kiki Pop‑Eleches, Columbia NBER, “The Consequences of High University and NBER, “The Effect of Eric P. Bettinger, Case Western School Exit Examinations for Struggling Computer Use on Child Outcomes” Reserve University and NBER, “Paying Urban Students: Evidence from to Learn: The Effect of Financial Massachusetts” Parag A. Pathak, Harvard University, Incentives on Elementary School Test and Tayfun Sonmez, Boston College, Scores” Esther Duflo, MIT and NBER; “Leveling the Playing Field: Sincere

The growing prominence of high- pass the retest. Furthermore, failing the assigned to a section. In the remaining 70 stakes exit examinations has made ques- eighth grade mathematics examination schools, students were ranked by prior tions about their effects on student out- reduces by 3 percentage points the prob- achievement (measured by their first term comes increasingly important. Exit ability that low-income urban students grades), and the top and bottom halves examinations can cause students to drop stay in school through tenth grade. There of the class were assigned to different out of school for several reasons: because are no such effects found for wealthier sections. In all 140 schools, the teach- they fear taking the test; because they fail urban students, or for suburban students ers were randomly assigned to a section. it and become discouraged; or because regardless of their family income. After 18 months, students in tracking they repeatedly retake the examination Duflo and her co-authors provide schools performed on average 0.13 stan- and cannot pass it. Papay and his co- experimental evidence on the impact of dard deviations higher than students in authors use a natural experiment to evalu- tracking primary school students by ini- the non-tracking schools. Furthermore, ate the causal effects of high-stakes testing tial achievement. In the presence of posi- students benefited from tracking at all lev- on high school completion for the cohort tive spillover effects from academically els of the distribution. In tracking schools, scheduled to graduate from Massachusetts proficient peers, tracking may be benefi- the endline test score of the initially- high schools in 2006. They find that, for cial for strong students but hurt weaker median student is as high when she was low-income urban students on the margin ones. However, tracking may help every- assigned to the “bottom” section as when of passing, failing the tenth grade math- body if heterogeneous classes make it dif- she was assigned to the “top” section. In ematics examination reduces the prob- ficult to teach at a level appropriate to non-tracking schools, where peers were ability of on-time graduation by approx- most students. The authors set up a ran- randomly assigned to each student, the imately 8 percentage points. Among domized evaluation in Kenya to evaluate authors find no impact of the average students who fail the tenth grade math- these competing claims. A total of 140 peer quality either, but they find evidence ematics examination, the low-income primary schools received funds to hire there as well that heterogeneity hurts test urban students are just as likely to retake an extra teacher and create an additional scores. the test as apparently equally skilled sub- section in first grade. In 70 (randomly Malamud and Pop‑Eleches exam- urban students, but are much less likely to selected) schools, students were randomly ine the impact of having access to a home

34 NBER Reporter • 2008 Number 2 computer on various child outcomes. To of a modified economy, where sincere stu- student information. avoid the bias attributable to non-ran- dents lose priority to sophisticated stu- In recent years, policymakers and aca- dom access to home computers, they dents. Any sophisticated student weakly demics have become increasingly inter- exploit a unique government program prefers her assignment under the Pareto- ested in applying financial incentives to that provided vouchers towards the pur- dominant Nash equilibrium of the Boston individual decisionmaking in education. chase of a personal computer for low- mechanism to her assignment under the Bettinger presents evidence from a pay- income children enrolled in Romanian recently adopted student-optimal stable for-performance program that took place public schools. Since the fixed number of mechanism. in Coshocton, Ohio. Since 2004, the vouchers were allocated based on a sim- Jackson analyzes a program that was Coshocton City Schools has provided ple ranking of family income, this pro- implemented in schools serving under- cash payments to students for successful gram allows comparisons across students privileged populations in Texas that pays completion of their standardized testing. very similar in family income and other both students and teachers for passing Students in third, fourth, fifth, and sixth respects, but who experienced markedly grades on advanced placement examina- grade who passed district and state-man- different access to having a computer at tions. Exploiting the fact that different dated standardized exams are eligible for home. In 2007, the authors conducted a schools adopted the program at differ- these rewards. Coshocton determined eli- household survey of children who par- ent times, he uses a difference-in-differ- gibility for the program using randomiza- ticipated in the program two years earlier. ences strategy. He compares the changes tion. By exploiting the randomization of Using these data, they show that children in aggregate student outcomes, before Coshocton’s incentive program, Bettinger who received a voucher were 25 percent and after adoption, for schools adopting attempts to identify the effects of stu- more likely to own a computer. Next, the AP incentive program to the changes dent incentive programs on students’ aca- they show that receipt of a voucher had a experienced over the same time period demic behavior. Additionally, the struc- large impact on time spent in front of the for carefully selected groups of compari- ture of Coshocton’s program creates some computer and decreased the amount of son schools. Adoption of the AP incen- “kinks” in student incentives. These kinks time spent watching TV and doing home- tive program is associated with a 30 per- reveal places where students who were eli- work. Children in households that won a cent increase in the number of students gible for the incentive program no longer voucher also report having lower school scoring above 1100 on the SAT or 24 on had any incentive for subsequent “high grades and lower educational aspirations. the ACT, and an 8 percent increase in stakes” work. The results present evidence Finally, there is some suggestive evidence the number of students who matriculate that young children respond to incen- that winning a voucher is associated with in college in Texas. The per-student costs tives. Math scores improved about 0.15 negative behavioral outcomes. of the program are very small relative to standard deviations higher for elemen- Empirical and experimental evidence reasonable estimates of the implied life- tary school students who were eligible for suggests different levels of sophistica- time benefits that accrue to affected stu- the program relative to the control group. tion among families in the Boston Public dents such that the APIP may amelio- There is little evidence that reading, social School student assignment plan. Pathak rate sub-optimal educational investments. science, and science test scores changed and Sonmez analyze the preference reve- Empirical evidence suggests that teachers in response to the incentive program. lation game induced by the Boston mech- and students were not simply aiming to However, students’ behavior at the spe- anism with sincere players who report maximize their rewards. Anecdotal evi- cific discontinuities in the cash incentive their true preferences and sophisticated dence suggests that the increases in AP program suggests that students respond to players who play a best response. They participation were due to better access to incentives even in ways which may not be characterize the set of Nash equilibrium AP courses, changes in teacher and peer desirable to educators. outcomes as the set of stable matchings norms towards AP courses, and better

