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 4

Program Report

IN THIS ISSUE

Program Report The Program on Children 1 The Program on Children

Research Summaries Management And Productivity 6 Banking Crises 10 Jonathan Gruber* The Economics of Charity 14 Prices, Productivity, and Innovation 17

NBER Profiles 20 Conferences 23 NBER News 40 The impact of public policy on the well-being of children contin- Program and Working Group Meetings 41 ues to be a major area of interest for policymakers. Likewise, the eco- Bureau Books 66 nomics of children’s issues continue to be a major area of research for members of the Children’s Program at the NBER. This program brings together a group of researchers from across many different fields, including labor economics, public economics, industrial organization, econometrics, and development economics. These researchers come together to work on a diverse set of issues related to the well-being of children, and to present their work in annual meetings each spring and at the NBER’s Summer Institute each summer. Since my last report on the Children’s Program four years ago, there has been rapid growth in our roster of members, including some of the most exciting young economists in the profession. The num- ber of Working Papers in the last full year before this report, 2007, was almost 40 percent higher than in the last full year before my pre- vious report, 2003. Moreover, there has been a growing diversity of topics addressed by Program researchers, making this summary even more challenging! In this report, I focus on the contributions of this group of researchers to eight areas over the past four years: intergen- erational linkages between parent and child; the impact of early life circumstances on later child outcomes; fertility and family structure; child health; children in developing economies; public policies (par- ticularly child care and preschool) and child welfare; risky behavior by and around youths; and education.

* Gruber directs the NBER’s Program on Children and is a Professor of Economics at MIT. In this article, the numbers in parentheses refer to NBER Working Papers.

NBER Reporter • 2008 Number 4 Intergenerational Linkages NBER Reporter One of the major determinants of the well- being of children is the decisions made by, and changes imposed on, their parents. Over the past four years, a number of researchers in the Children’s Program have explored very inter- esting aspects of this intergenerational linkage. The National Bureau of Economic Research is a private, nonprofit research orga- One important area is maternal time inputs to nization founded in 1920 and devoted to objective quantitative analysis of the newborns. John Cawley and Feng Liu (13609, American economy. Its officers and board of directors are: 13600) find that employed women spend less President and Chief Executive Officer — James M. Poterba time reading to their children and helping with Controller — Kelly Horak homework, as well as less time cooking and BOARD OF DIRECTORS more use of prepared meals. However, Michael Chairman — John S. Clarkeson Baker and Kevin Milligan (13188, 13826) find Vice Chairman — Kathleen B. Cooper that expanded maternity leave, while increasing Treasurer — Robert Mednick both maternal time at home and use of breast- DIRECTORS AT LARGE feeding, did not have measurable impacts on Peter Aldrich Jessica P. Einhorn Alicia H. Munnell child development outcomes or health. Elizabeth E. Bailey Mohamed El-Erian Rudolph A. Oswald Potential improvements in child out- Richard Berner Jacob A. Frenkel Robert T. Parry comes from additional parental education or John Herron Biggs Judith M. Gueron James M. Poterba John S. Clarkeson Robert S. Hamada John S. Reed income also are important. Lucia Breirova and Don R. Conlan Karen N. Horn Marina v. N. Whitman Esther Duflo (10513) find that more educa- Kathleen B. Cooper John Lipsky Martin B. Zimmerman tion for both mothers and fathers in Indonesia Charles H. Dallara Laurence H. Meyer George C. Eads Michael H. Moskow led to lower child mortality. Gordon Dahl and Lance Lochner (11279) find that family DIRECTORS BY UNIVERSITY APPOINTMENT income increases attributable to the Earned George Akerlof, California, Berkeley Joel Mokyr, Northwestern Income Tax Credit lead to improved student Jagdish W. Bhagwati, Columbia Andrew Postlewaite, Pennsylvania Glen G. Cain, Wisconsin Uwe E. Reinhardt, Princeton test scores. And Philip Oreopolous, Marianne Ray C. Fair, Yale Nathan Rosenberg, Stanford Page, and Ann Huff Stevens (11587) find that Franklin Fisher, MIT Craig Swan, Minnesota children with a parent displaced from a job Mark Grinblatt, California, Los Angeles David B. Yoffie, Harvard Saul H. Hymans, Michigan Arnold Zellner, Chicago have lower long-run earnings themselves. Marjorie B. McElroy, Duke In related studies, Janet Currie and Enrico DIRECTORS BY APPOINTMENT OF OTHER ORGANIZATIONS Moretti (11567) find strong intergenerational Jean Paul Chavas, American Agricultural Economics Association correlations in the birth weight of mothers Gail D. Fosler, The Conference Board and children. Bruce Sacerdote (10894) com- Martin Gruber, American Finance Association pares adoptees to their biological siblings and Timothy W. Guinnane, Economic History Association Arthur B. Kennickell, American Statistical Association finds that transmission from parent to child Thea Lee, American Federation of Labor and of education, income, height, and obesity are Congress of Industrial Organizations much higher for biological children, but that William W. Lewis, Committee for Economic Development Robert Mednick, American Institute of Certified Public Accountants transmission of drinking and smoking is com- Angelo Melino, Canadian Economics Association parable in both groups. Sandra Black, Paul Harvey Rosenblum, National Association for Business Economics Devereux, and Kjell Salvanes (13336) use data John J. Siegfried, American Economic Association from Norway to document that planned larger The NBER depends on funding from individuals, corporations, and private foun- family sizes have little impact on child IQ, but dations to maintain its independence and its flexibility in choosing its research activities. Inquiries concerning contributions may be addressed to James M. that unplanned increases in family size through Poterba, President & CEO, NBER 1050 Massachusetts Avenue, Cambridge, MA twin births do lead to lower IQ among existing 02138-5398. All contributions to the NBER are tax deductible. children. Patricia Anderson, Kristin Butcher, The Reporter is issued for informational purposes and has not been reviewed by and Diane Schanzenbach (13479) study the the Board of Directors of the NBER. It is not copyrighted and can be freely repro- intergenerational correlation of obesity and duced with appropriate attribution of source. Please provide the NBER’s Public find that it has increased since the 1970s. Information Department with copies of anything reproduced. And one study provides intriguing evidence Requests for subscriptions, changes of address, and cancellations should be sent of transmission in the other direction: Ebonya to Reporter, National Bureau of Economic Research, Inc., 1050 Massachusetts Avenue, Cambridge, MA 02138-5398. Please include the current mailing label. Washington (11924) finds that legislators who

2 NBER Reporter • 2008 Number 4 have daughters are more likely to vote in erwise in no worse health. Jessica W. Other studies in this area focus on favor of women’s rights than are those Reyes (13097) shows that the state-spe- the determinants of fertility and abortion who have sons! cific variations in the reduction of lead decisions. Alma Cohen, Rajeev Dehejia, used in gasoline is strongly associated 20 and Dmitri Romanov (13700) find that Long-Run Impacts of Early years later with variations in youth crime. government subsidies for childbirth in Childhood Life Events David M. Cutler, Winnie Fung, Michael Israel significantly increased fertility. Kremer, and Monica Singhal (13539) Melissa Kearney and Phillip B. Levine In addition to the intergenerational show that malaria eradication in India (13045, 13436) find that expanded linkages between parent and child, led to long-run gains in literacy and pri- access to family planning services under another major influence on the life mary school completion rates. Douglas public insurance programs reduced teen course of individuals is their early child- Almond, Lena Edlund, and Marten births by over 4 percent, and that this was hood experiences. A particular area of Palme (13347) show that Swedish chil- accomplished via greater use of contra- innovation over the past four years has dren exposed in utero to higher radiation ception; they also find that women who been the study of interesting early life levels from the Chernobyl accident had grow up disadvantaged are much more influences and how they affect the life much worse schooling outcomes later likely to give birth as teens. Elizabeth course. Here the research has focused in life. And Almond, Edlund, Hongbin Ananat, Levine, Douglas Staiger, and I on two questions: First, how do child Li, and Junsen Zhang (13384) find that (12150) show that cohorts with higher health conditions affect the life course? fetal exposure to acute malnutrition dur- rates of abortion have much improved Oreopolous, Mark Stabile, Randy Walld, ing the Chinese famine of 1959–61 was longer-term outcomes, including more and Leslie Roos (11998) compare sib- associated with a wide range of worse education and lower odds of being a lings to show that those in poorer health outcomes later in life. single parent, while Ananat and Dan as infants have lower educational attain- Hungerman (13402) find that the avail- ment and are more likely to end up on Fertility and Family Structure ability of birth control pills led to falls in welfare. Black, Devereux, and Salvanes short-term fertility for young women. (11796, 10720) use birthweight varia- Since the inception of the Children’s tion across twin pairs to show that the Program, a major focus of the research by Child Health twin weighing less at birth has lower IQ, Program affiliates has been the determi- education, and earnings later in life; they nants and implications of fertility deci- Another major area of activity for also show that increases in family size sions and family structure; innovative the Program has been extensive inves- because of twin births lead to lower edu- work in this area has continued over the tigation of the broad determinants of cational attainment, but this appears to past four years. A number of these studies child health. Almond and Joseph J. reflect birth-order effects (later children investigate the determinants and implica- Doyle, Jr. (13877) use the fact that a receive less education) rather than fam- tions of marriage and divorce decisions. child born just before midnight receives ily size effects per se. At a cohort level, Dahl and Moretti (10281) show that almost one full day less of hospital care Carlos Bozzoli, Angus S. Deaton, and a strong preference for sons leads cou- than one born just after midnight to Climent Quintana-Domeque (12966) ples with daughters to be more likely to show that there are no health benefits find that groups with higher mortality divorce, and that unwed mothers carry- to the infant from the longer hospital rates have lower height as adults, consis- ing a boy are more likely to marry in the stay. Currie and Stabile (10435) find tent with a “scarring” effect of youth dis- future. Dahl (11328) uses variation in that Attention Deficit Hyperactivity ease. And Anne Case, Christina Paxon, state laws regulating the age at which indi- Disorder (ADHD), the most common and coauthors (12667, 13495) continue viduals are allowed to drop out of school, child mental health problem, is asso- to provide evidence that the income work, and marry to show that women ciated with delinquency, grade repeti- gradient in children’s health becomes who marry earlier in life are more likely tion, and lower test scores; these findings steeper as children age. to live in poverty. Keith Finlay and David are confirmed and extended by Jason Second, what are the long-run effects Neumark (13928) show that for moth- Fletcher and Barbara Wolfe (13474). of negative childhood shocks? Abhijit ers who are less likely to marry because of Jens Ludwig, Dave Marcotte, and Karen Banerjee, Duflo, Gilles Postel-Vinay, higher incarceration rates among poten- Norberg (12906) use data across 26 coun- and Timothy M. Watts (12895) study tial spouses, there is little negative impact tries to show that increased sales of anti- the impact of a negative income shock (and perhaps positive impacts) on the depressants are associated with lower sui- from insect attacks in the wine-grow- outcomes of their children. And Betsey cide rates. Robert Kaestner and Xin Xu ing region of France in the late nine- Stevenson and Justin Wolfers (12944) (12113) show that the increase in female teenth century; they find that children provide a comprehensive review of trends sports participation mandated by Title born in regions that suffered income in marriage and divorce and their deter- IX had strongly beneficial health impacts losses were shorter later in life but oth- minants over the past 150 years. for adolescent girls. Currie and Matthew

NBER Reporter • 2008 Number 4 3 Neidell (10251) find that reductions and Rebecca Thornton (10971) find that to disabled children) leads to increased in air pollution in California during a merit scholarship program for ado- household income and a reduction in the 1990s saved over 1000 infant lives. lescent girls in Kenya led to significant child poverty. But Joyce, Diane Gibson, Anderson and Butcher (11177) show gains in exam scores that persisted even and Silvie Colman (10796) find rela- that schools under financial pressure are after the scholarship had ended. tively limited impacts of maternal par- more likely to adopt unhealthful food Other studies in this area have ticipation in the WIC program (which policies, resulting in an increase in body explored some of the central issues in provides food to low-income pregnant mass index for youths. development. Eric Edmonds (12926) women and children) on fetal growth. A particularly important determi- reviews the explosion of work on child A particular focus of research in nant of children’s health is public insur- labor in developing countries; he also this area has been on childcare and ance policies towards children: Anna shows (10265) that income transfers to pre-school availability. Erdal Tekin Aizer (12105) shows that information households in South Africa led to less (10459) finds that childcare subsidies and administrative costs are barriers to child labor, while he and Nina Pavcnik lead to higher labor supply among sin- low-income children receiving the public (10317) show that free trade leads to gle mothers. V. Joseph Hotz and Mo insurance to which they are entitled, and higher incomes and lower child labor Xiao (11873) find that minimum qual- that reducing those barriers increases in developing nations. Ashlesha Datar, ity regulations in childcare centers deter insurance coverage and improves health. Arkadipta Ghosh, and Neeraj Sood entry and reduce availability of childcare Aizer, Currie, and Moretti (10429) (13649) show that villages in India with options. Michael Baker, Kevin Milligan, show that mandatory managed care in higher mortality rates invest less in new- and I (11832) study the introduction Medicaid programs reduced the quality borns, while Duflo and Christopher of “$5/day child care” in Quebec, and of prenatal care and increased poor birth Udry (10498) show that a shift in fam- find that it led to more use of childcare, outcomes. ily resources from men towards women higher maternal labor supply, and worse leads to more expenditures on food child behavioral outcomes. Katherine Children in Developing and education and less on alcohol and Magnusun, Christopher Ruhm, and Jane Countries tobacco. And, Grant Miller (11704) Waldfogel (10452) find that pre-kin- finds that family planning in Colombia dergarten programs are associated with A particularly exciting development led to reduced fertility and education short-run improvement in test scores, in the Children’s Program over the past and income gains for young women. but also a long-run increase in behav- several years has been the growth in stud- ioral problems. On the other hand, Jens ies exploring the welfare of children in Public Policies and Ludwig and Douglas Miller (11702) find developing economies. As reviewed by Child Welfare that availability of publicly provided pre- Duflo, Rachel Glennerster, and Kremer school through the Head Start program (333T), a novel feature of many of these There are dozens of policy interven- led to lower child mortality at ages 5–9, studies is the use of field experiments tions in the United States and elsewhere and Ludwig and Phillips (12973) find to assess the causal impact of policy that have either intended or unintended that the benefits of Head Start exceed interventions. Felipe Barrera-Osorio, consequences for child development, and the costs. David Blau and Janet Currie Marianne Bertrand, Leigh Linden, and many studies in the Children’s Program (10670) provide a comprehensive review Francisco Perez-Calle (13890) evalu- have explored the impacts of these of childcare arrangements in the United ate cash transfers conditioned on school interventions. Some of these studies States. attendance in Colombia, and find that focus on interventions targeted to low- these transfers increase attendance and income families. Neeraj Kaushal, Qin Risky Behaviors Among graduation rates; incentives for gradua- Gao, and Jane Waldfogel (12624) study and Around Youth tion and matriculation are found to be the impact of welfare reform on fam- particularly important. Banerjee, Shawn ily expenditure patterns and show that The final topic of considerable focus Cole, Duflo, and Linden (11904) dem- there is an increase in spending on work- among Children’s Program researchers onstrate strong test-score improvements related items, but not on child learn- has been risky behaviors by youths — and among children in India from programs ing or enrichment activities. Ted Joyce, by those around them who affect their providing remedial education and com- Robert Kaestner, Sanders Korenman, lives. Michael S. Visser, William T. puter-assisted learning. Duflo and Rema and Stanley Henshaw (10214) show that Harbaugh, and Naci H. Mocan (12507) Hanna (11880) show that a financial “family caps” on welfare benefits have use experimental evidence from the lab- incentive to reduce teacher absenteeism little impact on fertility decisions. Mark oratory to show that young people are in India led to both increased teacher Duggan and Melissa Kearney (11568) rational and responsive to incentives in attendance and improved student out- show that participation in the child SSI their criminal decisions, such as how comes. And Kremer, Edward Miguel, program (which provides cash resources much to steal. Mocan and Erdal Tekin

 NBER Reporter • 2008 Number 4 (12019) show that being very attractive in life. Aizer (13773) shows that inter- the Chicago Public School system do not reduces a young adult’s propensity for acting with violent peers increases vio- see any improvement in performance, criminal activity, while being unattract- lence among urban youths. But Matthew while Joshua Angrist, Eric Bettinger, and ive increases it. Mark Coppejans, Donna Gentzkow and Jesse Shapiro (12021) Michael Kremer (10713) showed very Gilleskie, Holger Sieg, and Koleman find that increased preschool exposure to positive long-run effects of school choice Strumpf (12156) show that youths’ television (because of differential intro- in Colombia. demand for cigarettes is forward look- duction rates of television in the United Another set of papers provided com- ing, and that they are less likely to smoke States) actually raises average test scores pelling evidence on the impact of a host when the price of cigarettes is more vol- among youths, and Dahl and Stefano of interventions in the educational pro- atile. Sara Markowitz, Robert Kaestner, DellaVigna (13718) find that the release cess. For example, Cecelia E. Rouse, and Michael Grossman (11378, 10459) of violent movies lowers the incidence of Alan B. Krueger, and Lisa Markhamn show that increased alcohol use among youth violence. (10315) find that a popular computer teens does not change the odds of sexual program for improving writing skills had activity, but does lead to lower contra- Education little positive impact; Brian Jacob and ceptive use; they also find that policies Lars Lefgren (13514) find that grade to reduce drinking lead to lower rates The economics of education remains retention strategies lead to an increase of sexually transmitted diseases. Pinka the main area of focus for researchers in dropping out of school; and Black, Chatterji and Jeff DeSimone (11337) affiliated with the Children’s Program. Devereux, and Salvanes (10911) find find that binge drinking by youths is Since this work will be summarized that compulsory schooling laws lead to associated with lower probabilities of independently in a Program Report on reduced teenage childbearing in both the high school graduation, while DeSimone the Education Program by its Director United States and Norway. and Amy Wolaver (11035) find that Caroline M. Hoxby, I will only briefly In the last few years there was a par- binge drinking leads to lower grades. discuss the large amount of work in this ticular focus on the determinants and Christopher Carpenter and Carolos area. implications of segregation among stu- Dobkin (13374) find that there is a This past year researchers have made dents. David Card and Krueger (10366) sharp increase in mortality right at age important contributions in five central find that the abandonment of affirma- 21 when drinking becomes legal. Tekin areas. The most activity has been around tive action in higher educational admis- and Markowitz (11238) show that sui- evaluating state and federal efforts to sions decisions led to a very large drop in cidal urges among young people decrease increase school accountability and to minority acceptance rates, but had little the investment that they make in work promote school choice. Eric A. Hanushek impact on the college-going of highly or schooling. and Margaret Raymond (10591) pro- qualified minorities. Figlio (11195) finds Much of the risk facing youths also vide an overview of state accountability that across siblings, the one whose name comes from their environment and the efforts and find improved student per- is associated more typically with lower decisions of the adults around them. formance, although papers by Neal and socio-economic status tends to be graded Jeffrey R. Kling, Jens Ludwig, and Schanzenbach (13293), David N. Figlio worse by teachers. Thomas Dee (11660) Lawrence F. Katz (10777) study the (11193), and Brian Jacob (12817) find finds that assignment to same-gender important Moving to Opportunity dem- other responses by schools and students teachers improves the performance onstration that randomly moved fami- that undercut these gains. of both boys and girls in high school, lies out of public housing projects; they The next step in educational reform, while Florian Hoffman and Oreopolous find that moving to lower-poverty areas beyond accountability, is to move to free (13182) find similar effects in a col- reduces arrests among female youth for school choice and perhaps even vouch- lege setting. Card and Jesse Rothstein both violent and property crimes, while ers, whereby parents can bring their pub- (12078) find that the black-white test for males violent crimes fall but property lic tax dollars to whatever school they score gap is highest in the most segre- crimes rise. Currie and Tekin (12171) use choose rather than just the local school. gated school districts. a variety of data strategies to find strong Justine S. Hastings, Thomas J. Kane, Finally, the Program also contin- evidence that the mistreatment of chil- and Douglas O. Staiger (11805, 12145) ues to focus on the all important higher dren leads them to later engage in crimi- and Hastings, Richard VanWeeden, and education sector. Kane (10658) finds nal activities. Doyle (13291) uses the fact Jeffrey Weinstein (12995) use unique that allowing residents of the District that children at risk of foster care are ran- data from North Carolina on parent’s of Columbia to use in-state tuition in domly assigned to case managers, some school choice preferences to explore attending colleges in Maryland and of whom have a much higher propensity the role of information in public school Virginia doubled the number of D.C. than others to place them in foster care, choice. Julie B. Cullen and Brian Jacob residents attending college in those other to show that children placed in foster care (13443) find that students winning lot- states. Oreopolous, Till Van Wachter, have somewhat higher arrest rates later teries to attend higher quality schools in and Andrew Heisz (12159) find that

NBER Reporter • 2008 Number 4 5 graduating from college in a recession And, Susan Dynarski (10357, 10470) that dramatically advance our under- lowers earnings for eight to ten years discusses the pros and cons of tax subsi- standing of child well-being. Moreover, thereafter. Angrist, Daniel Lang, and dized savings accounts for college. the findings summarized here are only Oreopolous (12790) find that financial a part of the exciting research program incentives for college students in Canada Conclusions being carried out by Program members. led to improved grades and college com- This policy-relevant work will continue pletion for female students. Rothstein The past four years have seen con- to promote our understanding of chil- and Rouse (13117) find that when a tinued growth in the areas of interest dren’s well-being in the United States U.S. college moved from loans to direct and the depth of research produced by and around the world. grants to its students, students were will- members of the Children’s Program. The ing to take much lower paying jobs. results are an important set of findings