NBER Reporter • 2008 Number 2 35 Program Meeting on Children

The NBER’s Program on Children Drinking Age, Alcohol Consumption, the Effect of Family Size on Child met in Cambridge on May 2. These and Crime” Development” papers were discussed: Manuela Angelucci, University Dan Anderber, Arnaud Chevalier, and Anna Aizer, Brown University and of Arizona; Giacomo De Giorgi, Jonathan Wadsworth, Royal Holloway, NBER, and Pedro Dal Bó, Brown Stanford University; Marcos A. University of London, “Anatomy of a University, “Love, Hate, and Murder: Rangel, University of Chicago; and Health Scare: Education, Income, and Commitment Devices in Violent Imran Rasul, University College the MMR Controversy in the UK” Relationships”(NBER Working Paper London, “Family Networks and No. 13492) School Enrollment: Evidence from a J Peter Nilsson, Uppsala University, Randomized Social Experiment” “Does a Pint A Day Affect Your Child’s Christopher Carpenter, University Pay? The Effect of Prenatal Alcohol of California, Irvine and NBER, Jorge M. Aguero and Mindy S. Marks, Exposure on Adult Outcomes” and Carlos Dobkin, University of University of California, Riverside, California, Santa Cruz and NBER, “The “Using Infertility Shocks to Estimate

Many violent relationships are char- those just under age 21. This greater alco- influences household schooling choices acterized by a high degree of cyclicality: hol consumption results in a 6 percent relates to the redistribution of resources women who are the victims of domes- increase in arrests, which is largely attrib- within the family network towards those tic violence often leave and return multi- utable to robberies, assaults, DUI, drunk- households on the margin of enrolling ple times. To explain this, Aizer and Dal enness, and disorderly conduct. These children into secondary school. Bó develop a model of time-inconsistent results imply an elasticity of about .08, Aguero and Marks introduce a preferences in the context of domestic which suggests that crime will increase novel instrument to test for the quan- violence. This time inconsistency gen- significantly if proposals to reduce the tity/quality trade-off in child develop- erates a demand for commitment. The legal drinking age are adopted. ment. Specifically, they exploit exogenous authors present supporting evidence that Angelucci and her co-authors pres- changes in family size generated by infer- women in violent relationships display ent evidence on whether and how a house- tility shocks after the first child is born. time-inconsistent preferences by exam- hold’s behavior is influenced by the pres- They show that an indicator variable for ining their demand for commitment ence and characteristics of its extended the infertility status of women is a plau- devices. They find that “no-drop” poli- family. Using household panel data from sible instrument for family size. They cies — which compel the prosecutor to the Progresa social assistance program in further show that infertility is not cor- continue with prosecution even if the vic- rural Mexico, they first exploit information related with the child’s or mother’s back- tim expresses a desire to drop the charg- on the paternal and maternal surnames of ground characteristics and has a robust es — result in an increase in reporting. heads and spouses in conjunction with impact on family size. Using data from No-drop policies also result in a decrease the patronymic naming convention in Demographic and Health Surveys in in the number of men murdered by inti- order to identify the inter- and intra-gen- Latin America, they evaluate the possi- mates, suggesting that some women in erational family links of each household ble trade-off for different aspects of child violent relationships move away from an to others in the same village. They then development, including education and extreme type of commitment device when exploit the randomized research design health. Their results indicate that, after a less costly one is offered. of the Progresa evaluation data to identify instrumenting for the number of children Carpenter and Dobkin use the exog- whether the treatment effects of Progresa using the mother’s infertility status, there enous variation in alcohol consumption transfers on secondary school enrollment is a negative and significant relationship induced by the Minimum Legal Drinking vary according to the presence and char- between the quantity and quality of chil- Age (MLDA) and data from California to acteristics of extended family members. dren for health indicators (weight-for- determine how much an increase in drink- They find that Progresa only raises sec- age and breastfeeding status) but not for ing increases criminal behavior. They find ondary enrollment among households education. that individuals just over age 21 are 32 that are embedded in an extended fam- The MMR controversy in the United percent more likely to report having con- ily network. Eligible but isolated house- Kingdom provides a case where, for a sumed alcohol in the previous month holds do not respond. The key mecha- brief period of time, some highly pub- and drink on 70 percent more days than nism through which the extended family licized research suggested that a partic-

36 NBER Reporter • 2008 Number 2 ular multi-component vaccine (MMR), MMR, parents may purchase single vac- of a sharp increase in alcohol consump- freely provided to young children, could cines privately; the MMR is the only vac- tion in the experimental regions, particu- have potentially serious side-effects. As cine for which Anderber, Chevalier, and larly among youth. Nilsson finds that the the controversy set in, uptake of the Wadsworth observe a strong effect of cohort in utero during the experiment has MMR vaccine by more educated par- income on uptake. substantially reduced educational attain- ents decreased significantly faster than by Nilsson uses a Swedish alcohol policy ment, lower earnings, and higher welfare less educated parents, turning a signifi- experiment conducted in the late 1960s dependency rates at around age 30 than cant positive education gradient into a to identify the impact of prenatal alcohol the surrounding cohorts. His results indi- negative one. Somewhat puzzling, highly exposure on educational attainment and cate that investments in early-life health educated parents also reduced their labor market outcomes. The experiment may have far reaching effects on economic uptake of other noncontroversial child- started in November 1967 and was prema- outcomes later in life. hood vaccines. As an alternative to the turely discontinued in July 1968 because

Higher Education

The NBER’s Working Group on Apply? The Effect of Application Jason M. Fletcher, Yale University, and Higher Education met in Cambridge on Costs on College Applications for Marta Tienda, Princeton University, May 2. Charles T. Clotfelter of NBER Low‑Income Students” “High School Peer Networks and and Duke University organized the meet- College Success: Lessons from Texas” ing. These papers were discussed: Basit Zafar, Northwestern University, “College Major Choice and the Gender Stephen Desjardins and Brian McCall, Kalena E. Cortes, Syracuse University, Gap” University of Michigan, “The Impact of and Isaac McFarlin, Jr., University of the Gates Millennium Scholars Program Texas‑Dallas, “College Quality and the Scott E. Carrell, UC, Davis, and on the College Enrollment, Borrowing, Texas Top 10% Plan: Implications for James E. West, United States Air Force and Work Behavior of Low‑Income Minority Students” Academy, “Does Professor Quality Minority Students” Matter? Evidence from Random Amanda D. Pallais, MIT, “Why Not Assignment of Students to Professors”

Cortes and McFarlin analyze the the student’s expected utility from attend- change induced students of all income benefits of attending selective public col- ing minus the application cost. Thus, stu- levels, but especially low‑income students, leges in the context of race-sensitive admis- dents may be deterred from applying to to send more score reports and to raise sions: they estimate how college quality selective schools because they believe the quality of the best colleges to which and the admissions-policy change, from they’re unlikely to be admitted, even they sent their scores. Many of these extra affirmative action to the Top 10% Plan, though their expected utility from attend- score reports translated into applications. affect college completion. Their results ing would be high. A policy that induces Using conservative assumptions, and Dale show that all students who attended selec- students to apply to more colleges could and Krueger’s (2002) estimates on the tive colleges have higher college comple- encourage them to send applications with return to attending a more selective col- tion rates than their comparable counter- lower marginal benefits, including appli- lege for low‑income students, Pallais esti- parts who attended less selective colleges. cations to more selective colleges. In fact, mates that if the policy were limited to Moreover, there is little evidence to sup- an exogenous decrease in the cost of send- low‑income students, the resulting wage port the minority “mismatch hypothe- ing ACT-score reports to colleges had benefits that would have accrued to those sis.” The effect of the admissions-policy exactly this effect. Students who took the students would have been 223 times the change for non-top 10 percent students is ACT before the fall of 1997 were enti- implementation cost to the ACT. She negative and three times larger for minor- tled to three free score reports and paid argues that this small cost change had ities than for non-minorities. $6 for each additional report; students such large effects on application behav- Pallais argues that the value of apply- who took the test afterwards were allowed ior either because it was a salient decrease ing to a given college is the probability four free score reports, with the same mar- in part of the application cost to “free” that the student attends the college times ginal cost for each additional report. This or because it provided a signal about

NBER Reporter • 2008 Number 2 37 how many colleges students should apply of required coursework, Carrell and West peer networks influence college achieve- to. These mechanisms potentially can be examine how professor quality affects stu- ment. Their results show that students applied to other areas, such as job search dent achievement. Introductory course with larger peer networks upon entering and consumers’ search for goods and ser- professors significantly affect student college perform better than their counter- vices, to increase social welfare. achievement in the contemporaneous and parts with smaller networks at the begin- Zafar studies the question of how follow-on related courses, but these effects ning of their freshman year. The average college majors are chosen with a focus vary greatly across subjects. The authors effects of network size on college achieve- on explaining the underlying gender gap. find that the academic rank, teaching ment are small, but a marginal increase in Since observed choices may be consistent experience, and terminal degree status of the size of same-race peer networks raises with many combinations of expectations mathematics and science professors are GPA by 0.1 point. The authors also find and preferences, he collects a unique data- negatively correlated with contemporane- some suggestive evidence that minority set of Northwestern sophomores contain- ous student achievement, but positively students with large high school peer net- ing the students’ subjective expectations related to follow-on course achievement. works reap larger academic benefits than about choice-specific outcomes. He esti- Across all subjects, student evaluations of their white counterparts. mates a model where college major choice instructors are positive predictors of con- The Gates Millennium Scholars is made under uncertainty (about per- temporaneous student achievement but (GMS) program, funded by the Bill & sonal tastes, individual abilities, and real- are poor predictors of follow-on student Melinda Gates Foundation, was estab- izations of outcomes related to the choice achievement. lished in 1999 to improve access to and of major). Enjoying coursework, enjoying Fletcher and Tienda use adminis- success in higher education for low- work at potential jobs, and approval of trative data from the University of Texas- income and high-achieving minority stu- parents are the most important determi- Austin to examine whether high school dents by providing them with full tuition nants in the choice of college major. Males peer networks upon college entry influ- scholarships and other types of support. and females have similar preferences while ence college achievement, as measured The effects of programs such as the GMS in college, but differ in their preferences by grade point average (GPA) and per- are often difficult to assess, however, in the workplace: non-pecuniary out- sistence. For each freshman cohort from because of the non-random nature of comes at college are most important in 1993 through 2003, the authors calculate selection into the program. Desjardins the decision for females, while pecuniary the number and ethnic makeup of college and McCall use regression discontinu- outcomes at the workplace explain a sub- freshmen from each Texas high school, ity to ascertain the causal effect of the stantial part of the choice for males. Zafar which they then use as a proxy for fresh- scholarship on a number of important decomposes the gender gap into differ- men “peer network.” The empirical speci- educational outcomes. While they find ences in beliefs and preferences. Gender fications include high school fixed effects some limited evidence that retention is differences in beliefs about academic abil- to control for unobservable differences higher for GMS recipients and that GMS ity and future earnings explain a small and across schools that influence both college recipients, loan debt and work hours dur- insignificant part of the gap. Conversely, enrollment behavior and academic perfor- ing college and parental contributions most of the gender gap is attributable mance. Using an IV/fixed effects strategy towards college expenses are lower for to differences in beliefs about enjoying that exploits the introduction and expan- GMS recipients than non-recipients, and coursework, and preferences for various sion of the Longhorn Scholars Program, GMS recipients are more likely than non- outcomes. which targeted low income schools with recipients to report educational aspira- Using the random assignment of col- low college traditions, the authors also tions to obtain a Ph.D. degree after com- lege students to professors in a large body evaluate whether “marginal” increases in pleting their bachelor’s degree.