Research Summaries

Bossonomics? The Economics Of Management And Productivity

Nick Bloom and John Van Reenen*

Management and Productivity Despite this data constraint, hetero- • Monitoring: how well do com- geneity in managerial and entrepreneur- panies track what goes on inside their Economists have long speculated on ial ability has become the foundation firms, and use this for continuous why such astounding differences in pro- for a wide range of literatures.3 In many improvement? ductivity exist between firms and plants benchmark theories, it is assumed that • Target setting: do companies set within countries, even within the same management is an unobservable factor the right targets, track the right out- narrow sector.1 While the popular press that varies across firms, driving produc- comes, and take appropriate action if the and business schools have long stressed tivity differences. In parallel, Mundlak two don’t tally? the importance of different management labelled his fixed-effect differences in • Incentives: are companies promot- practices, empirical economists have had productivity as “unobserved managerial ing and rewarding employees based on relatively little to say about manage- ability.”4 performance, and systematically trying ment.2 A major problem has been the to hire and keep their best employees? absence of high quality management Measuring Management To obtain accurate responses from data that is measured in a consistent way using Double-Blind Surveys firms, we interview production plant across firms and countries. managers using a “double-blind” tech- To address this lack of management nique: managers are not told they are * Nick Bloom is a Research Associate in data, we have been developing a method- being scored or shown the scoring grid; the NBER’s Programs on Economic ology for measuring management prac- they are only told they are being “inter- Fluctuations and Growth, Monetary tices.5 We use an interview-based evalu- viewed about management practices for a Economics, and Productivity, and an Assistant Professor of Economics at Stanford. ation tool that defines and scores from research project.” To run this blind scor- John Van Reenen is a Research Associate in one (“worst practice”) to five (“best prac- ing, we use open questions. For exam- the NBER’s Program on Labor Economics tice”) 18 basic management practices. ple, on the first monitoring question, and a Professor of Economics at the London This evaluation tool was developed by an we ask “tell me how you monitor your School of Economics. Their profiles appear international consulting firm, and scores production process,” rather than asking later in this issue. these practices in three broad areas: a closed question such as “do you moni-

 NBER Reporter • 2008 Number 4 tor your production daily [yes/no]”. We The other side of the double-blind To ensure high sample response continue with open questions targeting technique is that interviewers are not told rates and skilled interviewers, we hired actual practices and examples until the in advance anything about the firm’s per- MBA students to run interviews. We interviewer can make an accurate assess- formance. They are only provided with also obtained government endorsements ment of the firm’s practices. For example, the company name, telephone number, for the surveys in each country covered, the second question on that performance and industry. Since we randomly sample and positioned it as a “Lean manufactur- tracking dimension is “what kinds of medium-sized manufacturers (employ- ing” interview with no requests for finan- measures would you use to track perfor- ing between 100 to 10,000 workers) who cial data. These steps helped to yield a 45 mance?”. The scoring grid for this perfor- are not usually reported in the business percent response rate which was uncorre- mance tracking dimension is shown here press, the interviewers generally have no lated with the (independently collected) as an example. preconceptions about these firms. performance measures. Management Practice Dimension 4 (“Performance tracking”) Score 1 Score 3 Score 5 Scoring grid Measures tracked do not indicate Most key performance indicators Performance is continuously directly if overall business objectives are are tracked formally. Tracking is over- tracked and communicated, both for- being met. Tracking is an ad-hoc process seen by senior management. mally and informally, to all staff using a (certain processes aren’t tracked at all). range of visual management tools.

Sample firm A manager tracks a range of measures At a firm every product is bar-coded A firm has screens in view of every when he does not think that output is suffi- and performance indicators are tracked line, to display progress to daily tar- cient. He last requested these reports about throughout the production process. get and other performance indicators. 8 months ago and had them printed for a However, this information is not commu- The manager meets daily with the shop week until output increased again. Then nicated to workers floor to discuss performance metrics, and he stopped and has not requested anything monthly to present a larger view of the since. company goals and direction. He even stamps canteen napkins with perfor- mance achievements.

Management Practices an important role in determining coun- ation is far greater than cross-country across Firms and Countries try-level productivity. At the firm-level, variation. This figure also highlights that better management practices are strongly U.S. firms have the highest average man- The bar chart in Figure 1 (page 8) associated with higher firm-level produc- agement score because they have almost plots the average management practice tivity, profitability, and survival 7, sug- no density of firms with management score across countries from the 6,000 gesting they could play an equally impor- practices below two. Thus, the infamously interviews we carried out in survey waves tant role in country-level productivity. badly managed paper-supply firm from 2004 to 2008. This shows the Better management is also linked with “Dunder-Mifflin” from the TV show United States has the highest manage- improved employee work-life-balance and “The Office” is thankfully a rare U.S. ment practice scores on average 6, with lower energy use, suggesting better man- exception.9 In comparison India, which the Germans, Japanese, and Swedes next, agement does not come at the expense of has the lowest cross-country management followed by a block of mid-European worker welfare or more pollution.8 score, has a large mass of firms with countries (France, Italy, the United Of course the key question is why are extremely poor management practices Kingdom, and Poland), with Southern management practices different across (scores of two or less). Europe and developing countries Brazil, countries? Figure 2 (page 8) plots the This raises two key questions which China, Greece, and India at the bottom. firm-level histogram of management prac- we are currently working on: Why do In one sense this is not surprising because tices by country, and shows that manage- these variations in management practices it approximates the cross-country spread ment practices display tremendous exist, and to what extent do variations of productivity. But in another sense it within-country variation. So, much like in management practices actually cause suggests management practices could play productivity figures, within-country vari- variations in productivity?

NBER Reporter • 2008 Number 4 7 Figure 1: Management Practice Scores Across Countries

US Germany Sweden Japan Great Britain France Italy Poland Brazil Greece China India

2.5 2.7 2.9 3.1 3.3 3.5 Country average management scores, from 1 (worst practice) to 5 (best practice)

Averages taken across all firms within each country. 4423 observations in total.

Figure 2: Management Practice Scores Across Firms

Brazil China France Germany 1 .5 0 s m

r Great Britain Greece India Italy i 1 F

f o

n o i t .5 c a r F 0

Japan Poland Sweden US 1 .5 0 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

Firm level average management scores, from 1 (worst practice) to 5 (best practice)

 NBER Reporter • 2008 Number 4 Why do Management the answer was often selection by eldest NBER Working Paper No. 6120, August Practices Vary So Much across son, while in others countries such as the 1997, and Review of Economics and Firms and Countries? United States and Sweden this was very Statistics, LXXXIII (3), (2001), pp.434– rare. A number of factors, including tradi- 45. We have identified three key factors tions over leadership succession, estate tax 3 For example, R. Lucas, “On the Size that appear to play an important role in breaks, and the external market for CEOs Distribution of Business Firms”, Bell shaping management practices — com- appear to drive these differences. Journal of Economics, IX (2), (1978), pp. petition, family ownership, and multina- Multinational and Export Status also 508–23; B. Jovanovic “Selection and the tional status. appears to play an important role in deter- Evolution of Industry”, Econometrica, L Product market competition is asso- mining a firm’s management practices. (3), (1982), pp. 649–70; H. Hoppenhayn, ciated with significantly better manage- One stylized fact is that multinationals “Entry, Exit and Firm Dynamics in Long- ment practices. In particular, the tail of have good management practices wherever Run Equilibrium”, Econometrica, LX badly managed firms shrinks in highly they are located — so multinationals in the (5), (1992), pp. 1127–50; or M. Melitz, competitive markets. Thus, the competi- United States, Brazil, and India all appear “The Impact of Trade on Intra-industry tive product markets of the United States to be well run. A second stylized fact is Reallocations and Aggregate Productivity explain much of its lack of badly managed that some countries have relative man- Growth”, NBER Working Paper No. 8881, firms. In contrast, many product markets agerial strengths and weaknesses — for April 2002, and Econometrica, LXXI in India have limited competition because example, the Japanese are better at moni- (6), (2003), pp. 1695–725. of entry barriers, trade regulations, and toring and the Americans at incentives/ 4 Y. Mundlak, “Empirical Production high transportation costs, enabling badly people management — and their multi- Function Free of Management Bias”, managed firms to persist. nationals take this with them abroad. In Journal of Farm Economics, XLIII (1) We are currently investigating the other work with Rafaella Sadun, we show (1961), pp. 44–56. mechanisms through which competition that U.S. multinational affiliates located 5 See N. Bloom and J. Van Reenen, works to improve management. One pos- in Europe were able to use their mana- “Measuring and Explaining Management sibility is Darwinian selection — high lev- gerial advantage to make better use of Practices across Firms and Countries”, els of competition should drive badly information technology to raise produc- NBER Working Paper No. 12216, managed firms out of business more tivity.11 We argue that these managerial May 2006, and Quarterly Journal of quickly. Another is by inducing higher differences could account for about half Economics, November 2007, pp. 1351– levels of effort — tough product market of the superior productivity growth per- 1408; and N. Bloom, R. Sadun, and J. competition may lead managers to work formance in the United States relative to Van Reenen, “Americans Do I.T. Better: harder as the stakes are higher (slacking Europe in the decade after 1995. A third U.S. Multinationals and the Productivity is more likely to lead to losses of market stylized fact is that among domestic firms, Miracle”, NBER Working Paper No. share and bankruptcy). As we follow up those that export are better managed than 13085, May 2007. the initial cross-sectional firm surveys to non-exporters. Interestingly, these styl- 6 We should note that all the co-authors convert this into panel data, we can inves- ized facts are broadly consistent with a of the management research are European. tigate these different mechanisms. wide range of recent trade models.12 7 See N. Bloom and J. Van Reenen, Firms that are both family owned and “Measuring and Explaining Management family managed tend to be badly run on 1 See, for example, L. Foster, J. Practices across Firms and Countries.” average.10 This is true even after includ- Haltiwanger, and C. Syverson, 8 See N. Bloom, T. Kretschmer, and J. ing a battery of econometric controls for “Reallocation, Firm Turnover, and Van Reenen, “International Differences country, industry, firm size, human and Efficiency: Selection on Productivity or in the Business Practices and Productivity physical capital intensity. Looking at these Profitability?” NBER Working Paper No. of Firms”, forthcoming in an NBER book, family firms in more detail, it appears the 11555, August 2005. R. Freedman and K. Shaw, eds.; and N. worse managed firms are those in par- 2 Notable exceptions include A. Bartel, Bloom, C. Genakos, R. Martin, and R. ticular that hand down the position of C. Ichinowski, and K. Shaw, “How Does Sadun, “Modern Management: Good for CEO using the ancient practice of pri- Information Technology Really Affect the Environment or Just Hot Air?”, NBER mogeniture (succession of the eldest son). Productivity? Plant-Level Comparisons of Working Paper No. 14393, October 2008. To elicit this information, we asked the Product Innovation, Process Improvement, 9 We should note “The Office” originated plant managers the question: “How was and Worker Skills”, NBER Working Paper in the UK, where a much larger tail of the CEO chosen, was he selected as the No. 11773, November 2005; and S. badly run firms exists. Interestingly, the eldest son or by some other mechanism?” Black and L. Lynch, “How to Compete: UK was also the source of “Fawlty Towers”, Surprisingly, in many countries, including The Impact of Workplace Practices and another classic show about bad manage- Brazil, India, and the United Kingdom, Information Technology on Productivity”, ment practices.

NBER Reporter • 2008 Number 4 9 10 M. Bennedsen, K. Nielsen, F. Perez- 11 See N. Bloom and J. Van Reenen, “Export Versus FDI with Heterogeneous Gonzales, and D. Wolfenzon, “Inside “Measuring and Explaining Management Firms”, NBER Working Paper No. 9439, the Family Firm: the Role of Families in Practices across Firms and Countries”, and January 2003, and American Economic Succession Decisions and Performance”, N. Bloom, R. Sadun, and J. Van Reenen, Review, Vol. 94, 2004, pp. 300–16; and NBER Working Paper No. 12356, July “Americans Do I.T. Better: US Multi­ A. Burstein and A. Monge, “Foreign 2006, uses gender of the founder’s first- nationals and the Productivity Miracle”. Know-how, Firm Control, and the Income born child as an instrumental variable for 12 See M. Melitz, “The Impact of of Developing Countries”, NBER Working family-run firms. They find that the OLS Trade on Intra-industry Reallocations Paper No. 13073, May 2007, forthcoming estimates actually underestimate how bad and Aggregate Productivity Growth”; in Quarterly Journal of Economics. family firms are. E. Helpman, M. Melitz, and S. Yeaple,

Banking Crises1

Charles W. Calomiris*

The current global financial crisis distress often do not coincide. For exam- clear (loan losses reflected complex con- grew out of banking losses in the United ple, in the rural United States during the nections to securities market transactions, States related to subprime lending. How 1920s, large declines in agricultural prices with uncertain consequences for some well do economists understand the ori- cause many banks to fail, often with high New York banks). gins of such crises and how they spread? losses to depositors, but those failures Sometimes, large loan losses and con- Was this crisis something new or a replay were not associated with systemic panics.2 fusion regarding their incidence occur of familiar historical phenomena? Will In 1907, the opposite pattern was visible. together. In Chicago in mid-1932, for policy interventions be able to mitigate its The United States experienced a systemic example, large losses resulted in many costs? The history of banking crises pro- panic, originating in New York, which was failures and also in widespread withdraw- vides informative perspectives on these precipitated by small aggregate shocks but als from banks that did not ultimately and other important questions. had large short-term systemic effects asso- fail. Despite the confusion about the ciated with widespread withdrawals of incidence of the shock, and the conse- Crises Are Not All the Same deposits. Although some banks failed in quent widespread temporary disruptions 1907, failures and depositor losses were to the financial system, the banks that When considering the history of not much higher than in normal times.3 failed were exogenously insolvent; solvent banking crises, it is useful to distinguish That crisis was resolved only after banks Chicago banks experiencing withdrawals between two phenomena associated with had suspended convertibility and after did not fail. In other episodes, however, banking system distress: exogenous shocks uncertainty about the incidence of the bank failures may have reflected illiquid- that produce insolvency, and pressures on shock had been resolved. ity resulting from runs, rather than exog- banks that arise from rapid withdrawals The central differences between these enous insolvency.4 of debt or failures to rollover debt dur- two episodes relate to the information Today’s financial turmoil is closer to ing “panics.” These two contributors to about the shocks producing loan losses. the Chicago experience in 1932 than to In the 1920s, the shocks were loan losses either the banking shocks of the 1920s or * Calomiris is a Research Associate in the 5 NBER’s Program on the Development in agricultural banks, geographically iso- those of 1907. The shock that prompted of the American Economy and the lated and fairly transparent. Banks failed the turmoil was of moderate size (sub- Henry Kaufman Professor of Financial without subsequent system-wide con- prime and Alt-A loans totaled roughly $3 Institutions at Columbia University’s cerns. During 1907, although the ulti- trillion, including those on the balance School of International and Public Affairs. mate losses for New York banks were sheets of Fannie Mae and Freddie Mac, His profile appears later in this issue. small, the incidence of the shock was not and total losses are likely to generate total

10 NBER Reporter • 2008 Number 4 losses of roughly half a trillion dollars), most of the world, banks were perceived costs of establishing additional branches), and its consequences were significant for as stable; large losses from failed banks and able to coordinate the banking sys- both solvent and insolvent banks. Unlike were uncommon; banking panics were tem’s response in moments of confusion the Chicago Panic, today’s turmoil prob- not seen as a great risk; and there was lit- to avoid depositor runs (the number of ably has produced the failures of finan- tle perceived need for formal safety nets banks was small, and assets were highly cial institutions that were arguably sol- (for example, deposit insurance). In many concentrated in several nationwide insti- vent prior to their liquidity problems (for countries, ad hoc policies among banks, tutions). Outside the United States, coor- example, Bear Stearns). and sometimes including central banks, dination among banks facilitated systemic Banking crises can differ according to to coordinate bank responses to liquidity stability by allowing banks to manage whether they coincide with other finan- crises (as, for example, during the failure incipient panic episodes to prevent wide- cial events. Banking crises coinciding with of Barings investment bank in London spread bank runs. In Canada, the Bank currency collapse are called “twin” cri- in 1890), seemed adequate for preventing of Montreal occasionally would coordi- ses (as in Argentina in 1890 and 2001, systemic costs from bank instability. nate actions by the large Canadian banks Mexico in 1995, and Thailand, Indonesia, The unusual experience of the United to stop crises before the public was even and Korea in 1997). A twin crisis can States was a contributor to changes in aware of a possible threat.8 reflect two different chains of causation: thinking which led to growing concerns The United States, however, was an expected devaluation may encourage about banks runs, and the need for aggres- unable to mimic this behavior on a deposit withdrawal to convert to hard cur- sive safety-net policies to prevent or mit- national or regional scale. U.S. law pro- rency before devaluation (as in the United igate runs. In retrospect, the extent to hibited nationwide branching, and most States in early 1933); or, a banking crisis which U.S. banking instability informed states prohibited or limited within- can cause devaluation, either through its thinking and policy outside the United state branching. U.S. banks, in contrast adverse effects on aggregate demand or States seems best explained by the size and to banks elsewhere, were numerous (for by affecting the supply of money (when a pervasive influence of the United States; example, numbering more than 29,000 in costly bank bailout prompts monetization in fact, the U.S. crises were unique and 1920), undiversified, insulated from com- of government bailout costs). Sovereign reflected peculiar features of U.S. law and petition, and unable to coordinate their debt crises can also contribute to bank banking structure. behavior to prevent panics. distress when banks hold large amounts The U.S. Panic of 1907 (the last of The structure of U.S. banking explains of government debt (for example, in the a series of similar U.S. events, including why the United States uniquely had bank- banking crises in the United States in 1857, 1873, 1884, 1890, 1893, and 1896) ing panics in which runs occurred despite 1861, and in Argentina in 2001).6 precipitated the creation of the Federal the health of the vast majority of banks. Reserve System in 1913 as a means of The major U.S. banking panics of the Shifting Perceptions enhancing systemic liquidity, reducing postbellum era (listed above) all occurred of Banking Crises and the probability of systemic depositor runs, at business cycle peaks, and were pre- the Desirability of and mitigating the costs of such events.7 ceded by spikes in the liabilities of failed This innovation was specific to the United businesses and declines in stock prices; Government Protection States (other countries either had estab- indeed, whenever a sufficient combina- The consensus views regarding bank- lished central banks long before, often tion of stock price decline and rising lia- ing crises’ origins (fundamental shocks with other purposes in mind, or had not bilities of failed businesses occurred, a versus confusion), the extent to which established central banks), and reflected panic always resulted.9 Owing to the U.S. crises result from unwarranted runs on the unique U.S. experience with panics — banking structure, panics were a predict- solvent banks, the social costs attending a phenomenon that the rest of the world able result of business cycle contractions runs, and the appropriate policies to limit had not experienced since 1866, the last that, in other countries, resulted in an the costs of banking crises (government British banking panic. orderly process of financial readjustment. safety nets and prudential regulation) For example, Canada did not suf- The United States, however, was not have changed dramatically, and more fer panics like those of the United States the only economy to experience occa- than once, over the course of the nine- and did not establish a central bank until sional waves of bank failures before World teenth and twentieth centuries. Historical 1935. Canada’s early decision to permit War I. Nor did it experience the high- experience played a large role in chang- branch banking throughout the country est bank failure rates, or banking system ing perspectives toward crises, and the ensured that banks were geographically losses of that era. None of the U.S. bank- U.S. experience had a disproportionate diversified and thus resilient to large sec- ing panics of the pre-World War I era saw influence on thinking. Although pan- toral shocks (like those to agriculture in nationwide banking distress (measured ics were observed throughout world his- the 1920s and 1930s), able to compete by the negative net worth of failed banks tory (as early as Hellenistic Greece, and in through the establishment of branches relative to annual GDP) greater than the Rome in 33 A.D.), prior to the 1930s, in in rural areas (because of low overhead 0.1 percent loss of 1893. Losses were gen-

NBER Reporter • 2008 Number 4 11 erally modest elsewhere, but Argentina ing restrictions on bank activities (the (an opposition shared by the Fed, the in 1890 and Australia in 1893, the most separation of commercial and investment Treasury, and Senator Carter Glass). severe cases of banking distress during the banking, subsequently reversed in the Deposit insurance was seen as undesirable 1875–1913 era, suffered losses of roughly 1980s and 1990s), and government insur- special-interest legislation designed to 10 percent of GDP. Losses in Norway ance of deposits. Targeted bank recapi- benefit small banks. Numerous attempts in 1900 were roughly 3 percent of GDP talizations were also implemented via the to introduce it failed to attract support and in Italy in 1893 roughly 1 percent Reconstruction Finance Corporation, in in the Congress.13 Deposit insurance of GDP. With the possible exception of an innovative program that proved quite removes depositors’ incentives to monitor Brazil (for which data have yet to be col- successful at little cost to taxpayers.11 and discipline banks, and frees bankers to lected to measure losses), there were no Academic perspectives on the take imprudent risks (especially when they other cases in 1875–1913 in which bank- Depression fueled the portrayal of banks have little or no remaining equity at stake, ing loss exceeded 1 percent of GDP. as crisis-prone. The most important of and see an advantage in “resurrection risk Loss rates tended to be low because these was the treatment of the 1930s taking”). The absence of discipline also banks structured themselves to limit banking crises by Milton Friedman promotes banker incompetence, which their risk of loss, by maintaining ade- and Anna J. Schwartz in their book, A leads to unwitting risk taking. quate equity-to-assets ratios, sufficiently Monetary History of the United States Empirical research on the banking low asset risk, and adequate asset liquid- (1963). Friedman and Schwartz argued collapses of the last two decades of the ity. Most importantly, market discipline that many solvent banks were forced to twentieth century has produced a consen- (the fear that depositors would withdraw close as the result of panics, and that fear sus that the greater the protection offered their funds) provided incentives for banks spread from some bank failures to pro- by a country’s bank safety net, the greater to behave prudently. The picture of small duce failures elsewhere. Their views that the risk of a banking collapse.14 Studies depositors lining up around the block banks were inherently unstable, that irra- of historical deposit insurance reinforce to withdraw funds has received much tional depositor runs could ruin a bank- that conclusion. Indeed, the basis for the attention, but perhaps the more impor- ing system, and that deposit insurance opposition to federal deposit insurance in tant source of market discipline was the was a success, were particularly influential the 1930s was the disastrous experimen- threat of an informed (often “silent”) run coming from economists known for their tation with insurance in several U.S. states by large depositors (often other banks). skepticism of government interventions. during the early twentieth century, which Banks maintained relationships with each Since the publication of A Monetary resulted in banking collapses in all the other through inter-bank deposits and History of the United States, however, other states that adopted insurance.15 the clearing of public deposits, notes, and scholarship has led to important qualifica- bankers’ bills. Banks often belonged to tions of the Friedman-Schwartz view of Macroeconomic clearinghouses that set regulations and 1930s’ bank distress, and particularly of Consequences monitored members’ behavior. A bank the role of panic in producing distress.12 that lost the trust of its fellow bankers Detailed studies of particular regions and As macroeconomists increasingly could not long survive.10 banks’ experiences do not confirm the have emphasized, when banks respond This perception of banks as stable, as view that panics were a nationwide phe- to losses, deposit outflows, and increased disciplined by depositors and inter-bank nomenon during 1930 or early 1931, or risk of loan loss by curtailing the supply arrangements to act prudently, and as an important contributor to nationwide of credit, that can aggravate the cycli- unlikely to fail, was common prior to the distress until very late in the Depression cal downturn, magnifying declines in 1930s. The banking crises of the Great (that is, early 1933). Regional bank dis- investment, production, and asset prices, Depression changed that perception. U.S. tress often was localized and traceable to whether or not bank failures occur. Recent bank failures resulted in losses to deposi- fundamental shocks to the values of bank research explores the linkages among bank tors in the 1930s in excess of 3 percent loans. Indeed, recent scholarship in bank- credit supply, asset prices, and economic of GDP. Bank runs, bank holidays (local ing has emphasized that government pro- activity, and focuses in particular on the and national government-decreed peri- tections of banks, including the U.S. fed- adverse macroeconomic consequences of ods of bank closure to attempt to calm eral deposit insurance, can undermine “credit crunches” that result from banks’ markets and depositors), and widespread market discipline of bank risk taking, and attempts to limit their risk of failure.16 bank closure suggested a chaotic and vul- contribute significantly to the risk of a This new literature provides evidence nerable system in need of reform. The banking crisis. in support of a “shock-and-propagation” Great Depression saw an unusual raft of Interestingly, the theory behind the approach to understanding the contribu- banking regulations and interventions, problem of destabilizing protection has tion of financial crises to business cycles. especially in the United States, many of been well-known for over a century, and This approach has empirical implications which have subsequently been discredited was the basis for Franklin Roosevelt’s that can distinguish it from other theo- as unwarranted and undesirable, includ- opposition to deposit insurance in 1933 ries of the origins, propagation, and con-

12 NBER Reporter • 2008 Number 4 sequences of bank distress. For example, Economies Need Deposit Insurance? 93, December 2003a, pp. 1615–47, and this approach helps us to understand why Lessons from the U.S. Agricultural Boom C. W. Calomiris and B.Wilson, “Bank it was that during previous severe banking and Bust of the 1920s,” in If Texas Were Capital and Portfolio Management: The panics in the United States, in the face of Chile: A Primer on Banking Reform, P. 1930s ‘Capital Crunch’ and Scramble to severe asymmetric-information problems Brock, ed., ICS Press, 1992, pp. 237–314, Shed Risk,” Journal of Business 77, July and associated adverse-selection costs of 319–28, and 450–8. 2004, pp. 421–55. A modern-day example potential bank equity offerings in the 3 C. W. Calomiris and G. Gorton, of market discipline (Argentina 1991–9) wake of banking crises, it was prohibitive “The Origins of Banking Panics: Models, is analyzed in C. W. Calomiris and A. for banks to issue new equity in support Facts, and Bank Regulation,” in Financial Powell, “Can Emerging Market Bank of continuing lending. Interestingly, fol- Markets and Financial Crises, R. G. Regulators Establish Credible Discipline?” lowing the subprime shock, nearly $500 Hubbard, ed., University of Chicago Press, NBER Working Paper No. 7715, May billion of new capital was raised prior to 1991, pp. 109–73. See also R.F. Bruner 2000, in Prudential Supervision: What any announcement of public injections and S. D. Carr, Money Panic: Lessons Works and What Doesn’t, F. S. Mishkin, of funds. This unusual behavior reflected from the Financial Crisis of 1907 (2007). ed., University of Chicago Press, 2001, improvements in the structure of the U.S. 4 C. W. Calomiris and J. R. Mason, pp.147–96. banking system since the 1980s, which “Contagion and Bank Failures During the 11 See J. R. Mason, “Do Lender of Last resulted from nationwide branching and Great Depression: The June 1932 Chicago Resort Policies Matter? The Effects of the diversification of banking income Banking Panic,” NBER Working Paper Reconstruction Finance Corporation through the deregulation of bank activi- No.4924, November 1994, and American Assistance to Banks During the Great ties, which mitigated problems of adverse Economic Review 87, December 1997, Depression,” Journal of Financial selection (in comparison with the 1930s pp. 863–83. Services Research 20, 2001, pp. 77–95; or the 1980s). Although it is sometimes 5 C. W. Calomiris, “The Subprime C. W. Calomiris and J. R. Mason, wrongly believed that deregulation pro- Turmoil: What’s Old, What’s New, and “How to Restructure Failed Banking moted the recent instability of banks, in What’s Next,” in “Maintaining Stability Systems: Lessons from the U.S. in the fact subprime lending and securitization in a Changing Financial System”, Federal 1930s and Japan in the 1990s,” NBER were in no way linked to deregulation, Reserve Bank of Kansas City Symposium Working Paper No. 9624, April 2003, and whatever prudential regulatory fail- Proceedings, 2008. and in Governance, Regulation, and ures attended the subprime boom and 6 For a discussion of the Argentine expe- Privatization in the Asia-Pacific Region, bust, the last decade has seen substantial rience, see C. W. Calomiris, “Devaluation T. Ito and A. Krueger, eds., University of increases in those regulations, not a relax- with Contract Redenomination in Chicago Press, 2004, pp.375–420. ation of prudential regulation.17 Argentina,” NBER Working Paper No. 12 J. L. Lucia, “The Failure of the The shock-and-propagation approach 12644, October 2006, and Annals of Bank of United States: A Reappraisal,” to understanding the origins and trans- Finance 3, January 2007, pp.155–92. Explorations in Economic History mission of banking crises also implies that 7 The national banking era panics 22, 1985, pp. 402–16; E. Wicker, The regulatory policy and policy interventions are discussed in Calomiris and Gorton Banking Panics of the Great Depression, that are targeted to respond to the shocks (1991); the Panic of 1857 is treated in C. Cambridge University Press, 1996; buffeting banks (like the bank recapi- W. Calomiris and L. Schweikart, “The Calomiris and Mason (1997); Calomiris talizations recently employed by the G7 Panic of 1857: Origins, Transmission, and Mason (2003a). countries) can be used very effectively to and Containment,” Journal of Economic 13 C. W. Calomiris and E. N. White, offset the harmful macroeconomic con- History, December 1991, pp. 807–34. “The Origins of Federal Deposit sequences of shocks to banks’ balance 8 C. W. Calomiris, US Bank Insurance,” in The Regulated Economy: sheets, just as the Reconstruction Finance Deregulation in Historical Perspective, A Historical Approach to Political Corporation’s preferred stock purchases Cambridge University Press (2000). Economy, C. Goldin and G. Libecap, eds., helped to stabilize the banking sector and 9 Calomiris and Gorton (1991). University of Chicago, 1994. restart the flow of credit after 1933.18 10 For a theoretical perspective on mar- 14 See, for example, J.Barth, G. ket discipline, see C. W. Calomiris and Caprio, and R.Levine, Rethinking 1 A longer treatment of many of the C. M. Kahn, “The Role of Demandable Bank Regulation: Till Angels Govern, issues surveyed in this article appears as C. Debt in Structuring Optimal Banking Cambridge University Press, 2006. W. Calomiris, “Bank Failures in Theory Arrangements,” American Economic 15 C. W. Calomiris, “Is Deposit Insurance and History: The Great Depression Review 81, June 1991, pp. 497–513. For Necessary? A Historical Perspective,” and Other ‘Contagious’ Events,” NBER historical evidence, see C. W. Calomiris Journal of Economic History, June 1990, Working Paper No. 13597, November and J. R. Mason, “Fundamentals, pp. 283–95. 2007. Panics and Bank Distress During the 16 B. S. Bernanke, “Nonmonetary Effects 2 C. W. Calomiris, “Do ‘Vulnerable’ Depression,” American Economic Review of the Financial Crisis in the Propagation

NBER Reporter • 2008 Number 4 13 of the Great Depression,” NBER Working Factors in the Great Depression,” Journal F. Allen and D. Gale, “Financial Fragility, Paper No. 1054, 1983, and American of Economic Perspectives, 7, Spring Liquidity, and Asset Prices,” Journal of Economic Review 73, June 1983, pp. 1993, pp. 61–85; C. W. Calomiris and the European Economic Association 2, 257–76; B. S. Bernanke and M. Gertler, J.R. Mason, “Consequences of Bank 2004, pp.1015–48. “Financial Fragility and Economic Distress During the Great Depression,” 17 Calomiris (2008). Performance,” NBER Working Paper No. NBER Working Paper No. 7919, 18 Mason (2001) and Calomiris and 2318, July 1987, and Quarterly Journal September 2000, and American Mason (2004). of Economics 105, February 1990, pp. Economic Review 93, June 2003b, pp. 87–114; C. W. Calomiris, “Financial 937–47; Calomiris and Wilson (2004);

Using Field Experiments in the Economics of Charity

John A. List*

The experimental approach in scien- economics, field experiments occupy an theory, tests that explore issues in cost/ben- tific inquiry is commonly traced to Galileo important middle ground between labo- efit analysis and preference elicitation, tests Galilei, who pioneered the use of quanti- ratory experiments and studies that use that explore competitive market theory tative experiments to test his theories of naturally occurring field data.3 This is con- in the field, tests of alternative incentive falling bodies.1 Extrapolating his experi- venient because, on the one hand, eco- schemes in developing nations, and tests mental results to the heavenly bodies, he nomic theory is inspired by behavior in the of information assimilation among profes- pronounced that the services of angels field, so we would like to know if results sional financial traders.4 were not necessary to keep the planets from the laboratory domain are transfer- In the remainder of this research sum- moving, enraging the Church and disciples able to field environments. Alternatively, mary, I will summarize field experiments of Aristotle alike. For his efforts, Galileo because it is sometimes necessary to invoke within the realm of the economics of char- is now viewed as the Father of Modern strict assumptions to achieve identification ity, with an emphasis on my work, com- Science. Since the Renaissance, fundamen- using naturally occurring data, we wonder pleted with several colleagues. tal advances making use of the experimen- whether similar causal effects can be found tal method in the physical and biologi- in studies that have different identification Charitable Fundraising cal sciences have been fast and furious.2 assumptions. Within economics, the use of controlled Field experiments can play an impor- The charitable marketplace represents experiments has steadily increased, fueled tant role in the discovery process by allow- an interesting set of actors, which might be by the exploration of important economic ing us to tackle questions that are quite usefully parsed into three distinct types.5 phenomena in the laboratory more than difficult to answer without use of ran- First, is the Government, which decides one half century ago. domization in a field setting. They also on tax treatment of contributions and the Although laboratory experiments can serve an important complementary level of grants to charities. This insight- have dominated the experimental land- role — similar to the spirit in which astron- ful literature includes studies that explore scape in economics, the past decade has omy draws on the insights from particle crowding out, and studies that measure witnessed a significant surge in studies physics and classical mechanics to make responsiveness of giving to price changes.6 that gather data via field experiments. In sharper insights, field experiments can sup- Second are the donors, who provide the plement insights gained from lab and nat- resources to produce public goods. The * List is a Research Associate in the NBER’s urally occurring data. To date, field experi- final set of actors is the charitable organi- Program on Public Economics and a Professor in Economics and the College and ments have shed insights on areas as diverse zations, which develop strategies to attract Co-Director of Graduate Admissions at the as tests of auction theory, tests of the theory resources to produce public goods. The University of Chicago. His profile appears of private provision of public goods, tests economic interplay of these three actor later in this issue. that pit neoclassical theory and prospect types represents a vibrant area of research.

14 NBER Reporter • 2008 Number 4 My own research has focused on and I used a natural field experiment, port the Center for Natural Hazards studying the relationship between indi- dividing 3000 direct mail solicitations Mitigation Research at East Carolina vidual donors and charities. In the United to Sierra Club supporters into four treat- University. Federal gambling laws pro- States alone, annual individual giving now ments and asking solicitees to support the hibited us from executing a mail solici- exceeds 2 percent of GDP, with roughly expansion of a K-12 environmental edu- tation, a phone-athon, or an email drive. 90 percent of people giving money to at cation program. We find that announce- We therefore turned to a door-to-door least one cause annually. Further, there is ment of seed money increases the par- fundraising drive, wherein we approached at least one capital campaign under way ticipation rate of potential donors by 23 nearly 5000 households, randomly in virtually every major population cen- percent and total dollar contributions assigned to standard voluntary contribu- ter in North America that has an objec- by 18 percent, compared to an identical tions mechanism treatments (VCM — in tive of raising between $25–$100 million. campaign in which no announcement of other words, merely knocking on the door Smaller capital campaigns are even more leadership gift is made. Other scholars and asking for a contribution) or lottery numerous.7 Increased individual wealth, also have shown the importance of seed treatments, where every dollar given pro- an aging population, and recent devolu- money.9 vides one chance at a lottery prize. tionary trends across governments world- Dean Karlan and I explore a differ- Some insights emerged from that wide combine to set the stage for contin- ent use of upfront monies: we extend this experiment. For example, the lottery treat- ued rapid growth in the sector. line of inquiry by soliciting contributions ments raised roughly 50 percent more in Even though the stakes are clearly using a conditional, rather than an uncon- gross proceeds than our VCM treatments. high, many economic facts concern- ditional, match.10 Soliciting contributions This result was largely driven by greater ing the interrelationship between solici- from more than 50,000 supporters of a participation rates in the lotteries: lotter- tors and solicitees remain unknown. My liberal organization, we randomize house- ies increase participation rates by roughly own interest in this area was shaped in holds into several different matching rate 100 percent. This finding highlights an the late 1990s, when I was an Assistant treatments to explore whether presence of attractive feature of lotteries: they pro- Professor at the University of Central a match, and the match rate, influence giv- vide fund-raisers with a tool to gener- Florida and my Dean asked me to help ing. We find that simply announcing that ate “warm lists,” or a larger pool of active him raise money for a proposed Center in a match is available considerably increases donors to draw from in future fund-rais- the Economics Department. He provided the revenue per solicitation — by 19 per- ing drives. Our experimental design also me with $5,000 as “seed money.” After cent. In addition, the offer of a match sig- permits an exploration of whether, and a significant amount of discussion, the nificantly increases the probability that an to what extent, individual solicitor char- administration agreed to allow me to run individual donates — by 22 percent. Yet, acteristics influence fund-raising success. a charitable fundraising field experiment while the match treatments relative to a In this regard, we find that a single-stan- with the $5,000. control group increase the probability of dard deviation increase in physical attrac- This first study, completed with donating, larger match ratios — for exam- tiveness of women solicitors increased the David Reiley,8 was an effort to test eco- ple a $3 match for every $1 donated, or $2 average gift by approximately 35 to 72 per- nomic theory, to learn how best to use the for every $1 donated — relative to smaller cent. This result highlights the power of seed funds, and to actually multiply the match ratios (such as one to one) have no social factors in effecting fundraising.13 $5,000 seed funds. We split the full capi- additional impact. Whether, and to what extent, these tal campaign into several smaller capital Other studies also have shed light factors influence long-run giving patterns campaigns to fund one of six computers on the value of using a match. For exam- is an open question. To move the explo- the Center needed, each of which served ple, Eckel and Grossman11 use lab experi- ration from one of measuring short-run as a separate experimental treatment. For ments to compare matching to an equiva- substitution effects to one that measures example, one household received a solici- lent rebate of one’s contributions in the dynamic effects, we use a second door-to- tation that noted we had already secured context of a dictator game; they find door study.14 In this study, we use detailed $1000 of the goal, and we were asking that matching contributions lead to sig- information on the households that were solicitees to make up the shortfall. We nificantly larger contributions than the previously approached and how they were solicited contributions from 3000 Central rebate mechanism. Rondeau and I also approached. A little more than one year Florida residents, randomly assigned to report evidence consonant with the posi- after the first drive, we randomly allocated six different groups of 500, with each tive effects of having a match available. previous givers (“warm” list agents) and group asked to fund a separate computer In a study that explores whether those who have never given (“cold”list for use at the Center. Most importantly, upfront money can be used more effec- agents) into one of three treatments: a we found that seed money increased the tively by purchasing lottery prizes for VCM, and two gift treatments — asking average gifts of donors: more seeds led to donors, several colleagues and I inves- solicitees for money, but giving a large more money. tigated the effects of using lotteries to or a small gift to the potential donors. In a similar spirit, Daniel Rondeau induce giving.12 Our efforts were to sup- We also randomized solicitor types across

NBER Reporter • 2008 Number 4 15 household — in some cases removing the use upfront monies efficiently to gener- North Holland (2006), pp. 1201–69. “beauty incentive” that was present in the ate the greatest level of gifts, and learn- 6 See the important work of Martin first drive, for others adding the “beauty ing about appropriate ask strategies for Feldstein and Charles T. Clotfelter. For incentive,” and for others keeping the the present and future are invaluable. In example, M. Feldstein, “The Income Tax “beauty incentive” the same. this regard, the data point to the fact that and Charitable Contributions: Part II- In that experiment, we find that the upfront money is important, and that The Impact on Religious, Educational warm-list enhances fund-raising success: it can be used effectively as announced and Other Organizations,” National donors who contributed in the first fund- seed money, a matching grant, to pur- Tax Journal 1975, 28(2): pp. 209–26; raising drive were twice as likely to give, chase donor gifts, or to purchase lottery C.T. Clotfelter, Federal Tax Policy and and provide average gifts that were roughly gifts. Critically, long-run fundraising suc- Charitable Giving, University of Chicago twice as large, as households that had cess depends on incentives used to attract Press, (1985), and more recently G. Auten, never given. We find, however, a differ- first-time donors. More specifically, some H. Sieg, and C.T. Clotfelter, “Charitable ence in the efficacy of how warm-list giv- types of incentives might crowd out rea- Giving, Income and Taxes: An Analysis ers were initially attracted: removing the sons for giving later. The literature also of Panel Data,” American Economic lottery incentive had no discernable pinpoints that non-price incentives can Review, 2002, 92: pp. 371–82. impact on the behavior of warm-list have important effects on givers; in some 7 See D. Rondeau and J.A. List, households, whereas removing the “beauty cases surprisingly large effects. For exam- “Matching and Challenge Gifts to incentive” leads to a significant reduction ple, giving small donor gifts to new solici- Charity: Evidence from Laboratory in average contribution levels. tees can substitute for a warm-list. and Natural Field Experiments, NBER Finally, concerning the effect of the I suspect that this line of research Working Paper No.13728, January 2008, gifts in this natural field experiment, we will continue to be a strong growth area. and “Special Issue on Field Experiments find that donor gifts can have important As fundraisers continue to recognize the in Charity,” Experimental Economics, effects on the intensive and extensive mar- value of experimentation, economists will 11(3), pp. 253–67. gins. Importantly, whether the gift is given increasingly be called upon to lend their 8 J.A. List and D. Lucking-Reiley, unconditionally (solicitee receives the gift services. Likewise, as economists continue “The Effects of Seed Money and Refunds regardless of whether she gives) or condi- to recognize the value of using naturally on Charitable Giving: Experimental tionally (solicitee receives the gift if she occurring settings as laboratories, such Evidence from a University Capital gives a pre-specified dollar amount) influ- domains increasingly will be used to gen- Campaign,” Journal of Political Economy ences the probability and size of the mon- erate new datasets. (2002), 110(1), pp. 215–33. etary contribution. 9 One unique recent example is S. Huck 1 S.D. Levitt and J.A. List, “Field and I. Rasul, “Comparing Charitable Lessons Learned Experiments in Economics: The Past, The Fundraising Schemes: Evidence from a Present, and The Future,” NBER Working Natural Field Experiment,” working paper, A first lesson that I take from this Paper No.14356, September 2008. UCL. body of research is that what we do not 2 For a more complete discussion see J.A. 10 D. Karlan and J.A. List. “Does Price know dwarfs what we know. So, what List and D. Reiley, “Field Experiments” Matter in Charitable Giving? Evidence have we learned from this body of work? in The New Palgrave Dictionary of from a Large-Scale Natural Field One feature of this line of research is Economics, S. N. Durlauf and L. Experiment,” NBER Working Paper that it has been able to lend insights into E. Blume, eds. Palgrave Macmillan No.12338, June 2006, and American the types of models that accurately pre- Publishing. Economic Review, (2007), 97(5), pp. dict giving behavior. In this sense, simple 3 See J.A. List, “Field Experiments: 1774–93. models that treat individual contributions A Bridge between Lab and Naturally 11 C.C. Eckel, and P.J. Grossman. as if they are identical to purchases of Occurring Data,” NBER Working Paper “Rebate versus Matching: Does How private goods should be reconsidered in No. 12992, March 2007, and Advances We Subsidize Charitable Contributions light of the findings from this literature. in Economic Analysis and Policy 6(2), Matter?” Journal of Public Economics, Additionally, work in this area that mea- Article 8. 87, 2003, pp. 681–701. See also B. Frey sures key parameters has provided econ- 4 See my website, www.fieldexperiments. and S. Meier, “ Social Comparisons and omists with useful information and lent com, which contains a catalogue of field Pro-social Behavior: Testing ‘Conditional guidance into policymaking discussions. experiments in these and several related Cooperation’ in a Field Experiment,” Equally as important, these studies have areas. American Economic Review 94(5), collected enough facts to help us con- 5 See the valuable work of J. Andreoni, 2005: pp.1717–22. struct new economic theories of giving. “Philanthropy,” In Handbook of Giving, 12 C.A. Landry, A. Lange, J.A. List, For practitioners, understanding Reciprocity and Altruism, S-C. Kolm M.K. Price, and N. Rupp, “Toward what motivates people to give, how to and J. Mercier Ythier, eds. Amsterdam: an Understanding of the Economics