38 NBER Reporter • 2008 Number 2 Behavioral Finance

The NBER’s Working Group on Discussant: Alex Edmans, University of Jialin Yu, Columbia University, Behavioral Finance met at Yale University Pennsylvania “Commonality in Disagreement and on May 3. NBER Research Associates Asset Pricing” Nicholas Barberis and Robert J. Shiller, Enrichetta Ravina, New York Discussant: Hongjun Yan, Yale both of Yale, organized the meeting. University “Love and Loans: The Effect University These papers were discussed: of Beauty and Personal Characteristics in Credit Markets” Joseph Chen, University of Southern Nicholas Barberis, and Wei Xiong, Discussant: Tanya Rosenblat, Wesleyan California; Samuel Hanson, Harvard Princeton University and NBER, University University; Harrison Hong, Princeton “Realization Utility” University; and Jeremy C. Stein, Discussant: Simon Gervais, Duke Markus K. Brunnermeier, Princeton Harvard University and NBER, “Do University University and NBER; Stefan Nagel, Hedge Funds Profit from Mutual Fund Stanford University and NBER; Distress?” Ingolf Dittmann, Erasmus University, and Lasse H. Pedersen, New York Discussant: Owen Lamont, DKR and Ernst Maug and Oliver Spalt, University and NBER, “Carry Trades Capital University of Mannheim, “Sticks or and Currency Crashes” Carrots? Optimal CEO Compensation Discussant: Nikolai Roussanov, when Managers are Loss Averse” University of Pennsylvania

Barberis and Xiong study the pos- dicts convex contracts with substantial points more than otherwise similar white sibility that, aside from standard sources option holdings that provide a stron- borrowers. However, in this sample such of utility, investors also derive utility from ger upside (“carrots”). By contrast, the personal characteristics are not, all else realizing gains and losses on individual optimal contract is concave for the stan- equal, significantly related to subsequent investments that they own. The research- dard Risk Aversion-model where it pro- delinquency rates, with the exception of ers propose a tractable model of this “real- vides a significant downside (“sticks”). beauty, which is associated with substan- ization utility,” derive its predictions, and These results suggest that loss aversion is tially higher delinquency rates. These find- show that it can shed light on a number a better paradigm for analyzing design ings suggest that the mechanism through of puzzling facts. These include the poor features of stock options and for devel- which personal characteristics affect loan trading performance of individual inves- oping preference-based valuation models supply is lenders’ preferences and percep- tors, the disposition effect, the greater than the conventional model used in the tion, rather than statistical discrimina- turnover in rising markets, the negative literature. tion, based on inferences from previous premium to volatility in the cross-section, Ravina examines whether easily experience. and the heavy trading of highly valued observable variables, such as the personal Brunnermeier and his co-authors assets. Underlying some of these applica- characteristics of a loan applicant and the document that carry traders are subject tions is one of their model’s more novel way he presents himself, affect lenders’ to crash risk: that is, exchange rate move- predictions: that, even if the form of real- decisions, once hard financial informa- ments between high-interest-rate and ization utility is linear or concave, inves- tion about credit scores, employment his- low-interest-rate currencies are negatively tors can be risk-seeking. tory, homeownership, and other financial skewed. The researchers argue that this Dittmann and his co-authors ana- information are taken into account. She negative skewness is attributable to sud- lyze optimal executive compensation studies an online lending market in which den unwinding of carry trades, which contracts when managers are loss averse. 7,321 borrowers posted 11,957 loan tend to occur in periods in which inves- They establish the general optimal- con requests that included verifiable finan- tor risk appetite and funding liquidity tract analytically and calibrate the model cial information, photos, an offered inter- decrease. Carry-trade losses reduce future to the observed contracts of 595 CEOs. est rate, and related context. Borrowers crash risk, but increase the price of crash They find that the Loss Aversion-model whose appearance was rated above aver- risk. The authors also document excess co- explains the observed structure of execu- age are 1.41 percentage points more likely movement among currencies with similar tive compensation contracts significantly to get a loan and, given a loan, pay 81 interest rate. Their findings are consistent better than the Risk Aversion-model. This basis points less than an average-looking with a model in which carry traders are holds especially for the mix of stock and borrower with the same credentials. Black subject to funding liquidity constraints. options. The Loss Aversion-model pre- borrowers pay between 139 and 146 basis Yu presents a dynamic model to dem-

NBER Reporter • 2008 Number 2 39 onstrate that, when differences of opinion variation of equity premium; and the own. The authors document two pieces over individual securities have a common common disagreement correlates with of evidence that are consistent with hedge component, the valuation of the aggre- discount-rate news rather than cash-flow funds taking advantage of this opportu- gate market can be higher than its funda- news and has explanatory power for the nity. First, in the time series, the average mental, even if all investors agree on the time-series variation of value premium. returns of long/short equity hedge funds market fundamental and the common Chen and his co-authors explore the are significantly higher in those months disagreement drives discount rate news. question of whether hedge funds engage when a larger fraction of the mutual- Using analyst forecast dispersion to mea- in front-running strategies that exploit the fund sector is in distress. Second, at the sure disagreement, Yu finds that: indi- predictable trades of others. One poten- individual-stock level, short interest rises vidual stock disagreements co-move and tial opportunity for front-running arises in advance of sales by distressed mutual the common component mean-reverts; when distressed mutual funds—those suf- funds. the common disagreement has substan- fering large outflows of assets under man- tial explanatory power for the time-series agement—are forced to sell stocks they

Health Care

The NBER’s Program on Health of Michigan, “ The Impact of Medicare “Using Price Variation to Estimate Care met in Cambridge on May 8. Kate Part D on Drug Utilization and Welfare in Insurance Markets” Bundorf, NBER and Stanford University, Out‑of‑Pocket Spending: Evidence from and Mark Pauly, NBER and University the Health and Retirement Study” Guy David, University of Pennsylvania, of Pennsylvania, organized the meeting. and Tanguy Brachet, The Children’s These papers were discussed: Jay Bhattacharya, Stanford University Hospital of Philadelphia, “Human and NBER, and Mikko Packalen, Capital Accumulation and Forgetting in Mark Duggan, University of Maryland University of Waterloo, “The Emergency Medical Services” and NBER, and Fiona Scott Morton, Other Ex‑Ante Moral Hazard in Yale University and NBER, “The Effects Health”(NBER Working Paper No. Tom Chang, MIT, and Mireille of Medicare Part D on Pharmaceutical 13863) Jacobson, University of California, Prices and Utilization” Irvine, “Ownership Status and Response Liran Einav, Stanford University; to Cost Shocks: Evidence from Helen Levy, University of Michigan Amy Finkelstein, MIT and NBER, California’s Seismic Retrofit Mandate” and NBER, and David Weir, University and Mark R. Cullen, Yale University,