16 NBER Reporter • 2008 Number 4 of Charity: Evidence from a Field interested reader to the clever work of Frey 14 C.A. Landry, A. Lange, J.A. List, Experiment,” NBER Working Paper No. and Meier (2005; cf. footnote 11) and M.K. Price, and N. Rupp, “Is a Donor 11611, September 2005, and Quarterly R. Croson and J. Shang, “The impact of in Hand Better than Two in the Journal of Economics, 121 (2): pp.747– downward social information on contri- Bush? Evidence from a Natural Field 82. bution decisions,” “Special Issue on Field Experiment,” NBER Working Paper No. 13 For more in the economics literature Experiments in Charity”, Experimental 14319, September 2008. on the effects of social factors, I direct the Economics, 11(3), pp. 221–33.

Prices, Productivity, and Innovation

David E. Weinstein and Christian Broda*

Economists have long known the economy. This research has produced a sured. Virtually all price indexes used importance of focusing on “real” as series of papers that have emphasized the by economists are essentially “common opposed to “nominal” variables in order macro implications of these micro biases. goods” price indexes. In other words, to understand a wide range of economic The principle macro implications of our most of these indexes compare the prices outcomes including growth, productiv- work are: of a common set of goods sampled in two ity, and welfare. While the distinction • Because trade provides consumers periods and then take a weighted average between real and nominal variables is with new goods, we have underestimated of those prices to obtain a single esti- simple in theory, in practice it is very the gains from globalization around the mate of inflation. In the case of the U.S. difficult for statistical agencies to mea- world over the last few decades.1 Consumer Price Index (CPI), we adjust a sure prices accurately. One of the main • We estimate the aggregate CPI small subsample of the prices for quality problems is that the set of goods in the bias for a large set of goods to be close to changes (for example, computers), but in economy is constantly changing because 1 percentage point per year and to have general no adjustments are made for the of the creation of new goods and quality a strong pro-cyclical component. The appearance of new goods. upgrading. How can we measure price cyclicality of the bias suggests that busi- The problem with this methodol- changes when the set of goods consumed ness cycles are more pronounced than is ogy is that the appearance of new goods in two periods is different? Much of typically reported in official statistics.2 has price implications for consumers. our research over the last few years has • Incorporating the effect of new To understand this, one needs to think focused on estimating the impact of new goods into the measurement of prices about how a new good affects a con- goods on our understanding of the U.S. suggests that real wages for the typical sumer. As John Hicks argued decades and world economies. worker in the United States have risen ago, the appearance of a new good can A hallmark of our approach has substantially over the last 30 years. It also be thought of as a drop in the price of the been to combine micro data with a rich suggests that poverty rates based on our good from its reservation price — that framework that allows the biases in the corrected CPI measurements have fallen is, the price at which demand equals price measurement of individual goods sharply since the late 1960s relative to zero — to the observed market price. to be aggregated over large sectors of the their official counterparts.3 Since official price indexes do not record these implied price drops, they overstate * Broda is an NBER Faculty Research New Goods and Inflation inflation. Fellow and a professor at the University of Chicago’s Graduate School of Business. Although this problem with conven- Weinstein is an NBER Research Associate The starting point for thinking about tional indexes is well known, prior work and a professor of economics at Columbia how new goods and higher quality goods has only been able to address it for a hand- University. Their profiles appear later in affect price measurement is an under- ful of products. The Boskin Commission, this issue. standing of how prices are currently mea- for instance, extrapolated the findings of

NBER Reporter • 2008 Number 4 17 a few studies to estimate the bias for the remains a substantial bias arising from siderable heterogeneity, even in goods entire CPI, but this extrapolation was new and higher quality goods in the categories that sound homogeneous, like based on studies that covered at most CPI. This upward bias averages 0.6– bottled water: Perrier can sell at an enor- 10 percent of the CPI. Moreover, these 0.9 percent per year depending on the mous premium over Poland Springs. studies were not comprised of a represen- aggregation methodology. The bias is Again, these differences are not random. tative sample of goods. One of the main also strongly pro-cyclical, which sug- In particular, households with higher differences between our work and that gests that business cycles are more pro- incomes tend to buy more expensive of prior researchers is that we examine a nounced than is typically reported in varieties of the same class of good. For unique dataset that covers the universe of official statistics. example, a household that spends twice products with barcodes — approximately as much as another household per capita 700,000 goods in a typical quarter, cov- International Price tends to pay 6 percent more for the items ering around 40 percent of all expendi- Implications of Product in a product group. Approximately 85 tures of goods in the CPI. Because man- Heterogeneity percent of this price difference is attrib- ufacturers almost never change a barcode utable to richer households purchasing unless they make some modification, and Barcode data also can help us to more expensive varieties within a prod- never make an important modification understand many of the puzzles in inter- uct group. These results hold even for to a good without changing the barcode, national economics.4 Consider two key seemingly homogeneous goods like eggs, our data enable us to observe virtually all results about international price devia- milk, and sodas, which establishes that changes in quality, or in the set of goods, tions: borders give rise to flagrant vio- tests of price convergence using aggre- in this sample. lations of the law of one price, and con- gate data collected across countries or The data reveal several important vergence rates back to purchasing power regions with different per capita incomes facts for understanding pricing. First, parity (PPP) are inconsistent with the are likely to falsely reject purchasing over 90 percent of product creation and evidence of micro studies on nominal price parity (PPP) because the quality of destruction happens within firm product price stickiness. A major problem with the goods will vary systematically with lines and because of firm entry and exit. these studies is that they compare goods income. To put this in perspective, we find four that are not identical. For example, if Taken together, these facts suggest times more entry and exit in product we look at product categories similar to that a major problem in examining PPP markets than is found in establishment those in the CPI — like fresh eggs and across borders is that the set of goods and labor market data. In a typical year, milk — we see that the disaggregate price consumed in different countries var- 40 percent of a household’s expenditures indexes across cities within the United ies substantially — even seemingly sub- are on goods that were created in the last States are comprised of samples of goods stitutable goods, like eggs, sell at very four years, and 20 percent of expendi- with little or no overlap across locations. different prices, even within the same tures are in goods that disappear in the This fact implies that using these price store.5 Thus, finding that egg prices do next four years. This implies that price indexes as the basis of studies of the law not move together across borders may indexes that do not adjust for the impor- of one price has clear limitations. be in part because consumers in differ- tant role played by new goods are likely Moreover, since Canada and the ent locations purchase similar goods that to be highly susceptible to new goods or United States use the same barcode sys- differ substantially in quality. Moreover, quality biases. tem, we can directly compare the similar- since there is substantial heterogeneity in Second, we find that net creation is ity of consumption bundles in the two the prices charged for the same barcode strongly pro-cyclical, with more prod- countries. In the typical bilateral city/ in different locations, even within the ucts being introduced in expansions and region comparison between the United United States, the simple fact that inter- in product categories that are booming. States and Canada, only 7.5 percent of national prices differ is not informative Destruction of goods is counter-cyclical, the goods are common. This is less than of the incremental border barrier. although its magnitude is quantitatively one third the common set of goods avail- Indeed, we find that if one restricts less important. This is suggestive of mod- able between city pairs of equal distance the analysis to the set of common goods els where firms have an incentive to defer within the United States. In other words, consumed in the United States and implementation of the product until a random sample of goods sold in the Canada, the border introduces only a aggregate demand is relatively high. United States is likely to differ substan- small distortion to relative prices, and Finally, we develop a methodology tially from the composition of a sample rates of convergence to PPP are relatively that enables us to estimate the aggregate of goods sold in Canada. By the same fast. However, if one runs the same anal- importance of price drops for consum- token, we find that more proximate loca- ysis on product groups — for example, ers. We show that since most product tions have more similar consumption milk — the border looms much larger creation and destruction is unobserved bundles than distant locations. because the prices of different goods do by the Bureau of Labor Statistics, there The next key fact is that there is con- not always move together.

18 NBER Reporter • 2008 Number 4 New Goods, Variety, the level of long-run growth is large. Our market power. The results are robust to and Growth results indicate that the increase in vari- the inclusion of political economy vari- eties we observed is likely to raise the ables and to a variety of model specifica- We next turn our attention to the present discounted value of the income tions. The effect is statistically and eco- implications of the increases in the set by 17 percent. Of this, only 1.3 percent- nomically significant, both relative to of available goods for economic growth. age points are attributable to the static other explanations and to the average Earlier work had demonstrated an gains from trade; the remainder is due tariff in the typical country. In short, we important fact for the United States. to the impact that new goods have on find strong evidence that countries have If we defined a “variety” as a narrowly the incentive to perform R and D and market power in imports and exploit it defined good exported by a particular to invest. In other words, semi-endog- in setting their trade policy. country, for example French red wine, enous growth models suggest that there we saw that there was a dramatic increase are very powerful growth effects caused 1 C. Broda and D. E. Weinstein, in the number of goods imported by the by trade liberalization that are ignored “Globalization and the Gains from United States between 1972 and 2001. by conventional static analyses. Variety”, Quarterly Journal of This analysis indicated that there were Economics, Vol. 121, Issue 2, May 2006. substantial gains to the U.S. economy Political Economy Issues 2 C. Broda and D. E. Weinstein, arising from the increased availability of “Product Creation and Destruction: foreign varieties in our markets. The previous work has focused on Evidence and Price Implications,” NBER In 2006, we reexamined this from the positive gains from new varieties, but Working Paper No. 13041, April 2007, a global perspective.6 A key feature of it is legitimate to ask, “if new varieties are and forthcoming in The American endogenous growth models is that the so good, why do countries restrict trade?” Economic Review. introduction of new varieties drives pro- We have focused on two key theories of 3 C. Broda and D. E. Weinstein, ductivity growth. Analyzing 6-digit bilat- trade barriers.7 First is the standard, but Prices, Poverty, and Inequality, 2008, eral trade flows over the period 1994– controversial, “optimal tariffs argument,” Washington: AEI Press. 2003, we document that the reason that that is, countries set tariffs to exploit 4 C. Broda and D. E. Weinstein, trade-to-GDP ratios have been rising in their market power in international mar- “Understanding International Price virtually all countries is that countries kets. Second, we examine more conven- Differences Using Barcode Data,” NBER are importing new varieties, not because tional political economy arguments. Working Paper No. 14017, May 2008. they imported more of existing varieties. In doing so, we make three contri- 5 C. Broda and D. E. Weinstein, In the typical developing country, virtu- butions. First, we estimate elasticities “Getting Real about Real Wages,” ally all of the growth in imports to GDP of export supply faced by 15 importer Columbia University mimeo, 2008. came from the import of new varieties. countries at a highly disaggregated 6 C. Broda, J. Greenfield, and D. These models also imply that new level. Second, we use these elasticities E. Weinstein, “From Groundnuts to varieties should affect the growth rate to provide evidence that, prior to con- Globalization: A Structural Estimate of productivity. The wider access to straints imposed by the World Trade of Trade and Growth,” NBER Working imported intermediate goods means that Organization (WTO), these countries Paper No. 12512, September 2006. R&D labor is more productive, which systematically set higher import tariffs on 7 C. Broda, D. E. Weinstein, and Nuno reduces the cost of generating new blue- goods in which they have market power. Limao, “Optimal Tariffs and Market prints for intermediate products. We Finally, we estimate similar elasticities for Power: The Evidence,” NBER Working find that the effect of varieties on the the United States and find that its trade Paper No. 12033, February 2006, and growth rate of an economy is relatively restrictions that are not constrained by forthcoming in The American Economic small, temporary, but persistent. Given the WTO are significantly higher in Review. this persistence, however, the impact on goods where the United States has more

NBER Reporter • 2008 Number 4 19 NBER Profile: Nick Bloom

Nick Bloom is a Research Associate (the U.K. Finance Ministry) from 2001– in the NBER Programs on Economic 2, McKinsey & Company (the manage- Fluctuations and Growth, Monetary ment consultants) from 2002–3, and at Economics, and Productivity. He is also the Centre for Economic Perfomance from an Assistant Professor in the Economics 2003–6. His work focuses on firms and Department at Stanford University and productivity, with three particular areas an Associate of the Centre for Economic of interest: the effects of uncertainty on Performance at the London School of economic activity; the causes and conse- Economics. quences of good management; and innova- Bloom received a B.A. from tion and technical change. Cambridge University in 1994, an M.Phil. Nick lives with his wife, Claire, in from Oxford University in 1996, and a Palo Alto, along with their two children, Ph.D. from University College London in Amelia (5) and Alex (3). He enjoys spend- 2001. He worked at the Institute for Fiscal ing his time trying to impress friends and Studies (a public-economics think-tank in family with his excellent British cooking London) from 1996–2001, HM Treasury and wine cellar of classic British vintages.

NBER Profile: Christian Broda

Christian Broda is an NBER Faculty financial contracts. He has written for Research Fellow and a Professor of publications ranging from the American Economics at the University of Chicago’s Economic Review and Quarterly Journal Graduate School of Business. He earned of Economics to the Financial Times and a bachelor’s degree in economics in 1997 New York Times. He also serves as associ- from Universidad de San Andres in ate editor of the Journal of Development Argentina. Four years later, he received his Economics and as a panel member of Latin PhD from MIT. Prior to joining Chicago’s America’s Economia journal. Business School faculty in 2005, Broda When he’s not doing research or worked in the research department of the teaching, Broda enjoys horse racing, cook- Federal Reserve Bank of New York. ing, and playing soccer, especially when Broda studies issues related to inter- followed by an Argentine-style asado. national finance and trade, and the impact of exchange rates on asset prices and

20 NBER Reporter • 2008 Number 4 NBER Profile: Charles W. Calomiris

Charles W. Calomiris is a Research and Ph.D. levels at Columbia. His research Associate in the NBER’s Program on the spans several areas, including banking, cor- Development of the American Economy porate finance, financial history, and mon- and the Henry Kaufman Professor of etary economics. Financial Institutions at the Columbia In 1999 and 2000, Calomiris served University’s Graduate School of Business. on the International Financial Institution He is also a Professor at Columbia’s School Advisory Commission, a Congressional of International and Public Affairs. commission to advise the U.S. government In 1995, Calomiris was named a on the reform of the IMF, the World Bank, University Scholar at the University of the regional development banks, and the Illinois, where he served as Associate WTO. He is now a member of the Shadow Professor of Finance and Co-Director Financial Regulatory Committee and the of the Office for Banking Research. He Financial Economists Roundtable, and was also has been a visiting faculty member at a Senior Fellow at the Council on Foreign Stanford University’s economics depart- Relations. He also has served as a con- research grants or awards from the National ment and at the finance department of sultant or visiting scholar for the Federal Science Foundation, the World Bank, the the Wharton School of Finance. Reserve Banks of New York, Chicago, Japanese Government, and others. Calomiris received a B.A. in eco- Cleveland, St. Louis, and Philadelphia, the Calomiris and his wife, Nancy Wolf nomics from Yale University in 1979 Federal Reserve Board, the World Bank, Calomiris, have two daughters: Zoi and a Ph.D. in economics from Stanford and the governments of Mexico, Argentina, Nicoletta, 18, and Eleni Sophia, 17. In his University in 1985. Currently, he teaches Japan, China, Colombia, and El Salvador, limited free time, Calomiris performs mod- courses on corporate finance, emerging the Attorneys General of Connecticut and ern Greek music and ancient Byzantine financial markets, corporate governance, Massachusetts, and numerous private sec- chants. and international banking at the MBA tor clients. Calomiris is the recipient of

NBER Profile: John A. List

John A. List is a Research Associate in List’s current research emphasizes the the NBER’s Program on Public Economics use of experimental methods to address both and a Professor in Economics and the College positive and normative issues. Much of his and Co-Director of Graduate Admissions time has been spent using field experiments at the University of Chicago. He received to explore economic aspects of environmen- a B.A. in Economics from the University tal regulations, incentives, preferences, values, of Wisconsin-Stevens Point in 1993 and a and institutions. Ph.D. in Economics from the University of List is Associate Editor of the American Wyoming in 1996. He has also taught at the Economic Review and the Journal of Economic University of Central Florida, the University Literature. He also serves on several journal of Arizona, and the University of Maryland. Editorial Boards, is a co-editor of the Journal From May 2002 to July 2003, List served of Environmental Economics and Management, as Senior Economist for Environmental and and is on the EPA Science Advisory Board. Resource Economics at the President’s Council List is married and has five children: of Economic Advisers in Washington. His pri- Annika, twins Eli and Noah, Greta, and mary work focused on multinational mar- Mason. In his spare time he enjoys taking his ket institutions to address climate change, kids to sports card shows, where he also exe- the Clear Skies Act, revising the Office of cutes field experiments, and traveling with his Management and Budget benefit cost guide- family. His wife, Jennifer List, recently started lines, and the softwood lumber trade dispute Seven Smooches, a company that designs and between the United States and Canada. makes children’s clothing.

NBER Reporter • 2008 Number 4 21 NBER Profile: John Van Reenen

John Van Reenen has been a Research Van Reenen was a senior advisor to the UK Associate in the NBER’s Program on Secretary of State for Health and Prime Labor Studies since 2006. He has also Minister. been Director of the Centre for Economic Van Reenen’s work focuses on the Performance and a professor in the causes and consequences of innovation Department of Economics at the London on economic performance, with a par- School of Economics (LSE) since 2003. He ticular emphasis on growth and the labor is currently the Denning Visiting Professor market. His research covers wage inequal- at the Stanford University Graduate School ity, econometrics, many areas of public of Business. policy, and the industrial organization of Van Reenen received his B.A. from high tech industries. Recently, he has been Cambridge University (UK) in 1988, his working on the economics of organization M.Sc. from the LSE, and his Ph.D. from and management. University College London in 1993. From Van Reenen, his wife Sarah Chambers 1998–9, he was a Visiting Professor of (an interior designer), and their daughter Economics at the University of California, Charlotte (1) live in San Francisco. Apart Berkeley. From 1994–2003, he was from “messing around with Charlotte”, his a Professor of Economics at University favorite pastimes are politics, music, poetry, College London. Between 1999 and 2002, and reading novels and comic books.

NBER Profile: David E. Weinstein

David E. Weinstein is an NBER Research Harvard University. He also served on the Associate, Director of the NBER’s Japan Council of Economic Advisors from 1989 Project, and the Carl S. Shoup Professor of the to 1990. Japanese Economy at Columbia University. Weinstein’s teaching and research inter- He is also the Associate Director of Research ests include international economics, macro- at the Center for Japanese Economy and economics, corporate finance, the Japanese Business at Columbia, and a member of the economy, and industrial policy. He earned Council on Foreign Relations. his Ph.D. and M.A. in Economics from the Previously, Weinstein was a Senior University of Michigan and his B.A. at Yale Economist at the Federal Reserve Bank of University. He is the recipient of numerous New York and a consultant for the Federal grants and awards, including four National Reserve Bank of San Francisco and the Science Foundation grants, an Abe Fellowship, Federal Reserve Board of Governors. Prior and a Japan Foundation Fellowship. to joining the Columbia faculty, he was the Weinstein lives with his family in Sanford R. Robertson Associate Professor of Manhattan. When not traveling to Asia, they Business Administration at the University of enjoy bicycling, tennis, theater, fine dining, Michigan’s School of Business Administration and the “gains from variety that New York and an Associate Professor of economics at has to offer.”

22 NBER Reporter • 2008 Number 4 Conferences

NBER’s 23rd Tax Policy and the Economy Conference Held in Washington

The NBER’s 23rd Conference on Tax the Labor Supply Effects of the Social Knittel, U.S. Department of Treasury, Policy and the Economy took place at the Security Earnings-Test” “Understanding U.S. Corporate Tax National Press Club in Washington on Losses” September 25. NBER Research Associate Jeffrey R. Brown and Don Fullerton, Jeffrey R. Brown of the University of University of Illinois and NBER, and James R. Hines, Jr., University of Illinois at Urbana-Champaign and Julia L. Coronado, Barclays Global Michigan and NBER, and Lawrence NBER President James M. Poterba of Investors, “Is Social Security Part of the H. Summers, Harvard University and MIT organized this year’s meeting. The Social Safety Net?” NBER, “How Globalization Affects Tax following papers were discussed: Design” Rosanne Altshuler, Rutgers University; Leora Friedberg, University of Virginia Alan J. Auerbach, University of A. Abigail Payne, McMaster University, and NBER, and Anthony Webb, California, Berkeley and NBER; “Empirical Analysis of Crowd-Out” Boston College, “New Evidence on and Michael Cooper and Matthew

Friedberg and Webb investigate the gains persisted as exposed cohorts aged the lifetime-income distribution. Third, impact on labor supply of changes in and were also observed at younger ages in cases where redistribution does occur, the Social Security earnings test in 1996 because of the shock to anticipated earn- it is not targeted efficiently: many high- and 2000. They highlight how inertia in ings-test rules. income households receive positive net labor supply choices can influence the Expanding on earlier work, Brown transfers, while many low-income house- responses to policy changes in two ways: and his co-authors develop a large sam- holds pay net taxes. Finally, the progres- First, they show that taking account of ple of individuals born at different times, sivity of Social Security changes over time previous employment status is important construct an entire lifetime earnings his- because of changing labor force patterns in estimating responses to any current tory for each individual, and then develop and other demographic patterns, but the earnings-test changes. Second, they test several measures by which to classify direction of these changes depends on the the effect of both actual and anticipated each individual’s lifetime economic sta- income concept used to classify someone earnings-test parameters that cohorts tus. Then, they calculate each individual’s as “poor.” Specifically, when comparing faced at earlier ages. This approach dem- Social Security taxes and retirement bene- individuals in pre-baby boomer cohorts onstrates that both past and anticipated fits, and use several measures of the impact to baby boomers, the authors find that future rules can influence current employ- of the Social Security retirement program Social Security is becoming slightly less ment and earnings. Thus, Friedberg and on “the poor.” Their four major find- redistributive when measured by indi- Webb identify an impact of earnings-test ings are: first, as the definition of income vidual income and slightly more redis- changes on employment not only contem- becomes more comprehensive, Social tributive when measured by household poraneously for those at the ages directly Security’s retirement program becomes, income. affected, but also for those at younger overall, less progressive. Indeed, by using Recent data on corporate tax losses ages and in the years that follow the direct “potential” labor earnings (defined as an present a puzzle: the ratio of losses to pos- change. Finally, the researchers show that individual’s labor endowment multiplied itive income was much higher around the earnings-test changes initiated in 1996, by the individual’s wage rate) at the house- recession of 2001 than in earlier reces- like the changes in 2000 that have been hold level rather than actual earnings at sions, even those of greater severity. Using studied by others, affected labor supply. the individual level, the researchers find a comprehensive sample of U.S. corporate Overall, Friedberg and Webb predict that that Social Security has virtually no effect tax returns for the period 1982–2005, the elimination of the earnings-test in on overall inequality. Second, this result Altshuler and her co-authors explore 2000 raised employment among Health is driven largely by the lack of redistribu- a variety of potential explanations for and Retirement Study respondents by tion across the middle and upper part of this surge in tax losses, taking account around 2 percentage points at ages 66 the income distribution, so it masks some of the significant use of stock options in to 69 and 3.5 points at age 65. These positive net transfers to the bottom of compensation packages beginning in the

NBER Reporter • 2008 Number 4 23 1990s and of recent temporary tax pro- neously making it more difficult to raise should they be concerned about whether visions that might have had important revenue because of the greater interna- their efforts are crowded out by changes effects on taxable income. They find that tional mobility of economic activities. In in behavior on the part of individuals and losses rose because the average rate of essence, the greater mobility of factors of institutions participating in the provision return of C corporations fell, rather than production means that even large coun- of the good or service? The theoretical because of an increase in the dispersion of tries like the United States are becoming literature suggests that crowd-out is an returns, or an increase in the gap between more like small open economies in terms issue. The empirical literature, however, corporate profits subject to tax and cor- of the responsiveness of the tax base to has largely failed to find a measurable porate profits as measured by the national changes in tax rates. The researchers doc- crowd-out effect. With better data and income accounts. This analysis also sug- ument that smaller countries tend to rely more sophisticated empirical techniques, gests that the increasing importance of more heavily on taxes on goods, services, a burgeoning literature now shows that S corporations may help to explain the and international trade than do larger crowd-out does exist. Payne examines recent experience within the C corporate countries, noting in particular the grow- the recent literature on crowd-out across sector, because S corporations have exhib- ing popularity of value-added taxes on a variety of venues to better understand ited a different pattern of losses in recent the international scene. Their work also the empirical estimation issues, as well as years. However, the researchers can iden- highlights the fact that the United States the institutional details that can lead to tify no simple explanation for the differing taxes both personal and corporate income an enhanced knowledge of the effects of experience of C and S corporations. Their at high rates relative to other countries, government programs on individuals and investigation raises new puzzles about and that the efficiency costs of this type of organizations. why rates of return of C corporations taxation will continue to grow relative to The proceedings of this conference declined so much early in the decade, and other forms of taxation as the world econ- are being reviewed for publication by the why the incidence of losses among C and omy becomes increasingly integrated. University of Chicago Press. The publi- S corporations has diverged. When governments introduce pro- cation’s availability will be announced in Hines and Summers suggest that grams, or funding for initiatives that are a future issue of the NBER Reporter. The globalization increases citizens’ demand partially provided by lower levels of gov- papers will also be available at “Books in for government services while simulta- ernments or in the private or third sectors, Progress” on the NBER’s web page.

The Great Inflation Conference Maintaining an environment of low The Great Inflation is considered the New Zealand, and Jacob Frenkel of the and stable inflation is widely regarded as most important macroeconomic policy Bank of Israel provided detailed accounts one of the most important objectives of failure in the United States since World of their experiences with combating infla- economic policy in general, and the single War II, and understanding its roots is tion in their respective countries. The dis- most important objective for monetary important to macroeconomists. To that cussion was focused most prominently policy in particular. An environment of end, an NBER conference on “The Great on inflation targeting and on the political price stability reduces uncertainty, Inflation” took place at the Woodstock battles involved in enacting inflation tar- improves the transparency of the price Inn on September 25–7. It was orga- geting in each of the countries. mechanism, and facilitates better plan- nized by NBER Research Associate The academic sessions that followed ning and the efficient allocation of Michael Bordo of Rutgers University and began with revisiting the ideas of Milton resources, thereby raising productivity. Athanasios Orphanides of the Central Friedman. The first paper, “The Great In the period from 1965 to 1982, Bank of Cyprus, and brought together Inflation: Did The Shadow Know Better?” known as the Great Inflation, inflation scholars and policymakers for this impor- by William Poole, Robert Rasche, and rose from near zero to above 15 percent tant discussion. David Wheelock of the Federal Reserve and caused significant damage to the U.S. The conference began with a panel Bank of St. Louis, focused on the Shadow economy, including distortions in finan- session, “A View from the Trenches,” in Open Market Committee. Founded in cial markets, a slumping stock market, which three former central bank gover- 1973, the shadow committee was based declining productivity, and social unrest. nors, on whose watch the Great Inflation on monetarist precepts and was formed Since the mid-1980s, monetary policy has was vanquished, reflected on their experi- in order to critically evaluate the Federal largely succeeded in controlling inflation ences. John Crow of the Bank of Canada, Open Market Committee’s actions and in the U.S. and other advanced countries. Donald Brash of the Reserve Bank of to formulate alternate policies to reduce

24 NBER Reporter • 2008 Number 4 inflation. These authors find, using -mod ployment as low as possible. the Evolution of Inflation Expectations,” ern techniques, that indeed “the Shadow The conference then switched topics conference organizer Athanasios knew better.” There followed a discus- to the role of a nominal anchor, and par- Orphanides and co-author John sion of the experience of Germany in the ticularly the role of inflation expectations Williams of the Federal Reserve Bank 1970s, and how policy there maintained in the formation of monetary policy. The of San Francisco noted that the policy lower inflation than most other advanced papers presented included “The Great of the 1960s and 1970s relied heavily on countries, and was able to avoid many of Inflation Drift” by Marvin Goodfriend, estimates of economic variables that often the mistakes that were made in the United NBER and Carnegie-Mellon University, were misguided and not measured cor- States. Andreas Beyer and Otmar Issing and Robert King, NBER and Boston rectly. This led to prescriptions for policy of the European Central Bank, Vitor University, and “Falling Behind the that also were misguided, and may have Gaspar of the Bureau of European Policy Curve: A Positive Analysis of Stop-Start contributed to the inflation that ensued. Advisers, and Christina Gerberding of Monetary Policies and the Great Inflation” Had policymakers been learning from the Deutsche Bundesbank, presented by Andrew Levin, Federal Reserve Board, or even noticing the mistakes they were their paper on the subject: “Opting Out and John Taylor, NBER and Stanford making, it might have been possible for of the Great Inflation: German Monetary University. If inflation expectations are them to readjust their models and to Policy after the Break Down of Bretton not anchored, the inflation itself will not reform policy in a more optimal way. Woods.” remain stable. It was argued that the The conference ended with a panel The discussion then turned to non- increase in expectations of inflation was a discussion involving Federal Reserve monetary explanations of the Great fundamental reason that inflation rose so Vice Chairman Donald Kohn, Harold Inflation. The first came from Alan dramatically, often because of the misper- James of Princeton University, and Anna Blinder, NBER and Princeton University, ceptions of policy by the Fed and a lack of J. Schwartz of NBER. Kohn empha- and Jeremy Rudd, Federal Reserve Board, coordination and communication. sized some lessons that central banks need in “The Supply-Shock Explanation of the International perspectives of the time to learn after experiences like the Great Great Stagflation Revisited.” The authors surrounding the Great Inflation in the Inflation. For example, central banks are focused on the commodity price shocks United States are important, because the the policy actors with key responsibility of the 1970s and the price-wage con- United States was not the only country for price stability; inflation expectations trols enacted by the government. The experiencing high inflation at the time. are critical; both vigorous debate and second paper “Stepping on a Rake: The In their paper, “The Great Inflation in the alternative viewpoints are essential for Role of Fiscal Policy in the Inflation of United States and the United Kingdom: creating better policy; shortcuts to price the 1970s” by Christopher Sims, NBER Reconciling Policy Decisions and Data stability are a recipe for disaster; and, cen- and Princeton University, regarded fiscal Outcomes,” Riccardo DeCecio and tral bankers must always exercise humil- policy as too accommodative at that time, Edward Nelson, both of the Federal ity when forming policy, since much of it aiding in the rapid increase in price pres- Reserve Bank of St. Louis, showed how depends on variables that no one can esti- sures. The consensus of both papers was the United Kingdom suffered with infla- mate. James highlighted four questions we that, had policy been more combative tion because of poor ideas and economic still have yet to answer: What is the dating against the forces at the time, the inflation theory underlying their monetary policy of the Great Inflation? What caused the might not have been avoided but poten- decisions. In a second paper — “Bretton Great Inflation? How can inflations be tially could have been far less severe. Woods and the Great Inflation” by ended? Why do we care about inflations? Next on the agenda was the issue of conference organizer Michael Bordo Finally, Schwartz returned to the lessons time consistency and central bank inde- and co-author Barry Eichengreen of of her colleague Milton Friedman. One of pendence. Alex Cukierman, Tel Aviv NBER and the University of California, the most important lessons of the Great University, presented “Misperceptions Berkeley — the authors posited that, Inflation was that the Fed incorrectly esti- about the Frequency of Price Adjustments before 1965, concern over external bal- mated the costs of disinflation for the sta- and Asymmetric Fed’s Preferences — An ance constrained Fed policy and anchored bilization of the real economy, and this Assessment of their Impact on Inflation expectations of inflation. After 1965, forced it to delay deflationary policy to and Monetary Policy Under Burns and when the Treasury increasingly handled the point where the costs to deflate only Miller.” Without a doubt, the Federal policy towards defense of the dollar, the became higher. Reserve Board of Governors of the 1970s, Fed paid more attention to the domestic These papers will be considered for under Chairman Arthur Burns, was influ- goal of full employment. This weakened publication in an NBER Conference enced heavily by political pressure; the the Bretton Woods nominal anchor. Volume by the University of Chicago ability to use its independence to fight The final session revolved around Press and will be posted at “Books in inflation was masked by the unfounded learning, expectations, and policy mis- Progress” on the NBER’s website. need to focus solely on keeping unem- takes. In “Monetary Policy Mistakes and

NBER Reporter • 2008 Number 4 25 American Universities in a Global Market An NBER conference on “American citations to decisions issued by our own E. Han Kim and Min Zhu, University Universities in a Global Market,” orga- U.S. Supreme Court. of Michigan, “Universities as Firms: The nized by NBER Research Associate This conference examined aspects of Case of U.S. Overseas Programs” Charles T. Clotfelter of Duke University, American higher education today that Discussant: Peter Doeringer, Boston was held on October 2–4. Since World will affect its global standing, in order to University War II, American universities have occu- provide insight about the likely future of pied an unchallenged position of global American universities. Will American uni- Devesh Kapur, University of preeminence. Owing to vigorous support versities sustain their leading role? Surely Pennsylvania, “Indian Higher Education” of scientific research by the federal gov- first-mover advantages and agglomera- Discussant: Arvind Panagariya, Columbia ernment, high rates of educational attain- tion economies, not to mention the tradi- University ment, and a massive influx of scholars tions of free expression and decentralized from Europe seeking refuge, America sup- competition, will continue to redound to Ofer Malamud, University of Chicago planted Europe as the home of the world’s our advantage. But the outcome seems far and NBER, “The Structure of Higher leading universities. Today, American from clear. Education in the U.S. and Europe: institutions dominate the highest rungs of The following papers were presented Reflections on the Nature of the Bologna the various world rankings of great univer- and discussed: Reforms” sities, routinely occupying the majority of Discussant: Michael Rothschild, the every top-twenty list. James D. Adams, Rensselaer Polytechnic Princeton University and NBER But this position of preeminence may Institute and NBER, “Is the U.S. Losing now be in jeopardy. The flow of foreign its Preeminence in Higher Education?” Lex Borghans and Frank Corvers, graduate students and scholars, while still Discussant: Charles Phelps, University of Maastricht University, “Mobility and massive, shows signs of slowing, in the Rochester the Changing Structure of Research and wake of heightened security concerns and Higher Education in Europe” competition from foreign universities. John Bound, University of Michigan and Discussant: Hugo Sonnenschein, Not only are European universities gird- NBER, and Sarah Turner, University University of Chicago ing themselves for more vigorous interna- of Virginia and NBER, “Coming to tional competition, but those in China, America: Where Do International Sunwoong Kim, University of India, and other parts of Asia have sig- Doctorate Students Study and How Do Wisconsin, Milwaukee, “Brain Drain, naled their intention to become major U.S. Universities Respond?” Brain Gain, and Brain Competition: players in the global higher education Discussant: Michael Teitelbaum, Sloan Changing Employment Opportunities market. Foundation and Career Patterns of U.S.-Trained Meanwhile, America’s own produc- Korean Ph.Ds” tion of university graduates has slowed rel- Paula Stephan, Georgia State University Discussant: Elizabeth U. Cascio, ative to that of other developed nations. and NBER, and Grant Black, Indiana Dartmouth College and NBER Whereas in 1990 the rate of college com- University, South Bend, “The Economics pletion in the United States (30 percent) of University Lab: Science and the Richard B. Freeman, Harvard University was 11 percentage points higher than the Role of Foreign Graduate Students and and NBER, “What Does Global median of a group of other developed Postdoctoral Scholars” Expansion of Higher Education Mean for countries; by 2004 the U.S. rate was 7 per- Discussant: Harold Shapiro, Princeton the U.S.?” centage points lower. And, there are University Discussant: Bruce Sacerdote, Dartmouth unmistakable signs that America’s position College and NBER of leadership in the world — financial, Haizheng Li, Georgia Institute of military, intellectual, and moral — is Technology, “Higher Education in Eric Bettinger, Case Western Reserve increasingly being challenged. One per- China — Complement or Competition University and NBER, “To Be or Not to haps small but telling example is the to U.S. Universities?” Be: Major Choices in Budding Scientists” recently reported decline among supreme Discussant: Debra Stewart, Council of Discussant: Caroline M. Hoxby, Stanford courts in other countries in number of Graduate Schools University and NBER

Adams asks whether the United higher education. While the growth of Asia since the 1980s has exceeded that States is losing its predominance in scientific research in Europe and East of the United States — suggesting con-

26 NBER Reporter • 2008 Number 4 vergence in world science and engineer- a U.S. university. They conclude that economy. Wage premiums are higher; ing — the slowdown of U.S. publica- international graduate students and few faculty members are being trained tion rates in the late 1990s is a different postdocs are not only important in (with implications for higher education matter. Using a panel of U.S. universi- staffing university labs, but also play a in the future); and disciplines outside ties, fields, and years, Adams uncovers leading role in university research. the professions (especially in the liberal evidence of allocative inefficiency in In his paper, Haizheng Li exam- arts) are collapsing. Yet professional public universities and university fields ines the Chinese university system as training, workforce training, and other in the middle and bottom 40 percent a possible competitor to the United highly specialized forms of education of their disciplines. Private/Top 10 uni- States. Enrollment in Chinese universi- are thriving. Does India face a future versities and top-20-percent university ties has increased rapidly since 1978, as with a workforce that’s reasonably well fields, however, do not display this pat- has the enrollment of Chinese students trained but narrow in its outlook and tern. Slow growth in tuition and state abroad. Will Chinese universities con- possibly less liberal? appropriations, compared to revenue tinue to serve as a complement to U.S. Europe is adopting some of the growth for private universities, are pos- universities by sending their best stu- unique aspects of the American sys- sible causes of this divergence. dents to the United States, or will they tem of higher education, including a It is well known that the represen- gradually become competitors to U.S. standardized degree structure and a tation of foreign doctorate students at universities? Though China has made system of transferable academic cred- U.S. universities, particularly in science great progress in improving its univer- its. Malamud reflects on how the char- and engineering fields, has increased sity system (China is now ranked fifth acteristics of American higher edu- dramatically over the last three decades, as a destination country for interna- cation are responsible for its success, rising to about 50 percent today from tional students, behind only the United and the possible consequences of these about 25 percent in the early 1970s. States, the United Kingdom, France, European reforms for the future of Much of that growth has come from and Germany), the author concludes American higher education. This paper new demand for advanced study in sci- that the relative status gap between explores the issues from a theoretical ence and engineering from countries Chinese and American universities and empirical perspective, focusing on rapidly climbing a development trajec- will remain until further reforms are the benefits of flexibility and competi- tory, particularly South Korea, India, implemented. tion associated with the American sys- Taiwan, and — more recently — China. E. Han Kim and Min Zhu inves- tem of higher education. Bound and Turner describe the trends tigate the overseas programs of U.S. Borghans and Corvers document in the distribution of foreign doctorate universities. They discover two waves changes in research and higher educa- students across programs of different of overseas programs: a supply driven tion in Europe and investigate poten- levels of quality and institutional con- wave during the late 1980s to the mid- tial explanations for the strong increase trol and resources, and at both public 1990s, and the current wave begin- in its international orientation. While universities and universities outside the ning in the early 2000s. Their data higher education started to grow sub- most highly ranked. Expansion of doc- reveal that tuition-dependent univer- stantially around 1960, only a few torate education propelled by the flow sities are more likely to offer over- decades later, research and higher edu- of foreign students has been particu- seas programs, and real GDP per cap- cation transformed gradually to the larly marked outside the most highly ita and tertiary school-age populations American standard. Decreased com- ranked programs and at public sector are two key determinants of the loca- munication costs are the likely causes universities. These differential trends tion choice. Also, U.S. universities of this trend. This transformation is highlight heterogeneity in the supply offer lower tuition discounts in coun- most clearly revealed in the change in response among different universities. tries with higher real GDP per capita. the language used in research from the Whether changes in doctorate produc- The authors conclude that universities national language/Latin, German, and tion are complementary to other uni- behave much like multinational corpo- French to English. This suggests that versity outputs, such as undergraduate rations in their overseas investments: returns to scale and the transferabil- education, helps to explain the pattern economics, not altruism, is the key ity of research results strongly influ- observed across universities. driver of U.S. overseas programs. ence adaptation of the international Black and Stephan study the role In his investigation of Indian standard. of foreign graduate students and post- higher education, Kapur finds an edu- Sunwoong Kim examines changes doctoral scholars (postdocs) in univer- cational system that, paradoxically, is in migration patterns of U.S.-trained sity lab science. To do this, they analyze simultaneously collapsing and thriving. Korean PhDs over the last several authorship patterns for a six-month Traditional multi-disciplinary univer- decades. Roughly speaking, three differ- period for articles published in Science sities are in a deplorable state, reflect- ent periods are identified: brain drain with a last author who is affiliated with ing cleavages afflicting India’s political (1950–70), brain gain (1970–97), and