The federal government began pro- specific prices and quantities sold in the medication use and out-of-pocket spend- viding insurance coverage for Medicare United States, the researchers find that ing. They compare changes in use and recipients’ prescription drug expendi- Part D substantially lowered the average spending for those who gained new Part tures on January 1, 2006 through a pro- price and increased the total use of pre- D coverage to changes for those who were gram known as Medicare Part D. Rather scription drugs by Medicare recipients. consistently covered by employer-spon- than setting pharmaceutical prices itself, The results further suggest that the mag- sored insurance or Medicare HMOs, and the government contracted with private nitude of these average effects varies across for those who had no drug coverage in insurance plans to provide this cover- drugs, as economic theory predicts. 2004 or 2006. Their results suggest that age. Enrollment in Part D was voluntary, Levy and Weir use data from the there was substantial selection into Part with each Medicare recipient allowed to 2004 and 2006 waves of the Health and D: among Medicare beneficiaries with choose from one of the private insurers Retirement Study to estimate the extent no drug coverage in 2004, those with with a contract to offer coverage in her of adverse selection into Part D — that high use and/or spending in 2004 were geographic region. Duggan and Scott is, whether beneficiaries with high exist- most likely to be enrolled in Part D in Morton evaluate the effect of this pro- ing demand for prescription drugs dis- 2006. On average, the use of prescription gram on the price and use of pharmaceu- proportionately choose to enroll in the drugs (number of prescriptions taken) tical treatments. Using data on product- program — and the impact of Part D on did not change dramatically in response

40 NBER Reporter • 2008 Number 2 to Part D. Monthly out-of-pocket drug externality of obesity on normal weight duct both individual and firm-level anal- spending for previously uninsured, newly individuals is positive as well. yses to separate skill decay from labor enrolled beneficiaries decreased, though. Einav and his co-authors show how turnover effects. They find that greater The median decrease was $30, compared standard consumer and producer theory individual experience is robustly related to median baseline spending of $100 per can be applied to estimating welfare in to improved performance at the trauma month. In contrast, median out-of-pocket insurance markets with selection. Their scene. Their estimates suggest that about spending for those consistently covered by key observation is that the same varia- a third of human capital accumlated in employer-sponsored insurance, Medicaid, tion in prices needed to trace out the a given quarter is eroded by skill decay. Medicare HMOs, or privately purchased demand curve in any applied welfare anal- They find that turnover increases organi- prescription drug insurance did not ysis also can be used to trace out how costs zational forgetting by about 25 percent change between 2004 and 2006. These vary as market participants endogenously Chang and Jacobson use a unified results are consistent with respondents’ respond to the price of insurance. With theoretical framework to model three subjective perception that Part D did not estimates of both the demand and cost popular theories of not-for-profit hos- change their use of drugs but did reduce curves, welfare analysis is straightforward. pital behavior: 1) “for-profits in dis- their out-of-pocket spending. Somewhat Moreover, because endogenous costs are guise”; 2) social welfare maximizers; and surprisingly, Medicare Part D does not the distinguishing feature of selection 3) perquisite maximizers. The authors seem to have reduced the extent of cost- models, the analysis of the cost curve also develop testable implications of a hospi- related non-compliance among those who provides a direct test for the existence tal’s response to a fixed-cost shock under previously had no drug coverage. of selection. The researchers discuss the each of these theories. They then exam- It is well known that public or data required to implement this approach ine the effect of a recent unfunded man- pooled insurance coverage can induce and then apply it using individual-level date in California that requires hospitals a form of ex-ante moral hazard: people data from a large private employer in the to retrofit or rebuild in order to comply will make inefficiently low investments United States on the health insurance with modern seismic safety standards. in self-protective activities. However, options, choices, and medical expendi- Because the majority of hospitals in the Bhattacharya and Packalen point out tures of its employees and their depen- state were built between 1940 and 1970, another ex-ante moral hazard that arises dents. They detect adverse selection in well before a sophisticated understand- through an induced innovation external- this market and estimate that its effi- ing of seismic safety, a hospital’s compli- ity: this mechanism will cause people to ciency cost, if these choices occurred in ance cost is plausibly exogenously prede- devote an inefficiently high level of self- a free market setting, would be about 0.2 termined by its underlying geologic risk. protection. As an example, they analyze percent and 2 percent of the surplus that The authors present evidence that within the innovation induced by the obesity could be generated from efficient pricing. counties seismic risk is uncorrelated with a epidemic. Obesity is associated with an They estimate that the social cost of the host of hospital characteristics, including increase in the incidence of many dis- subsidy needed to achieve the efficient ownership type. They show that hospitals eases. The induced-innovation hypothesis outcome is about an order of magnitude with higher seismic risk experience larger says that an increase in the incidence of a higher than the social welfare gain from increases in the category of spending that disease will increase technological innova- correcting the market failure. should be affected by retrofitting. Further, tion specific to that disease. The empirical David and Brachet examine the hospitals facing higher compliance costs economics literature has produced sub- underlying channels through which orga- are more likely to shut down, irrespective stantial evidence in favor of this hypoth- nizational forgetting is achieved and their of ownership type. In contrast, private esis. Bhattacharya and Packalen estimate magnitude. They develop a framework not-for-profits alone increase their mix the associations between obesity and dis- for studying the relative contributions of of profitable services, such as neonatal ease incidence; they then show that if labor turnover and skill decay to organi- intensive care days, obstetrics discharges, these associations are causal, and the phar- zational forgetting. They then apply the and MRI minutes. Government hospitals maceutical reward system is optimal, then framework to the performance of para- respond by decreasing the provision of the magnitude of the induced innovation medics by using the universe of trauma- charity care. As expected, for-profit hos- externality of obesity roughly coincides related ambulance runs in Mississippi pitals do not change their service mix in with the Medicare-induced health insur- betgween 1991 and 2005. Because para- response to this shock. These results are ance externality of obesity. The current medics are dispatched based on proxim- most consistent with the theory of not- Medicare subsidy for obesity thus appears ity and not on reputation, the relation- for-profit hospitals as perquisite maximiz- to be approximately optimal. They also ship between experience and performance ers; it allows the authors to reject two of show that the pattern of diseases for obese plausibly operates through learning-by- the leading theories of not-for-profit hos- and normal weight individuals are simi- doing rather than through selection on pital behavior — “for-profits in disguise” lar enough that the induced innovation unobservable quality. The authors con- and “pure altruism.”