NBER Reporter • 2008 Number 4 27 brain competition (since 1997). The lege in advanced countries has risen in U.S.-born candidates for advanced first period is typical of low-income above those in the United States in degrees does not reflect educational countries in which talented students some countries, while higher educa- deficiencies but other factors, nota- go to a rich country for further educa- tion in developing countries is increas- bly the attraction of careers in higher- tion and stay there after their training; ing at a rapid pace, greatly diminishing paying fields such as business or law. the later two periods reflect a more bal- the U.S. share of the world’s university Though fewer U.S. citizens are pursu- anced circulation between Korea and students and graduates. Freeman pre- ing doctoral degrees in science, tech- the United States. The author finds dicts a continued decrease in the U.S. nology, engineering, and mathematics, the Korean experience useful as a lead- share of global university enrollment, the United States continues to lead the ing-edge example of the international- while deferring judgment as to whether world in production of doctorates, and a ization of U.S. higher education with the United States should respond by significant proportion of these students China and India. encouraging immigration, or rather stay in the United States. Bettinger Between 1970 and 2006 the num- through off-shoring of some university concludes that the shortage of scien- ber of students enrolled in institu- graduate-level work overseas. tists and engineers is overstated. tions of higher education increased Over the last forty years, the supply These papers will be considered for from 29 million to over 141 million. of U.S.-born scientists and engineers publication in an NBER Conference Freeman examines the implications of has dropped dramatically. Bettinger Volume by the University of Chicago this expansion in foreign higher edu- sees worries of shortages as misplaced, Press and will be posted at “Books in cation for the United States. The pro- however. Interest in the sciences among Progress” on the NBER’s website. portion of young people going to col- undergraduates remains high. The fall

✴

28 NBER Reporter • 2008 Number 4 Europe and the Euro An NBER conference on “Europe Discussant: Martin Feldstein, Harvard Università di Bologna and the Euro” took place on October University and NBER 17 and 18 at Bocconi University in Anil K Kashyap, University of Milan. The conference highlighted Antonio Fatas and Ilian Mihov, Chicago and NBER, and Reint the success of the first ten years of INSEAD, “The Euro and Fiscal Gropp, European Business School, “A the Euro in terms of product mar- Policy” New Metric for Banking Integration ket reforms, increased trade between Discussant: Roberto Perotti, in Europe” members, successful monetary pol- Università Bocconi Discussant: Loretta Mester, Federal icy by the European Central Bank Reserve Bank of Philadelphia (ECB) — both in normal times and Jeffrey A. Frankel, Harvard in the most recent turbulent months, University and NBER, “The Alberto Giovannini, Unifortune progress towards more financial lib- Estimated Effects of the Euro Asset Management, Milan, “The eralization, and integration. However, on Trade: Why Are They Below Integration of European Financial the conference also noted the need for Historical Evidence on Effects of Markets: Why Has It Not Happened further progress in terms of labor mar- Monetary Unions among Smaller Yet?” ket reforms, banking supervision and Countries?” Discussant: Richard Portes, London regulation on the European level, inte- Discussant: Silvana Tenreyo, London Business School gration of financial markets, and syn- School of Economics chronization of business cycles. Lucrezia Reichlin, European Central Organizers Alberto F. Alesina, Ulf Soderstrom, Universita Bocconi, Bank, and Domenico Giannone, NBER and Harvard University, “Re-Evaluating Swedish Membership Universite Libre de Bruxelles, and Francesco Giavazzi, NBER and in the EMU: Evidence from an “Business Cycles in the Euro Area” Bocconni University, chose these Estimated Model” and Riccardo Discussant: Olivier J. Blanchard, papers to discuss: DiCecio and Edward Nelson, Federal MIT, NBER, and the International Reserve Bank of St. Louis, “Britain Monetary Fund Silvia Ardagna, Harvard University; and the Euro: Re-considering the Pros Alberto F. Alesina; and Vincenzo and Cons” Steven G. Cecchetti, Bank for Galasso, Università Bocconi “The Discussant: Carlo Favero, Università International Settlements and Euro and Structural Reforms” Bocconi NBER, and Kermit L. Schoenholtz, Discussant: Otmar Issing, European Citigroup, “ECB Policy: The First Central Bank Matteo Bugamelli and Roberta Decade and Beyond” Zizza, Bank of Italy; and Fabiano Discussant: Pervenche Beres, Barry Eichengreen, University of Schivardi, Università di Cagliari, European Parliament California, Berkeley and NBER, “Is “The Euro and Firm Restructuring” the Euro Forever?” Discussant: Gianmarco Ottaviano,

Alesina and his co-authors inves- connection between the adoption of rency should not be minimized. Nor is tigate whether the adoption of the the Euro and labor market reforms, at it obvious that the economic problems Euro has facilitated structural reforms, least as measured by changes in the leg- of the participating member states can which they define as deregulation in islation of primary labor markets. Their be significantly ameliorated by aban- the product markets and both liberal- paper also discusses issues concerning doning the Euro, although that possi- ization and deregulation in the labor the sequencing of labor market versus bility cannot be dismissed either. markets. After reviewing theoretical product market reforms. The creation of a single currency arguments linking the adoption of the In his paper, Eichengreen con- in Europe was accompanied by signif- Euro to structural reforms and investi- cludes that it is unlikely that one or icant changes in the institutional set- gating the empirical evidence, they find more members of the Euro area will ting for fiscal policy. Fatas and Milhov that the adoption of the Euro has been leave in the next ten years and that the ask whether these institutional changes associated with an acceleration of the total disintegration of the Euro area is have led to a change in the conduct of pace of structural reforms in the prod- even more unlikely. The technical diffi- fiscal policy by the members of the Euro uct market. In contrast, they find no culties of reintroducing a national cur- area. The authors review the behavior

NBER Reporter • 2008 Number 4 29 of fiscal policy after the introduction tries and all years, thereby bringing to that the Euro has been accompanied of the Euro in several dimensions: pro- bear as much information as possible by a reallocation of activity within, cyclicality, volatility, coordination, and on questions such as the proper coef- rather than across, sectors. Since adop- the role of automatic stabilizers. They ficients on common borders and lan- tion of the Euro, productivity growth characterize how the common currency guages in a gravity model, then the has been relatively stronger in country- and the constraints associated with the effect of the Euro in the five-to-eight sectors that once relied more on com- Stability and Growth Pact have shaped year interval is large and comparable to petitive devaluations to regain price fiscal policy among the members of the effect of the other non-Euro mon- competitiveness. Firm-level evidence the union. In order to provide a more etary unions. from Italian manufacturing confirms complete picture of fiscal policy, they Soderstrom revisits the potential that low-tech businesses, which argu- also report results related to the behav- costs and benefits for Sweden of join- ably benefited most from devaluations, ior of fiscal policy at the national level. ing the Economic and Monetary Union have been restructuring more since the Their results show that despite the sig- (EMU) of the European Union. He adoption of the Euro. Restructuring nificant change in the institutional set- shows that the Swedish business cycle has entailed a shift of business focus ting, the behavior of fiscal policy in the since the mid-1990s has been closely from production to upstream and Euro area is mildly procyclical and has correlated with that of the Euro area downstream activities, such as prod- not changed much since the introduc- economies, more so than that of some uct design, advertising, marketing, tion of the new currency. In contrast, current EMU members. Then, using an and distribution, and a corresponding U.S. fiscal policy has become distinctly estimated model of the Swedish econ- reduction in the share of blue-collar countercyclical. Also, there has been omy, he shows that while asymmetric workers. a broad-based decline in the volatil- shocks have been important fluctua- Most observers have concluded that ity of discretionary fiscal policy in all tions in the Swedish economy since while money markets and government of the major economies. Furthermore, 1993, to a large extent the exchange bond markets are rapidly integrating the discrepancy of fiscal policy across rate has acted to destabilize, rather following the introduction of the com- Euro-area countries — measured by the than stabilize, the economy. Finally, mon currency in the Euro area, there is dispersion of cyclically adjusted bal- his model predicts that Swedish infla- little evidence that a similar integration ances — has decreased threefold since tion and GDP growth might have been process is taking place for retail banking. 1999. slightly higher if Sweden had been a Data on cross-border retail bank flows, Frankel seeks to address two ques- member in EMU since 1999, but also cross-border bank mergers, and the law tions in his paper. First, do the effects that GDP growth might have been of one price reveal no evidence of inte- on intra-Euroland trade estimated in more volatile. EMU membership also gration in retail banking. Kashyap and the Euro’s first four years hold up in the could have implied higher inflation in Gropp show that the previous tests of second four years? The answer is yes. Sweden in 2004–5, when inflation was bank integration are weak, in that they Second, and more complicated, what exceptionally low. Thus, the evidence is are not based on an equilibrium con- is the reason for the large discrepancy not conclusive. cept and are neither necessary nor suf- vis-à-vis other currency unions, a gap Developments in open-econ- ficient statistics for bank integration. of between 15 and 200 percent? First, omy modeling, and the accumula- They propose a new test of integration the Euro is still very young. Second, tion of experience with the mone- based on convergence in banks’ prof- the European countries are much big- tary policy regimes prevailing in the itability. The new test emphasizes the ger than most of those who had formed United Kingdom and the Euro area, role of an active market for corporate currency unions in the past. Third, have increased our ability to evalu- control and of competition in bank- there is endogeneity about the decision ate the effects that joining a monetary ing integration. European listed banks’ to adopt an institutional currency link: union would have on the U.K. econ- profitability appears to converge to perhaps the high correlations estimated omy. DiCecio and Nelson analyze the a common level. There is weak evi- in earlier studies were spurious, an arti- debate on the U.K.’s monetary policy dence that competition eliminates high fact of reverse causality. Contrary to options using a structural open-econ- profits for these banks, and underper- expectations, though, Frankel finds omy model. forming banks tend to show improved no evidence that any of these factors Bugamelli and his co-authors test profitability. Unlisted European banks explains any share of the gap, let alone whether and how the adoption of the differ markedly. Their profits show no all of it. Instead, surprisingly, the dis- Euro, narrowly defined as the end of tendency to revert to a common target crepancy appears to stem from sample competitive devaluations, has affected rate of profitability. Overall, the bank- size. If one estimates the effects of the member states’ productive structures, ing market in Europe appears far from Euro versus other monetary unions in distinguishing between within- and being integrated. In contrast, in the a large sample that includes all coun- across-sector reallocation. They find United States both listed and unlisted

30 NBER Reporter • 2008 Number 4 commercial banks profits converge to no significant change can be detected vantage points. The researchers share the same target, and high profit banks since 1999. For the other countries, the assessment of their interviewees see their profits driven down quickly. there is a lot of uncertainty. No clear that the ECB has enjoyed many more Giovannini explains what a single, change since the EMU can be identi- successes than disappointments. These integrated European securities market fied in either group. As for the Euro successes reflect the ECB’s design and is and why we do not have it yet. He area business cycle, per capita GDP implementation. Looking forward, the argues that any market, including a secu- growth since 1999 has been lower than authors highlight the unique challenges rities market, is defined by the arrange- what could have been predicted on posed by enlargement and, especially, ments put in place to ensure delivery of the basis of historical experience and by the Euro area’s complex arrange- goods and payments to the counterpar- U.S. observed developments. The gap ments for guarding financial stability. ties in each trade (post-trading arrange- between U.S. and Euro area GDP per In the latter case, the key issues are ments). Analyzing these arrangements capita has been 30 percent on aver- coordination in a crisis and harmoni- is the most reliable way to assess the age since 1970, and there is no sign of zation of procedures. As several inter- extent to which there is integration in a catching up or of further widening. viewees suggested, in the absence of a geographic area like the EU or the Euro In their history of the first decade new organizational structure for secur- Area. In his paper, Giovannini analyzes of European Central Bank (ECB) pol- ing financial stability, the current one post-trading arrangements in the EU icy, Cecchetti and Schoenholtz dis- will need to function as if it were a sin- and discusses their reform, the objec- cuss key challenges for the next decade. gle entity. tive of which is to obtain a single EU Beyond the ECB’s track record and an These papers will be considered for securities market. array of published critiques, their anal- publication in an NBER Conference Reichlin and her co-authors note ysis relies on unique source material: Volume by the University of Chicago that in those EMU countries that extensive interviews with current and Press and will be posted at “Books in started from similar initial conditions, former ECB leaders and with other pol- Progress” on the NBER’s website. in terms of real activity in the 1970s, icymakers and scholars who viewed the business cycles are very similar and evolution of the ECB from privileged

✴

NBER Reporter • 2008 Number 4 31 Understanding Long-Run Economic Growth: A Conference Honoring the Contributions of Kenneth Sokoloff The NBER, the All-UC Economic a Time in the Americas: Land and University and NBER History Group, and the University of Immigration Policy in the New World” California, Los Angeles (UCLA) jointly Discussant: Peter Lindert, University of Robert W. Fogel, University of Chicago sponsored “Understanding Long-Run California, Davis and NBER, “The Impact of the Asian Economic Growth: A Conference Miracle on the Theory of Economic Honoring the Contributions of Kenneth James A. Robinson, Harvard University Growth” Sokoloff ” on November 7 and 8 in Los and NBER, and Camilo Garcia, MIT, Angeles. NBER Research Associates “The Myth of the Frontier” B. Zorina Khan, Bowdoin College and Dora L. Costa and Naomi R. Lamoreaux Discussant: Ronald Rogowski, NBER, “Premium Inventions: Patents of UCLA served as co-organizers of this University of California, Los Angeles and Prizes as Incentive Mechanisms memorial tribute to Sokoloff, who died in Britain and the United States, in May 2007 at the age of 54. He had Stephen Haber, Stanford University 1750–1930” been a member of the NBER’s Program and NBER, and Victor Menaldo, Discussant: Manuel Tratjenberg, Tel on the Development of the American Stanford University, “Do Natural Aviv University Economy and a professor of economics Resources Fuel Authoritarianism? A at UCLA for decades. Reappraisal of the Resource Curse” Sukkoo Kim, Washington University More than 150 scholars attended Discussant: Jeffrey Frieden, Harvard in St. Louis and NBER, and Sebastian the conference, with participants com- University Galiani, Washington University in St. ing from five different continents (North Louis, “The Law of the Primate City in and South America, Europe, Asia, and Daniel Kaufmann, World Bank, the Americas” Australia). In addition to the formal aca- “Corruption and Capture in Economic Discussant: Edward Leamer, University demic program, Douglass C. North, co- Development: A Few Reflections and of California, Los Angeles and NBER winner of the 1993 Nobel Prize in eco- Some Evidence” nomics, gave an after-dinner address that Discussant: Daniel Treisman, University John Majewski, University of interwove reminiscences about Sokoloff of California, Los Angeles California, Santa Barbara, and Daniel with an impassioned brief for the impor- Bogart, University of California, tance of studying economic history and, Claudia Goldin and Lawrence F. Katz, Irvine, “Population Densities, Political in particular, the history of economic Harvard University and NBER, “Mass Structures, and the Early Corporation: institutions. Robert W. Fogel, who Secondary Schooling and the State: The The Transportation Revolution in the shared the Nobel Prize with North, gave Role of State Compulsion in the High U.K. and United States” a paper at a session immediately preced- School Movement” Discussant: John J. Wallis, University of ing the dinner, and the two great scholars Discussant: David Card, University of Maryland and NBER enlivened the evening with their ongo- California, Berkeley and NBER ing repartee about how economic history Jean-Laurent Rosenthal and Philip T. should be written. Naomi R. Lamoreaux; Kenneth L. Hoffman, Caltech, and Gilles Postel- These papers were presented and Sokoloff; and Dhanoos Sutthiphisal, Vinay, EHESS, “History, Geography, discussed: McGill University and NBER, “The and the Markets for Mortgage Loans in Reorganization of Inventive Activity in 19th Century France” Stanley L. Engerman, University the United States in the Early Twentieth Discussant: Joel Mokyr, Northwestern of Rochester and NBER (joint with Century” University Kenneth L. Sokoloff), “Once Upon Discussant: Ariel Pakes, Harvard

Engerman and Sokoloff examine tematic patterns in the evolution of insti- the same time, and were extremely diverse land and immigration policies across the tutions. The case of the Americas pro- with respect to many aspects of their range of colonies/societies established by vides an excellent natural laboratory for endowments. Second, although their the Europeans in the New World over studying the effects of different forces on initial conditions differed enormously the sixteenth, seventeenth, eighteenth, the development of institutions. First, in some respects, nearly all of these New and nineteenth centuries to improve our these societies were all settled by a lim- World societies had a relative abundance understanding of whether there are sys- ited set of European countries at roughly of land and natural resources, and came to

32 NBER Reporter • 2008 Number 4 specialize quite early in their histories in that fragility is the use of pooled cross- state laws. The expansion of state compul- agriculture and mining. The policies they sectional data to address a question about sory schooling and child labor laws from adopted toward the ownership and use of change over time. Haber and Menaldo 1910 to 1939 can, at best, account for 6 to 7 land, and the openness toward labor flows, suggest that when testing theories about percent of the increase in high school enroll- had very significant implications for their processes that take place within countries ments and can account for about the same long-run paths of development. This com- over time, assembling time-series data- portion of the increase in the eventual edu- parative study not only helps to establish sets designed to operationalize explicitly cational attainment for the affected cohorts a systematic record, but also attempts to specified counterfactuals better matches the over the period. The “state,” in the form of identify salient factors in accounting for theory and empirics than do regressions localities, already was providing educational the variation over time and place in the centered on the cross-sectional analysis of resources in the United States. Compulsory design of strategic institutions. longitudinally truncated data. education laws had larger impacts in other Does the presence of an open fron- Although little known to many, Ken nations where the laws compelled the state tier explain why the United States became Sokoloff ’s intellectual and geographical to expand educational resources. democratic and, at least implicitly, prosper- adventures also led him to contribute to The standard view of U.S. technologi- ous? In their paper, Robinson and Garcia the field of development. Through his cal history is that the locus of invention begin with the contradictory observation, deep historical understanding of the role shifted during the early twentieth century that almost every Latin American country of endowment and institutions, he pro- to large firms whose in-house research lab- had a frontier in the nineteenth century as vided insights for Latin America today, oratories were superior sites for advancing well. They show that while the data does and for other areas. Kaufmann briefly the complex technologies of the second not support the Frontier thesis, it is con- reviews some of Sokoloff ’s contributions industrial revolution. In recent years, this sistent with a more complex “conditional to development. Then he focuses on a view has been subject to increasing criti- Frontier thesis”: that the effect of the fron- study of one narrow link between gover- cism. At the same time, new research on tier is conditional on the way it was allo- nance and development, which intrigued equity markets during the early twentieth cated, and that this in turn depends on Ken as a scholar and teacher: namely, the century suggests that smaller, more entre- political institutions at the time of fron- role of state capture in development. He preneurial enterprises were finding it easier tier expansion. They show that for coun- summarizes recent research on the effects to gain financial backing for technological tries with the worst political institutions, of capture on development, and goes on discovery. Lamoreaux and her co-authors there is a negative correlation between the to present new empirical results based on use data on the assignment (sale or transfer) historical extent of the frontier and con- a micro-dataset comprised of thousands of patents to explore the extent to which, temporary income per capita. For coun- of investment projects in developing coun- and how, inventive activity was reorganized tries with better political institutions, this tries whose performance has been rated by during this period. They find that two alter- correlation is positive. The effect of the the World Bank’s evaluation unit. He native modes of technological discovery frontier on democracy is positive irrespec- investigates the links between state cap- developed in parallel during the early twen- tive of initial political institutions, but it is ture, administrative corruption, and other tieth century. The first, concentrated in larger the better were these institutions. factors and controls and the performance the Middle Atlantic region, centered on Is there a relationship between “point of such investment projects. The results large firms that seem to have owed their source” natural resource dependence and suggest non-trivial effects of capture on prominence less to R and D labs than to authoritarianism? In order to answer this investment project performance. their superior access to the region’s rapidly question, Haber and Menaldo develop In the three decades after 1910, the growing equity markets. The other, located unique datasets that allow them to focus fraction of U.S. youths enrolled in public mainly in the East North Central region, on within-country variance in resource and private secondary schools soared from consisted of smaller, more entrepreneur- dependence and regime types. Their 18 to 71 percent and the fraction graduat- ial enterprises that drew primarily on local results indicate that dependence on oil ing increased from 9 to 51 percent. At the sources of funds. Both modes seem to have and minerals is not associated with the same time, state compulsory education and made roughly equivalent contributions to undermining of democracy or less com- child labor legislation became more strin- technological change during this period. plete transitions to democracy. These gent. It might appear from the timing that Fogel began by reviewing the concept results are at variance with a large body the laws caused the increase in education of growth theory, tracing its origins and its of scholarship that uses pooled cross- rates. Goldin and Katz evaluate that pos- evolution. After World War II, there were sectional techniques and finds a nega- sibility using contemporaneous evidence wide-ranging debates about the future of tive relationship between natural resource on enrollments and the microdata from economic development for the American dependence and democracy. The authors the 1960 census to examine the effect of economy, the economies of Europe, and subject those cross-sectional results to a the laws on overall educational attainment. the economies of less developed nations. battery of standard diagnostics and find Their estimation approach exploits cross- The pessimistic views of stagnation for that they are very fragile; the source of state differences in the timing of changes in the United States and other OECD econ-