NBER Reporter • 2008 Number 2 41 Insurance Workshop

The NBER’s Working Group on Dwight Jaffe, University of California, Discussant: Richard Thomas, American Insurance, directed by Kenneth A. Froot Berkeley; Howard Kunreuther; and International Group of the Harvard Business School and Erwann Michel Kerjan, University of Howard C. Kunreuther of the Wharton Pennsylvania, “The Development of Paul A. Raschky, University of School, University of Pennsylvania, met Long Term Insurance (LTI) To Address Innsbruck, “Natural Hazards, Growth, in Cambridge on May 8. These topics Catastrophe Insurance Market Failure” and Risk Transfer: An Empirical were discussed: Discussant: Robert J. Shiller, Yale Comparison between Risk‑Transfer University and NBER Mechanisms in Europe and USA” John Major, Guy Carpenter & Discussant: Jeffrey R. Brown, University Company, Inc., “On a Connection Neil A. Doherty and Anastasia of Illinois at Urbana‑Champaign and between Froot‑Stein and the de Finetti Kartasheva, University of Pennsylvania, NBER Optimal Dividends Models” and Richard D. Phillips, Georgia State Discussant: Kenneth A. Froot University, “Competition among Rating Christian Laux, Goethe University Agencies and Information Disclosure” Frankfurt, and Alex Muermann, Vienna Thomas Davidoff, University of Discussant: Christopher Lewis, The University of Economics and Business California, Berkeley, “Illiquid Housing Hartford Financial Services Group Administration, “Financing Risk as Self‑Insurance: the Case of Long Transfer under Governance Problems: Term Care” Martin F. Grace and Robert W. Klein, Mutual versus Stock Insurers” Discussant: David Moss, Harvard Georgia State University, “The Perfect Discussant: Keith Crocker, Pennsylvania University Storm: Hurricanes, Insurance and State University Regulation”

Costly external capital is one of the evidence supporting this. Households uncertainty as to whether their insur- frictions that violates the premises of the exposed to large increases in home ance will be cancelled unexpectedly, Modigliani and Miller irrelevance theo- equity in the recent housing boom were or if premiums are increased signifi- rems. Froot et. al. developed a three- relatively unlikely to add LTCI coverage cantly following the next disasters. A stage model to explain how costly exter- and relatively likely to drop coverage. two-period model illustrates when an nal capital, along with other frictions, Jaffe and his co-authors propose LT contract would be attractive to both provides an opportunity for risk man- long-term insurance (LTI) as an alter- insurers and consumers under competi- agement to create value. Major extends native to the standard annual policies tive market conditions. The authors their model to analyze risk management for homeowners’ coverage. They under- show that insurers will be willing to (in particular, reinsurance) in the con- score the need for such a contractual offer such a policy if they can charge a text of a going concern. He relates the arrangement by focusing on the chal- high enough penalty for cancellation. extended Froot model to a 50-year-old lenges facing homeowners who live in Doherty and his co-authors ana- stochastic dividend optimization prob- high-risk areas and insurers today, hav- lyze why a rating agency pools different lem, introduced by Bruno de Finetti, ing experienced the significant increase credit risks in one credit grade and how which has received growing interest in and variability in losses from natural information disclosure depends on the the recent decade. disasters in recent years. Lessons from value of information to the market. The Long-term care is one of the few the mortgage market provide a bench- authors build a model to analyze the observable triggers for home sale among mark for the development of LTI. More optimal disclosure policy of a monop- the elderly, Davidoff notes. Combined specifically, the authors here propose oly rating agency depending on the with a thin reverse mortgage mar- fixed and adjustable rate insurance con- value of information to investors; they ket, this helps to rationalize the weak tracts that have features similar to those then describe the potential market and demand for long-term care insurance in the mortgage market with penalties the strategy of the entrant. They find (LTCI). Home equity tapped in the for canceling a long-term (LT) contract that entry of symmetric rating agencies event of long-term care reduces the gain before it expires. For insurers to want results in asymmetric rating scales. This to insurance transfers during healthy to offer LTI, premiums have to reflect justifies why some companies obtain periods, in which home equity typi- the risk including the cost of capital. multiple ratings and suggests that sim- cally goes unspent. The Health and Consumers are likely to prefer LTI to ilar ratings from different agencies Retirement Study provides empirical annual policies if there is considerable may mean different credit risks. The

42 NBER Reporter • 2008 Number 2 researchers then empirically test the market adjustments. In this context, it is United States reduces personal income qualitative predictions of their model. important to understand how property by 0.3-0.4 percentage points. In addi- They use Standard and Poor’s entry insurance markets have been changing tion, the results for the U.S. sample into the insurance market that was pre- and governments have been responding suggest that counties participating in viously covered by a monopoly agency, to increased catastrophe risk. Grace and the NFIP follow a less volatile growth A.M. Best, as a natural experiment to Klein examine important market devel- path in subsequent years. Appropriate study the impact of competition on the opments and evaluate associated gov- ex-ante risk-transfer policies can largely information content of ratings. ernment policies. They find interest- mitigate these effects, while ex-post gov- The risk and cost of natural disas- ing similarities and contrasts between ernmental disaster relief tends to even ters, their effects on insurance markets, Florida and other coastal states. They enlarge the negative impact of natural and associated government policies comment on how government policies hazards on income. These results pro- yield an interesting and important story are affecting the equilibration of insur- vide useful implications for adaptation about the interplay of economics and ance markets and offer opinions on strategies against the adverse effects of politics. The intense hurricane seasons actions that are helpful and those that climate change. of 2004 and 2005 caused considerable are likely to have a negative impact on Laux and Muermann note that instability in property insurance mar- the supply of insurance and to under- mutual insurance companies and stock kets in coastal states with the greatest mine the efficient management of catas- insurance companies are different forms pressure in Florida and the Southeast. trophe risk. of organized risk sharing: policyholders Insurers have raised their rates substan- An analysis of the effects of natu- and owners are two distinct groups in tially and decreased their exposures. ral hazards on society does not depend a stock insurer, while they are one and While no severe hurricanes have struck solely on a region’s topographic or the same in a mutual. This distinction is the United States since 2005, market climatic exposure, but rather on the relevant to raising capital and to selling pressures remain strong given the high region’s institutional resilience to natu- policies under governance problems. In risk still facing coastal states. These ral processes that ultimately determine the presence of an owner-manager con- developments have generated consid- whether they will result in a natural flict, capital is costly. Free-rider and erable concern and some controversy hazard or not. Raschky provides an commitment problems limit the degree among various groups of stakehold- institutional comparison between dif- of capitalization that a stock insurer can ers. Government responses have var- ferent societal risk-transfer mechanisms obtain. The mutual form, by tying sales ied. In Florida, political pressures have against floods in Europe and the United of policies to the provision of capital, prompted a wave of legislation and reg- States. In the short run, a major flood can overcome these problems at the cost ulations to expand government under- event in a European region reduces the of less diversified owners. writing and subsidization of hurricane regional GDP by 0.4-0.6 percentage risk and to constrain insurers’ rates and points; an average flood event in the