NBER Reporter • 2008 Number 4 33 omies — because there was something that the award of prizes tended to be less their transport links. inherently unstable in the operation of systematic than that of patents and more Rosenthal, Hoffman, and Postel- capitalistic economies — were proved to susceptible to misallocation, but the results Vinay take two themes in Kenneth be unfounded. However, the growing gap varied by institutional context. If inventors Sokoloff ’s research —the negative effects in income between developed and less respond to expected benefits, these findings of inequality and the positive effects of developed nations led to an emphasis on imply that prizes may offer fewer incentives large, competitive markets — and exam- cultural and ideological barriers to eco- for investments in inventive activity. ine their impact on mid-nineteenth-cen- nomic growth in poor countries, wor- Kim and Ghaliani explore the causes tury French mortgage markets. In partic- ries about over-saving, and concerns that of urban primacy in the Americas using the ular, they document a negative externality export-led growth was exploitative. The insight that primate cities are often politi- from inequality: when the distribution remarkable growth of the four Asian cal capitals. Using extensive data on cities, of wealth is too skewed, the middle class dragons — Hong Kong, Singapore, they estimate the impact of capital city sta- is excluded from the market. But there is South Korea, and Taiwan — and, more tus on urban concentration after control- also a second, positive externality gener- recently, of Malaysia, China, Indonesia, ling for geographic, climatic, and economic ated by the geographical density of mar- and Thailand was not only not forecast, factors. They find that political capitals, kets. The authors distinguish its effects but also was believed to be a tempo- both national and provincial, contribute in the credit market itself from possible rary fluke. Fogel concluded by predict- significantly more to urban concentration spillovers from the comparable external- ing that by 2030 China and Southeast in Latin America than in North America ity in product markets. They proceed in Asia combined would have a total GDP although there are important country varia- four steps: 1) summarizing the sources that exceeds that of the United States tions within these areas. They suggest that of their data and their aggregate findings and the five largest European countries one possible cause of the differing patterns on French credit markets; 2) analyzing combined. Nevertheless, U.S. per capita of urban development in the Americas is local credit markets, which suggest that income in 2030 very likely will still be sev- the differences in the centralization of polit- inequality had adverse effects on lend- eral times greater than that of China. ical power in the Americas, a factor which ing; 3) searching for the positive exter- A number of economists have been has deep colonial roots. nality by examining loans between inhab- persuaded by the implications of theoreti- Majewski and Bogart compare the itants of different cities and towns; and cal models of prizes and subsidies and have evolution of transportation organizations 4) showing that there were indeed posi- begun to lobby for these policies as supe- in the United Kingdom and the United tive network externalities in credit mar- rior alternatives to patent institutions. The States — both world leaders in transport kets, which are consistent with a queuing late-eighteenth and nineteenth centuries development by the mid-nineteenth centu- model. Although most loans were local, provide a natural experiment for studying ry — with a focus on the differences in their the credit network allowed for significant the emergence, evolution, and effects of chartering regimes. They show that U.S. state inter-market flows of resources, they find. different incentive systems for technologi- governments incorporated far more trans- These papers will be considered for cal innovations. Khan’s analysis is based on portation companies per person at far lower publication in an NBER Conference samples of “great inventors” and “ordinary fees than did the U.K. Parliament. Their ini- Volume by the University of Chicago inventors” in Britain and the United States, tial investigation suggests that the key dif- Press and will be posted at “Books in and includes their patents and inventions, ference was the greater degree of democ- Progress” on the NBER’s website. as well as prizes granted during the criti- racy in the United States and its competitive cal transition from the First to the Second economic environment in which cities and Industrial Revolutions. The results suggest localities were engaged in a race to improve

34 NBER Reporter • 2008 Number 4 Microeconomics, Efficiency, and Welfare: Lessons for Latin America

The NBER and Universidad de Marcela Eslava, Universidad de Los Rodrigo Soares, University of los Andes in Santiago, Chile co-spon- Andes; John Haltiwanger, University Maryland and NBER, and Romero sored this year’s Inter-American Seminar of Maryland and NBER; Adriana Rocha, PUC, “Evaluating the on Economics on “Microeconomics, Kugler, University of Houston and Impact of Community-Based Health Efficiency, and Welfare: Lessons for NBER; and Maurice Kugler, Wilfrid Interventions: Evidence from Brazil’s Latin America.” The conference was held Laurier University, “Lower Barriers to Family Health Program” in Santiago on November 10 and 11. Entry, Experimentation and Aggregate Discussant: Ernesto Schargrodsky Organizers Sebastián Edwards of NBER Productivity: Lessons from Colombian and University of California, Los Angeles, Manufacturing Plants” Cesar Martinelli, ITAM, “Press and and Alexander Galetovic of Universidad Discussant: Roberto Álvarez, Banco Good Government” de los Andes chose the following papers Central de Chile Discussant: Alexander Galetovic to discuss: Fernando Diaz, Alexander Galetovic, Alejandra Mizala, Universidad de Stephen H. Haber, Stanford University and Ricardo Sanhueza, Universidad de Chile, and Miguel Urquiola, Columbia and NBER, and Aldo Musaccio, los Andes, “Mobile-Fixed Substitution University and NBER, “School Stanford University, “Foreign Entry and Telecomm Liberalization” Markets: The Impact of Information and the Mexican Banking System, Discussant: Claudio Agostini, Approximating Schools’ Effectiveness” 1997–2007” Universidad Alberto Hurtado (NBER Working Paper No. 13676) Discussant: Alejandro Micco, Chilean Discussant: Francisco Gallego, Ministry of Finance David MacKenzie, The World Universidad Católica Bank, and Ernesto Schargrodsky, Hoyt Bleakley, University of Chicago Universidad di Tella, Juan Eberhard, Yale University, and and NBER, and Kevin Cowan, Banco “Buying Less, but Shopping More: The Eduardo Engel, Yale University and Central de Chile, “Mishmash or Use of Non-market Labor during a NBER, “The Educational Transition Mismatch: Balance Sheet Effects and Crisis” and Decreasing Wage Inequality in Emerging Market Crises” Discussant: Harald Beyer, Centro de Chile” Discussant: Sebastián Edwards Estudios Públicos Discussant: Andrea Repetto, Universidad Adolfo Ibáñez

How does the entry of foreign banks ing going on because foreign-owned banks Eslava, Haltiwanger, Kugler, and affect pricing and the availability of credit may wish to avoid Mexico’s difficult prop- Kugler use plant output and input prices in developing economies? The Mexican erty rights environment. to decompose the profit margin into three banking system provides a quasi-experi- Bleakley and Cowan critically assess parts: productivity, demand shocks, and ment to address this question, because in the recent empirical literature on the input costs. They find that market-ori- 1997 the Mexican government radically importance of dollar debt and balance- ented reforms increase the effect of mar- changed the laws governing the foreign sheet effects in the emerging-market finan- ket fundamentals on market entry. Their ownership of banks; the foreign market cial crises of the 1990s. Using a simple definition of a market is one industry in share increased five-fold between 1997 model, they discuss which specifications a particular region. Prior to reforms, they and 2007 as a result. Haber and Musaccio are theoretically appropriate, and they pro- find, market fundamentals often work in construct and analyze a panel of Mexican vide additional insights as to the proper the wrong direction with respect to the bank financial data covering this period interpretation of the reduced-form evi- determinants of entry. These findings sug- and find no evidence that foreign entry dence in the literature. Using harmonized gest that frictions from poor market-ori- increases the availability of credit. Their microdata on corporations across a dozen ented institutions have distorted the pro- analysis also indicates that foreign banks countries in Latin America and Asia, they cess of entry. The authors also find that screen borrowers more closely and charge replicate common specifications for the market reforms increased the marginal higher lending rates than domestic banks. effect of dollar debt on investment and effect of productivity, and of other mar- One of their most robust findings is that output at the firm level. In this light, they ket fundamentals, on plant entry. There is foreign ownership is associated with a suggest that the literature on corporate also evidence that market reforms yielded decrease in housing lending. This sug- level currency mismatch is hardly a mish- an environment conducive to greater gests that there may be less housing lend- mash, but in fact is quite concordant. market experimentation.

NBER Reporter • 2008 Number 4 35 The development and diffusion of mainly at earlier ages. Municipalities in a sharp regression discontinuity to esti- mobile telephony has been fast and uni- the poorest regions of the country partic- mate the effect that being identified as a versal, but whether mobile and fixed ularly benefit from the program. For these SNED winner has on schools’ enrollment, phones are substitutes is still a conten- regions, implementation of the program tuition levels, and socioeconomic compo- tious issue. Most studies tend to find is also robustly associated with increased sition. Through five applications of the positive cross-price elasticity; some sug- labor supply of men between ages 41 program, they find no consistent evidence gest that there is complementarity. In any and 55, and reduced fertility of women that winning a SNED award affects these case, it seems fair to say that the evidence between ages 31 and 40. outcomes. is still not compelling. Consequently, tar- Martinelli presents a model of (pos- Eberhard and Engel find that iff liberalization of fixed telephony is still sibly hidden) capture of a democratically between 1975 and 1990 the wage of the an issue and it seems that regulation will elected government by the elite. He 90th (richest) percentile in Chile increased be with us for some time in most coun- shows that there is a U-shaped relation- faster than the median wage and the wage tries. Diaz, Galetovic, and Sanhueza ship between government turnover and of the 10th percentile. By contrast, from argue, however, that under particular cir- the extent of capture. When politicians 1990 on the wage of the 10th percentile cumstances one can perform a simple are very patient, economic conditions are and the median wage grew faster than the (non-statistical) test of substitution that bad, and press coverage of government wage of the 90th percentile. This is one of relies only on the standard properties of behavior is not reliable, the best equilib- many findings showing that wage inequal- demand curves. They perform this test rium for the majority of citizens requires ity in Chile has been falling, first slowly, for Chile and find compelling evidence of punishing governments with no re-elec- then faster, over the last two decades. The mobile-fixed substitution. tion, even if citizens know there is no cap- pattern emerges most clearly once cycli- Households that allocate time ture in equilibrium. If politicians are not cal components of wage inequality mea- between market and non-market uses very patient, then the best equilibrium for sures are removed. To understand this should respond to income variations by the majority may not imply government decrease in wage inequality, the authors adjusting the time they devote to shopping turnover, regardless of the quality of press group workers according to their cohort and other home production activities. coverage, even if citizens know that there and educational attainment and decom- MacKenzie and Schargrodsky exploit is a positive probability of hidden capture. pose the variance of (log) wages into high-frequency data on household expen- Finally, if politicians are quite impatient, within- and between-group variances. ditures to examine the use of changes in and when economic conditions are good They find that the significant decrease shopping intensity as a method of miti- and press coverage is not reliable, the best in wage inequality is largely attributable gating the effects of the 2002 Argentine equilibrium for the majority requires not to a decrease in between-group variance. economic crisis. They find that although re-electing governments. The evidence suggests that this decrease the total quantity and real value of goods The impact of competition on aca- in wage inequality was caused by a secular purchased fell during the crisis, consumers demic outcomes likely depends on whether decline in the skill premium for tertiary were doing more shopping. This increase parents are informed about schools’ effec- education resulting from the deregulation in shopping enabled households to seek tiveness or value added (which may or of this market in 1980. Somewhat surpris- out lower prices and to locate substi- may not be correlated with absolute mea- ingly, the marked decrease in the skill pre- tutes, allowing a given level of expendi- sures of their quality), and on whether mium can be found in the data one decade ture to buy more goods. The magnitude this information influences their school earlier, yet the corresponding decrease and prevalence of these effects suggest choices, thereby affecting schools’ market in wage inequality was canceled by a that this non-market use of labor can be outcomes. To explore these issues, Mizala Kuznets-type effect, where an increasing an important coping strategy for house- and Urquiola consider Chile’s SNED but still small fraction of workers with holds during a recession. program, which seeks to identify effec- tertiary education led to more inequality, Soares and Rocha analyze the direct tive schools, selecting them from within during the 1985–95 period. and indirect impacts of Brazil’s Family “homogeneous groups” of arguably com- These papers will be considered for Health Program. They estimate the effects parable institutions. Its results are widely publication in an NBER Conference of the program on mortality and on disseminated, and the information it gen- Volume by the University of Chicago household behavior related to child labor erates is quite different from that con- Press and will be posted at “Books in and schooling, employment of adults, and veyed by a simple test-based ranking of Progress” on the NBER’s website. fertility. They find that the program has schools (which in Chile, turns out to consistent effects on reductions in mor- largely resemble a ranking based on socio- tality throughout the age distribution, but economic status). The authors rely on

36 NBER Reporter • 2008 Number 4 Economic Aspects of Obesity Genetic factors cannot account for the tastes for physical activities choose to locate Lisa M. Powell, University of Illinois rapid increase in obesity since 1980 — these in areas with better access to these facilities. at Chicago, and Frank J. Chaloupka, factors change slowly over long periods of The economic consequences of obesity University of Illinois at Chicago and time. Therefore, economists can play a role are complex. One paper found that over- NBER in examining the determinants and con- weight teens have about the same levels of Discussant: Tinna Asgeirsdottir, University sequences of this trend, although the fac- educational attainment as teens of normal of Iceland tors at work are complex, and the policy weight. One explanation is that the over- prescriptions are by no means straightfor- weight offset the factors associated with “The Relationship between Neighborhood ward. To increase our understanding of poor health and discrimination by allocat- Quality and Obesity among Children” this subject, the NBER held a conference ing more time to schoolwork and less time Bisakha Sen, Stephen Mennemeyer, and on the “Economic Aspects of Obesity” at to sports and other leisure time activities. Lisa C. Gary, University of Alabama at Louisiana State University in Baton Rouge Another paper reported that the observed Birmingham on November 10 and 11. The organizers negative relationship between obesity and Discussant: Susan Averett, Lafayette were Michael Grossman, City University of wages is not caused by obesity per se but College New York Graduate Center and NBER, and rather by a factor such as physical attrac- Naci H. Mocan, Louisiana State University tiveness or discrimination. One paper con- “Studying the Child Obesity Epidemic and NBER. tained suggestive evidence that this nega- with Natural Experiments” Several of the conference papers tive relationship between obesity and wages Robert Sandy, Gilbert Liu, John addressed the role of reductions in real food could be attributable in part to low self- Ottensmann, Jeff Wilson, and O.T. Ford, prices in the period at issue, and reported esteem on the part of the obese. Indiana University; and Rusty Tchernis, that these declines could account for some These topics were discussed at the Indiana University and NBER of the increase in obesity. These papers bear conference: Discussant: Barbara Schone, Agency for on the question of whether taxes on food, Healthcare Research and Quality especially dense and high-caloric fast food, “Effects of the Minimum Wage on provide an effective public policy tool for Obesity” “Food Stamp Program and Consumption addressing obesity. The case for such taxes David O. Meltzer, University of Choices” is weakened if fully informed consumers Chicago and NBER, and Zhuo (Adam) Neeraj Kaushal, Columbia University and are taking account of all the costs of their Chen, Centers for Disease Control and NBER, and Qin Gao, Fordham University food choices, and strengthened if the obese Prevention Discussant: Kerry Anne McGeary, Drexel do not pay for their higher medical expen- Discussant: Kristin Butcher, Wellesley University ditures through differential payments for College health care and health insurance, and if body “Physical Activity: Economic and Policy weight decisions are responsive to the inci- “Outcomes in a Program that Offers Factors” dence of the medical care costs associated Financial Rewards for Weight Loss” Melayne M. McInnes, University of with obesity. One paper found that health John Cawley, Cornell University and South Carolina, and Judy A. Shinogle, insurance does indeed “make you fat.” But NBER, and Joshua A. Price, Cornell University of Maryland others at the conference argued that a tax University Discussant: James Sallis, San Diego State on fast food could actually increase caloric Discussant: Dhaval Dave, Bentley College University intakes via substitution towards non-taxed and NBER food. An alternative policy might be finan- “Effects of Weight on Adolescent cial rewards for weight loss. But another “Food Prices and Body Weight Among Educational Attainment” paper found very small average weight loss Older Americans” Robert Kaestner, University of Illinois at associated with worksite programs with this Dana Goldman and Darius Lakdawalla, Chicago and NBER; Michael Grossman; feature. Still other papers found evidence RAND and NBER, and Yuhui Zheng, and Ben Yarnoff, University of Illinois at that access to parks, gymnasiums, and other RAND Chicago recreational facilities increase exercise and Discussant: Michael Anderson, University Discussant: Jeffrey DeSimone, University reduce obesity. Whether this finding bears of California at Berkeley of Texas at Arlington and NBER directly on the suggestion that such facilities should receive public subsidies depends on “Economic Contextual Factors and Child “Where Does the Wage Penalty Bite?” the extent to which people with unobserved Body Mass Index” Christian A. Gregory, University of (continued)

NBER Reporter • 2008 Number 4 37 North Carolina at Greensboro, and Jay Bhattacharya and M. Kate Bundorf, “Obesity, Self-esteem, and Wages” Christopher J. Ruhm, University of Stanford University and NBER; Noemi Naci H. Mocan, and Erdal Tekin, North Carolina at Greensboro and NBER Pace, University College of London; and Georgia State University and NBER Discussant: Roy Wada, University of Neeraj Sood, Rand and NBER Discussant: Sara Markowitz, Emory California, Los Angeles Discussant: Alan C. Monheit, University University and NBER of Medicine and Dentistry of New Jersey “Does Health Insurance Make You Fat?” School of Public Health and NBER

Since the 1970s, the real minimum motion program that offered financial markets, grocery stores, and convenience wage in the United States has declined rewards for weight loss. The interven- stores. The researchers draw on longitu- by nearly half, and obesity has increased. tion varied by employer, in some cases dinal data from the Child Development Because consumption of food away from offering steady quarterly rewards and in Supplement of the Panel Study of home as been suggested as a major cause other cases requiring participants to post Income Dynamics, combined at the zip- of increasing obesity, and minimum wage a bond that would be refunded at year’s code level with food price data from labor is a major contributor to the cost end conditional on weight loss. Still oth- the American Chamber of Commerce of “fast food”, Meltzer and Chen exam- ers received no financial incentives at all Researchers Association and food-related ine whether changes in the minimum and serve as a control group. The authors outlet density data obtained from Dun & wage are associated with changes in body- find far higher attrition in this real-world Bradstreet. Their results show that higher weight over this period. Using data from intervention than has been experienced in fruit and vegetable prices are related to the Behavioral Risk Factor Surveillance similar interventions designed and oper- a higher body mass index (BMI) among System from 1984–2006, they test whether ated by researchers. Moreover, average children, with greater effects among those variations in the real minimum wage were weight loss attributable to the financial in low socioeconomic status (SES) fami- associated with changes in body mass rewards is small: after one year, it is 1.7 lies. Higher fast food prices are negatively index (BMI). They also examine whether pounds for those who posted a refund- related to child weight only among low- this association varied by gender, educa- able bond and is not significantly differ- income children. Increased supermarket tion, and income, and they test whether ent from zero for those facing steady quar- availability also lowers children’s weight the association varied over the BMI distri- terly rewards. outcomes among the low-SES children. bution. They find that a $1 decrease in the Goldman and his co-authors exam- These results provide evidence of the real minimum wage was associated with a ine the associations between food prices of potential effectiveness of using fiscal pric- 0.06 decrease in BMI. This relationship various kinds and body weight, for adults ing interventions, such as taxes and subsi- was significant across gender and income age 50 and over in the United States. The dies, and other interventions to improve groups and largest among the highest per- authors use the Health and Retirement supermarket access as policies to address centiles of the BMI distribution. Real Study, a longitudinal and nationally rep- childhood obesity. minimum wage decreases can explain 10 resentative survey of older Americans, Sen and co-authors explore whether percent of the change in BMI since 1970, along with data on geographical and tem- maternal perceptions of neighborhood they conclude. The declining real mini- poral variations in food price inside the quality affect children’s bodyweight out- mum wage has contributed to the increas- United States, to identify the associations. comes and, moreover, whether racial and ing rate of overweight and obesity in the They find that doubling the price of calo- ethnic differences in such perceptions may United States. ries or the price of fat reduces body weight explain any of the hitherto unexplained Obesity rates in the United States of low-income persons by 5 to 7 per- gap in bodyweight and obesity prevalence have doubled since 1980. Given the medi- cent. This implies that if food taxes were among whites and minorities. Their results cal, social, and financial costs of obesity, a imposed to reduce caloric intake, a sub- indicate that overall neighborhood qual- large percentage of Americans are attempt- stantial share of the reduction in obesity ity is not significantly related to children’s ing to lose weight at any given time but would occur among the population that bodyweight, but one particular character- the vast majority of weight loss attempts currently qualifies for Medicaid. isti — whether there is enough police pro- fail. Researchers continue to search for Powell and Chaloupka examine the tection in the neighborhood — is indeed safe and effective methods of weight loss, relationship between child weight and: related. Lack of police protection has and Cawley and Price evaluate one prom- the prices of energy-dense foods, such as robust and significant effects on the chil- ising method: offering financial rewards fast food, versus healthy foods, includ- dren’s BMI, although it has less robust for weight loss. They study data on 2,351 ing fruits and vegetables; the availability effects on the risk of becoming obese per employees in 15 worksites who partici- of fast food versus full-service restaurants; se. Additionally, lack of police protec- pated in a year-long worksite health pro- and access to food outlets, including super- tion appears to be associated with chil-