NBER Reporter • 2008 Number 2 43 Market Microstructure

The NBER’s Working Group Shmuel Baruch, University of Utah, Amy K. Edwards and Kathleen on Market Microstructure met “Random Limit‑Orders” Weiss Hanley, U.S. Securities and in Cambridge on May 9. Working Discussant: Ioanid Rosu, University Exchange Commission, “Short Selling Group Director Bruce Lehmann of Chicago in Initial Public Offerings” of University of California, San Discussant: Rajdeep Singh, University Diego; Eugene Kandel, of Hebrew Stefan Frey, University of Tubingen, of Minnesota University, Jerusalem; and Avanidhar and Patrik Sandas, University of Subrahmanyam, of University of Virginia, “The Impact of Hidden Alessandro Beber, University of California, Los Angeles, jointly orga- Liquidity in Limit Order Books” Lausanne; Michael W. Brandt, nized the meeting. These papers were Discussant: Uday Rajan, University of Duke University and NBER; and discussed: Michigan Kenneth A. Kavajecz, University of Wisconsin, “What Can Equity Steven L. Heston, University of Andrew Ellul, Indiana University, Orderflow Tell Us about the Maryland; Robert A. Korajczyk, and Marios Panayides, University Economy?” Northwestern University; and of Utah, “Do Financial Analysts Discussant: Shane Underwood, Rice Ronnie Sadka, University of Restrain Insiders’ Informational University Washington, “Intraday Patterns in the Advantage?” Cross‑Section of Stock Returns” Discussant: Paul Irvine, University of Discussant: Julie Wu, Texas A&M Georgia University

Microstructure effects, such as bid/ these persistent intraday patterns. ciated with greater liquidity in the ask bounce, induce short-run nega- Baruch studies a pure limit-order order books, greater trading volume, tive autocorrelation patterns in asset market in which strategic risk-neutral and smaller price impact. Limit and returns while longer horizons exhibit traders can post bids and offers at vari- market order submission behavior momentum effects. Heston and his ous prices, although they run the risk changes when hidden liquidity is pres- co-authors study the term structure of of being picked off by informed trad- ent, consistent with at least some trad- microstructure effects using half-hour ers. He shows that, in equilibrium, ers being able to detect hidden liquid- observation intervals in the post-deci- limit-order traders post offers at ran- ity. The researchers estimate a model malization period. The microstructure dom prices and for random quantities. of liquidity provision that allows them induced reversal is pronounced within In equilibrium, a trader can increase to measure variations in the marginal 24 hours. Notably, they find signifi- his/her chance of making the market and total payoffs from liquidity provi- cant continuation of returns at inter- by submitting price-aggressive offers. sion in states with and without hidden vals that are multiples of a day and this The payoff associated with an aggres- liquidity. They measure the expected effect lasts for over twenty trading days. sive offer is skewed to the left (that is, surplus to providers of visible and hid- Trading volume exhibits similar pat- there are small chances of large losses den liquidity and show that often the terns, but does not explain the return to informed traders) while the payoff surplus for both types is positive. They patterns. Additionally, bid/ask spreads associated with a non-aggressive offer approximate the difference in the sur- and order imbalances do not explain is skewed to the right (that is, there are plus of liquidity demanders in order the return pattern. The return continu- small chances of large gains from unin- books with visible and hidden liquidity. ation at daily frequencies is more pro- formed traders). However, in equilib- Their results suggest that, collectively, nounced for the first and last half-hour rium, expected payoffs are the same, the liquidity providers and demand- periods. These effects are not driven by so it is optimal to choose the offer- ers may obtain a higher expected sur- firm size, systematic risk premiums, or ing price randomly. Baruch also shows plus when there is hidden liquidity inclusion in the S&P 500 index. The that a trader who exposes his/her offer in the order books. The magnitude pattern is also not driven by particular expects to lose, even though rational of the estimated gain is between 25 months of the year, days of the week, traders don’t “penny jump” exposed and 35 percent of the quoted bid-ask or turn-of-the-month effects. This sug- offers. spreads per trade. The authors inter- gests that traders may wish to time Frey and Sandas report that the pret their results as evidence that a less portfolio rebalancing to account for presence of hidden liquidity is asso- than fully transparent limit order book

44 NBER Reporter • 2008 Number 2 may enhance the welfare of market becomes larger and their trades become more likely to occur in IPOs that are participants. more profitable. They also find that an hypothesized to be price supported. Ellul and Panayides investigate important role of analysts is their abil- Thus, they conclude that short selling the competitive relationship between ity to make price-sensitive information in IPOs is neither as constrained as sug- financial analysts and firm insiders available to the market, suggesting that gested by the literature nor the result of for price-sensitive information and its analysts make a significant contribu- “naked” short sales. influence on liquidity and price dis- tion to market quality by competing Investors rebalance their port- covery. Without the presence of ana- with insiders for information. folios as their views about expected lysts, insiders have complete monopoly A number of academic papers have returns and risk change. Beber and over information, influencing market used short-sale constraints in the imme- his co-authors use empirical measures equilibrium and liquidity. If analysts diate aftermarket of IPOs to explain of portfolio rebalancing to back out really compete for information (as in short-term pricing anomalies that are investors’ views, specifically about the Fishman and Hagerty, 1992) they can subsequently reversed in the long term. state of the economy. They show that reduce insiders’ informational advan- Using newly available data, Edwards aggregate portfolio rebalancing across tage with a consequent improvement and Hanley document that short sell- equity sectors is consistent with sector of traders’ welfare (as in Holden and ing is prevalent early in the aftermarket rotation, an investment strategy that Subrahmanyam, 1992). The authors trading of IPOs. Greater short selling is exploits perceived differences in the empirically investigate this hitherto observed in IPOs with positive changes relative performance of sectors at dif- ignored role of analysts by using a sam- in offer price, high initial returns, and ferent stages of the business cycle. This ple of stocks that lost all analyst cov- large trading volume. Although large empirical foot-print of sector rota- erage, thus giving insiders complete levels of underpricing may be indicative tion has strong predictive power for monopoly over price-sensitive informa- of overvaluation, the authors find that the evolution of the economy, future tion. The departure of analysts leads short selling does not appear to limit stock returns, and future bond returns, to important changes in liquidity and observed underpricing. This is incon- even after controlling for relative sec- market equilibrium. Using a match- sistent with theoretical models that tor returns. Contrary to many theo- ing-firm methodology to address pos- predict high initial returns, in part, by ries of price formation, trading activity sible endogeneity, the authors find that assuming constraints on short sales. thus contains information that is not liquidity decreases significantly, price Using data on failures to deliver, they entirely captured by resulting relative efficiency deteriorates rapidly, infor- also test whether short sellers appear price changes. Moreover, the authors mation asymmetries increase, and insti- to avoid the perceived restrictions of find that a portfolio that mimics the tutional shareholders and liquidity- the equity lending market by “naked” observed aggregate rebalancing across motivated traders leave the stock. The short selling. However, the level of fail- equity sectors dominates the market impact of insiders’ trading on adverse ures to deliver are uncorrelated with portfolio, particularly during economic selection costs and price efficiency the amount of short selling and are downturns.

NBER Reporter • 2008 Number 2 45 Bureau Books

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

International Financial Issues in the Pacific Rim: Global Imbalances, Financial Liberalization, and Exchange Rate Policy

International Financial Issues in the the United States enjoys a seemingly sym- nomic policy issues in Asia and beyond. Pacific Rim: Global Imbalances, Financial biotic relationship with its trading part- Conference organizers and volume Liberalization, and Exchange Rate Policy, ners despite stark inequalities in the trade editors Takatoshi Ito and Andrew K. which is Volume 17 in NBER’s East Asia balance, especially with Asia. This timely Rose are NBER Research Associates in Seminar on Economics series, is now and well-informed study also debunks the the Program on International Finance and available from the University of Chicago assumed link between economic openness Macroeconomics. Ito is also a Professor Press. The price of the clothbound volume and low inflation in the region, identifies at the Graduate School of Economics, is $99.00 the serious gap between academic and pri- University of Tokyo. Rose is the B.T. The imbalanced, yet mutually ben- vate-sector researchers’ understanding of Rocca Professor of Economic Analysis eficial, trading relationship between the exchange rate volatility, and analyzes the and Policy at the Haas School of Business, United States and Asia has long been one liberalization of Asian capital accounts. University of California, Berkeley. of international finance’s most perplex- As a result, the book will have broad ing mysteries. This volume considers why implications for global trade and eco-

The following three volumes may be ordered from the University of Chicago Press Journals Division, P.O. Box 37005, Chicago, IL 60637. Via email, write to: [email protected]

Two Volumes from NBER’s International Seminar on Macroeconomics Available this Summer

The NBER’s International Seminar and productivity in the OECD; mone- are Lucrezia Reichlin, director of general on Macroeconomics (ISoM) has met tary policy’s impact on inflation and out- research at the European Central Bank and annually in Europe for the past thirty put; implications of rising government a full professor of economics at Université years. The proceedings of the 2006 con- debt; the relationship between consump- Libre de Bruxelles, and Kenneth West, the ference are now available, and the 2007 tion and labor market tightness; variation Ragnar Frisch Professor of Economics at proceedings will appear later this summer. in real wages over the business-cycle; pro- the University of Wisconsin, Madison, This series of annual volumes is now pub- duction sharing and business cycle syn- and an NBER Research Associate in the lished by the University of Chicago Press chronization in the accession countries; Programs on Monetary Economics and Journals Division. The books are priced at and pension systems and the allocation of Asset Pricing. $90.00 for clothbound editions. macroeconomic risk. The topics included inISoM 2007 are: The papers included in ISoM 2006 The editors of the 2006 volume, who interest setting and central bank transpar- discuss: the relationship between prices also organized the conference that year, ency; expectations, monetary policy, and

46 NBER Reporter • 2008 Number 2 traded goods prices; public investment omies; capital flows and asset prices; and of Economics and professor of interna- and the golden rule; the role of institu- welfare implications of financial global- tional affairs at Columbia University, and tions, confidence, and trust in financial ization without financial development. Francesco Giavazzi, professor of econom- integration within EU countries; inter- The editors of that volume, who ics at Bocconi University and a regular vis- national portfolios with supply, demand, also organized last year’s conference, are iting professor at MIT. Both are members and redistributive shocks; transmission NBER Research Associates Richard H. of the NBER’s Program on International and stabilization in closed and open econ- Clarida, the C. Lowell Harriss Professor Finance and Macroeconomics.

Tax Policy and the Economy

Tax Policy and the Economy, Volume 22 includes issues related James M. Poterba, the conference Volume 22, is now available from the to savings through tax-deferred retire- organizer and editor of this volume, is University of Chicago Press Journals ment programs, consumer choice on a professor of economics and head of Division for $60.00 (clothbound). This high-deductible health plans, financial the Economics Department at MIT, the annual series presents current academic aid applications and the tax filing pro- Director of NBER’s Public Economics research findings on taxation and gov- cess, and recent developments in corpo- Research Program, and President of the ernment spending that have both imme- rate income tax reform in the European NBER. diate bearing on policy debates and lon- Union and possible implications for the ger-term interest. United States.

NBER Reporter • 2008 Number 2 47 Nonprofit Org. U.S. Postage NBERReporter PAID National Bureau of NATIONAL BUREAU OF ECONOMIC RESEARCH Economic Research

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

Change Service Requested