38 NBER Reporter • 2008 Number 4 dren spending more time watching tele- dence that the policies affected expendi- men are more similar to those for women: vision. Finally, there are differences in tures on specific food items. the findings for females (and the IV esti- perceptions about adequate police protec- McInnes and Shinogle examine cor- mates for males) suggest that it is not obe- tion between whites and minorities that relates and predictors of physical activity sity but rather some other factor — such as remain after controlling for other socio- over time with emphasis on economic fac- physical attractiveness — that may be pro- economic factors, and these differences tors. Using data for adults from the 2000– ducing the observed relationship between do explain part of the gap in bodyweight 5 Behavioral Risk Factor Surveillance BMI and wages. Using data from the between white and minority children. System (BRFSS) survey, they analyze the Medical Expenditure Panel Survey, the Sandy and his co-authors use clini- characteristics of individuals and their authors also provide estimates of the asso- cal records of successive visits by chil- environments that determine their level ciation between BMI and health expen- dren at pediatric clinics in Indianapolis of activity. Because BRFSS includes state ditures. These cast further doubt on the to estimate the effects on their body mass and county codes for each individual, the hypothesis that the wage penalties associ- of changes in the environment near their researchers are able to include additional ated with increasing BMI occur because homes. The sample is limited to children information regarding economic vari- the latter serve as an index for underlying who resided at the same address before ables, such as area unemployment, as well medical costs. and after the environmental change. The as price and supply variables. They find Bhattacharya and his co-authors use environmental factors are: fast food res- that certain area characteristics — such data from the Rand Health Insurance taurants, supermarkets, parks, trails, and as access to gyms, parks, and other recre- Experiment, in which people were ran- violent crimes, and 13 types of recre- ational facilities — increase the probabil- domly assigned to varying levels of health ational amenities derived from the inter- ity of exercise. It is possible that those peo- insurance, to examine the effect of insur- pretation of annual aerial photographs. ple with unobserved tastes for prevention ance coverage on body weight along the The researchers’ cross-sectional estimates activities, such as physical activity, may intensive-coverage margin. Then, they use are quite different from their fixed-effects choose to live in areas with better access nonlinear IV methods to estimate the (FE) estimates of the impacts of amenities to these facilities. Yet the results persist effect of type of insurance coverage (pri- locating near a child. In the cross section, when the authors attempt to control for vate, public, and none) on body weight, nearby fast food restaurants are associated this unobserved taste with a dichotomous in order to examine the effects of coverage with higher BMI and supermarkets with indicator for the receipt of a flu shot. along the extensive margin in the 1989– lower BMI. These results are reversed in Kaestner and his co-authors inves- 2004 waves of the National Longitudinal the FE estimates. The recreational ameni- tigate the association between weight Survey of Youth 1979. They find weak ties that appear to lower children’s BMI and adolescent educational attainment, evidence that more generous insurance are baseball and softball fields, fitness as measured by highest grade attended, coverage increases BMI. They find stron- areas, kickball diamonds, and volleyball highest grade completed, and drop out ger evidence that being insured leads to courts. status. Data for this study come from the a greater body mass index and a higher Using the Consumer Expenditure 1997 cohort of the National Longitudinal probability of obesity. Surveys for 1994–2004, Kaushal and Survey of Youth, which contains a large, Mocan and Tekin find that, among Gao study the effect of the Food Stamp national sample of teens between the ages a nationally representative sample of Program (FSP) on consumption patterns of 14 and 18. After controlling for a young American adults aged 21 to 26 in in families headed by low-educated single variety of observed characteristics, the 2001–2, body weight has an indepen- mothers in the United States. Their analy- researchers conclude that, in general, dent impact on self-esteem after control- sis suggests that the food stamp caseload teens who are overweight or obese have ling for a host of personal attributes, does not have any statistically significant levels of attainment that are about the including education, health status, and association with per capita expenditure same as teens with average weight. test scores. Specifically, being overweight on food in families headed by low-edu- Using data from the 1986 and 1999– or obese has a negative influence on the cated single mothers. The researchers find 2005 Panel Study of Income Dynamics, probability of having high self-esteem for that state and federal welfare reforms dur- Gregory and Ruhm estimate models that females (both white and black) and for ing the 1990s lowered the food stamp allow earnings to vary with BMI in a black males. There is no evidence of an caseload by approximately 18 percent, and highly flexible manner. Their results show impact of body weight on self-esteem the introduction of the Electronic Benefit that, for women, earnings peak at levels in case of white men. Wages of black Transfer cards and simplified reporting far below the clinical threshold of “obe- men and black women are influenced by procedures for re-certification of food sity” or even “overweight.” For men, the their body weight. Although not uni- stamps increased participation by about estimates suggest a reasonably flat BMI- formly statistically significant, there is a 7 percent. However, there is no evidence wage profile that peaks early in the “over- wage penalty for being obese, and there that these policies had any effect on total weight” category. However, the results of is some evidence of a wage penalty for food expenditure, nor any consistent evi- instrumental variables (IV) models for the being underweight in case of blacks. Self-

NBER Reporter • 2008 Number 4 39 esteem has no impact on black wages. weight does not affect white men’s wages. through this second channel is small. The authors’ results indicate that for both The results suggest that obesity has the most These papers will be considered for white men and women, self-esteem has serious impact on white women’s wages, publication in an NBER Conference an impact on wages. Having high self- because their wages are affected directly by Volume by the University of Chicago esteem is associated with a 3 percent wage obesity and indirectly through the impact of Press and will be posted at “Books in premium. White women’s wages are also obesity on self-esteem, although the mag- Progress” on the NBER’s website. influenced by their bodyweight, while body nitude of the wage penalty that emerges

NBER News

NBER Researcher Wins Nobel Prize in Economics

NBER Research Associate Paul standing of the determinants of trade E. Stiglitz (shared with Michael Spence), Krugman of Princeton University is and the location of economic activity. 2001; James J. Heckman and Daniel L. the winner of the 2008 Nobel Prize in His seminal papers … were instrumental McFadden, 2000; Robert C. Merton and Economics. Krugman has been affili- to the development of the new trade the- Myron S. Scholes, 1997; Robert E. Lucas, ated with the NBER since 1979 and is a ory … and inspired the new approach to Jr., 1995; Robert W. Fogel, 1993; Gary member of the Programs on International economic geography.” S. Becker, 1992; and the late George Trade and Investment and International Krugman joins a long list of cur- J. Stigler, 1982, Theodore W. Schultz, Finance and Macroeconomics. The rent and past NBER affiliates who have 1979, Milton Friedman, 1976, and Simon Prize Committee of the Royal Swedish received the Prize, including: Edward C. Kuznets, 1971. Academy of Sciences summarized his Prescott and Finn Kydland, 2004; Robert contributions as “deepen[ing] our under- F. Engle, 2003; George Akerlof and Joseph

Two New Directors-at-Large The NBER’s Board of Directors CEO of Citicorp and Citibank. He also At the same meeting, John S. elected two new at-large members at its served as chairman of the New York Clarkeson succeeded Elizabeth Bailey as September 8 meeting: Mohamed El-Erian Stock Exchange from September 2003 Chairman of the Board, and Kathleen and John S. Reed. El-Erian is CEO of until April 2005. He holds a Master B. Cooper was elected Vice Chairman. PIMCO, the world’s largest bond inves- of Science degree from MIT’s Sloan Clarkeson is Chairman Emeritus of the tor; he previously worked as the invest- School. Boston Consulting Group, Inc. Cooper ment manager of Harvard University’s These two new directors fill an open was Undersecretary for Economic Affairs endowment fund. Before moving to the slot and replace Roger W. Ferguson, Jr., of the U. S. Department of Commerce from private sector, El-Erian spent 15 years at who resigned as a director after becom- 2001–5 and is currently a Senior Fellow the International Monetary Fund. He ing CEO of TIAA-CREF earlier this at the Tower Center for Political Studies holds a Ph. D. in Economics from Oxford year. NBER President James M. Poterba, at Southern Methodist University. University. a professor of economics at MIT, also Reed is the former chairman and becomes a director-at-large.

40 NBER Reporter • 2008 Number 4 Program and Working Group Meetings

China Working Group Meets in Cambridge The NBER’s Working Group on University of California, San Diego and Jie Gan, Hong Kong University of China, directed by NBER Research NBER Science and Technology; Yan Guo, Associate Shang-Jin Wei of Columbia Peking University; and Chenggang Xu, University, met on October 3 and 4 in Li Jin, Harvard University, and Joseph University of Hong Kong, “What Makes Cambridge. These papers were presented P.H. Fan and Guojian Zheng, Chinese Privatization Work? The Case of China.” and discussed: University of Hong Kong, “Internal Discussant: Gary Jefferson, Brandeis Capital Market in Emerging Markets: University Shing-Yi Wang, New York University, Expropriation and Mitigating Financing “Credit Constraints, Job Mobility, and Constraints.” Irene Brambilla, Yale University and Entrepreneurship: Evidence from a Discussant: Wei Jiang, Columbia NBER; Galina Hale, Federal Reserve Property Reform in China.” University Bank of San Francisco; and Cheryl Discussant: Lakshmi Iyer, Harvard Long, Colgate University, “Foreign University Yan-Leung Cheung, City University of Direct Investment and Incentives to Hong Kong; Ping Jiang, University of Innovate and Imitate.” Lena Edlund, Columbia University; International Business and Economics; Discussant: C. Fritz Foley, Harvard Hongbin Li, Tsinghua University; and and Kai Li and Tan Wang, University University and NBER Junjian Yi and Junsen Zhang, Chinese of British Columbia, “Privatization and University of Hong Kong, “Sex Ratios Risk Sharing: Evidence from the Split Kalina Manova, Stanford University and Crime: Evidence from China’s One- Share Structure Reform in China.” and NBER, and Zhiwei Zhang, IMF, Child Policy” Discussant: Zhiwu Chen,Yale University “China’s Exporters and Importers: Firms, Discussant: Avi Eberstein, Harvard Products, and Trade Partners.” University Lakshmi Iyer; Xin Meng, Australian Discussant: Amit Khandelwal, Columbia National University; and Nancy Qian, University Shang-Jin Wei, and Xiaobo Zhang, Brown University, “Property Rights and IFPRI, “Sex Ratio Imbalances Stimulate Household Decisions: The Impact of Li Han, Stanford University, “Marketing Savings Rates: Evidence from the China’s Urban Housing Reforms.” Politics? Economic Reforms and the ‘Missing Women’ in China.” Discussant: Marcos Chamon, IMF Selection of Political Elites in China.” Discussant: Roger H. Gordon, Discussant: Yasheng Huang, MIT

Wang provides new evidence on the ing them to capitalize on the value of the plus of sons. While a national policy, its impact of private property rights on entre- real estate. The unbundling of housing implementation was local. At the pro- preneurship. He explores this issue in the benefits from state employment also con- vincial level, implementation was unre- context of a housing reform in urban tributed to the increase in entrepreneur- lated to contemporaneous economic China that allowed state employees rent- ship by facilitating labor mobility out of characteristics of the province. Instead, ing state-owned housing the opportunity the state sector. individual characteristics of the provin- to buy their homes at subsidized prices. Sex ratios (males to females) rose cial party secretary influenced the tim- Using the reform as an exogenous change markedly in China in the last two decades, ing. Moreover, leaders were systemati- in the capital constraints and mobility and crime rates nearly doubled. In their cally rotated such that ten years on, leader costs that influence individuals’ entry paper, Edlund, Li, Yi, and Zhang exam- characteristics were serially uncorrelated. into entrepreneurship, he estimates that it ine whether the two are causally linked. Using annual province-level data for the increased self-employment. He develops High sex ratios imply fewer married men, period 1988–2004, the authors show that a model of job choice to test two mecha- and marriage has been conjectured to a 0.01 increase in the sex ratio raised vio- nisms that might explain how the reform be a socializing force. The paper exploits lent and property crime rates by some 3 increased entrepreneurship, noting that the quasi-natural experiment generated percent, suggesting that the rise in excess it increased the ability of individuals to by the Chinese one-child policy, a policy males may account for up to one-seventh finance entrepreneurial ventures by allow- that is widely held to be behind the sur- of the overall rise in crime.

NBER Reporter • 2008 Number 4 41 Wei and Zhang note that Chinese ket to mitigate financing constraints, the cal burdens of disposing of laid-off work- households save about 25 percent of their efficiency of the internal capital market ers, tend to use the MBO method to income, contributing to one of the world’s is highest. privatize. highest current-account surpluses. The A fundamental question in finance Brambilla, Hale, and Long explore a life-cycle theory and precautionary sav- is whether and how removing barriers is new channel for FDI spillovers on domes- ings motive provide only an incomplete associated with efficiency gains. Cheung, tic firms in the host country that operates explanation for this. This paper proposes Jiang, Kai Li, and Wang study this through imitation of original products. a new hypothesis: China’s high and ris- question using share issue privatization They develop a model of heterogeneous ing sex ratio imbalance — too many boys in China that took place through the firms and allow domestic firms to choose relative to girls at birth — attributable to split-share structure reform. Prior to the among three alternatives: 1) not introduce a combination of a strict family planning reform, domestic A-shares were divided any new products, 2) introduce a new policy, parental preference for sons, and into tradable and non-tradable shares product line (innovate), or 3) develop inexpensive abortion technologies, may with identical cash flow and voting rights. a variety that is a very close substitute have induced the Chinese to postpone Under the reform, non-tradable share- to an existing product line developed consumption in favor of wealth accumu- holders negotiate a compensation plan by another firm (imitate). The presence lation. To avoid condemning their sons with tradable shareholders in order to of foreign firms generates spillovers via to life-long bachelorhood, families with a make their shares tradable. The key pre- increased incentives for imitation, as for- boy raise their saving rates. Other families dictions are: 1) the size of compensation eign firms introduce a range of original do not reduce their savings rates because made by the non-tradable shareholders products that are vertically differentiated there is a spillover channel. Across prov- to the tradable share holders is negatively with respect to original domestic prod- inces, local saving rates are strongly posi- correlated with the bargaining power of ucts. The model generates testable impli- tively associated with local sex-ratio imbal- non-tradable share holders; 2) the size cations that allow the authors to distin- ances, the authors find, after accounting of compensation is positively correlated guish empirically between the effects of for demographics and social safety nets. with the gain in risk sharing; and 3) the FDI on true innovation and on imitation. This effect is stronger in rural areas than size of compensation is negatively corre- They test the model’s predictions using in urban areas. Household-level data also lated with firm performance; results are firm-level panel data for China and find support this hypothesis: families with a consistent with our model’s predictions. that, consistent with the model, increased son tend to save more in regions with a The authors conclude that better risk FDI presence in a given industry leads to more skewed sex ratio, holding constant sharing is an important consideration in more imitation, but not necessarily more various household features. Households China’s share issue privatization. innovation, by domestic firms. with daughters do not reduce their sav- Iyer, Meng, and Qian analyze the Manova and Zhang provide a ings in these regions. The increase in impact of housing reforms in China. Their detailed overview of China’s participa- sex ratios accounts for about half of the preliminary results indicate that urban tion in international trade. Using newly increase in household savings nationally. housing reforms are associated with more available data on the universe of globally Using Chinese data, Fan, Jin, and households working in the private sector; engaged Chinese firms over the 2003– Zheng examine the internal capital mar- households eligible for housing loans are 5 period, they document the distribu- ket in emerging market business groups. more likely to be running private busi- tion of trade flows, product- and trade- They focus on two aspects that are less nesses after the reforms. These reforms do partner intensity across both exporting prominent in the developed markets: not increase households’ access to non- and importing firms, and they study the cross-financing to get over severe financ- housing loans, or the propensity of house- relationship between firms’ intensive and ing constraints that are often prevalent in holds to invest in housing improvements. extensive margins of trade. They also com- emerging market economies, and the ram- Gan, Guo, and Xu use a unique pare trade patterns across firms of different pant expropriation of minority sharehold- hand-collected nationwide survey to organizational structure, distinguishing ers under the weak corporate governance study China’s privatization, the largest in between domestic private firms, domestic environment. The authors find that, from human history. They find that privatiza- state-owned firms, foreign-owned firms, the perspective of the collection of firms tion in China has improved performance, and joint ventures. Exploring the varia- affected by the internal capital market, the but only for firms bought out by managers tion in foreign ownership across sectors, market is the least inefficient when weak (MBOs). Consistent with improved per- they find results consistent with recent corporate governance induces more tun- formance, MBO firms are less likely to be theoretical and empirical work on the neling activities and there is no big need influenced by the state in their daily oper- role of credit constraints and contractual to mitigate financing constraints. On the ation and are more likely to take various imperfections in international trade and other hand, when corporate governance is restructuring measures. The authors also investment. relatively stronger, and firms have a press- find that city governments with stronger Han argues that economic liberaliza- ing need to use the internal capital mar- fiscal discipline, and with fewer politi- tion, by reducing the extent to which an

42 NBER Reporter • 2008 Number 4 autocrat can directly control economic bership increased more among educated in the growing private sector. Such a strat- resources, induces democratization. This individuals with greater private-sector egy, of co-opting a new economic elite, paper suggests that in post-reform China, opportunities. Exploiting exogenous vari- could help to increase the Party’s survival the composition of the ruling Communist ation in college graduates’ labor market probability and strengthen its commit- Party membership was altered so as to keep outside options reveals that such a change ment to economic reforms. political and economic control aligned. is driven mainly by the Party’s increased National survey data show that mem- demand for educated individuals working

Entrepreneurship Working Group The NBER’s Working Group on Larry Chavis, University of North University Entrepreneurship met in Cambridge on Carolina at Chapel Hill; and Leora October 3. Group Director Josh Lerner, Klapper and Inessa Love, The World Ola Bengtsson, Cornell University, and of NBER and Harvard University, orga- Bank, “Entrepreneurial Finance around Berk A. Sensoy, University of Southern nized the meeting. These papers were the World: The Impact of the Business California, “Investor Abilities and discussed: Environment on Financing Constraints” Financial Contracting: Evidence from Venture Capital” Luis Cabral, New York University, and James E. Rauch, University of Zhu Wang, Federal Reserve Bank of California, San Diego and NBER, Yael V. Hochberg, Northwestern Kansas City, “Spin-Outs: Theory and “Spinout Entrepreneurship, Crony University; Alexander Ljungqvist, New Evidence” Capitalism, and Development” York University; and Annette Vissing- Discussant: Chris Woodruff, University Jorgenson, Northwestern University Mark Doms, Federal Reserve Bank of of California, San Diego and NBER, “Informational Hold-Up San Francisco; Ethan Lewis, Dartmouth and Performance Persistence in Private College; and Alicia Robb, University Jerry G. Thursby, Georgia Institute of Equity” of California, Santa Cruz, “Local Labor Technology, and Marie C. Thursby, Discussant Morten Sorensen, Columbia Market Endowments, New Business Georgia Institute of Technology and University and NBER Characteristics, and Performance” NBER, “Faculty Participation in Discussant: Bart Hamilton, Washington Licensing: Implications for Research” University Discussant: Lee Fleming, Harvard

Cabral and Wang develop a passive not to compete” laws. view, the authors also find that, condi- learning model of firm entry by spin-off:a It is often asserted that a highly edu- tional on the owner’s education, higher firm’s employees leave their employer and cated workforce is vital to improving the education levels in the local market are create a new firm when they learn that they competitive position of American busi- positively correlated with improved busi- are good entrepreneurs (type I spin-offs), nesses, especially by boosting entrepre- ness outcomes. or they learn that their employer’s pros- neurship. Doms, Lewis, and Robb eval- Chavis, Klapper and Love use a pects are bad (type II spin-offs). Here, the uate this contention using population unique database that contains over 70,000 theory predicts a high correlation between Census data and a nationally representa- firms in over 100 countries to system- spin-offs and parent exit, especially when tive panel of startup firms. They examine atically study the use of different financ- the parent is a low-productivity firm. This how the education and skill level of the ing sources for new and young firms. As correlation may correspond to two types local labor force are related to the creation expected, they confirm that, in all coun- of causality: either spin-off causes firm exit and success of new businesses. They find tries, younger firms rely less on bank (type I spin-offs) or firm exit causes spin- that areas with more skilled labor also have financing and more on informal financ- off (type II spin-offs). The authors test higher rates of self-employment and more ing. However, they also find that younger and confirm this and other predictions of skilled entrepreneurs, and that the edu- firms have better access to bank finance in the model on a unique dataset of the U.S. cation of the business owner is strongly countries with better rule of law and bet- automobile industry. Finally, they discuss linked to improved business outcomes. ter credit information, and that the reli- policy implications regarding “covenant Potentially consistent with the popular ance of young firms on informal finance

NBER Reporter • 2008 Number 4 43 decreases with the availability of credit Thursby and Thursby exploit a to their portfolio companies. The analysis information. Overall, these results suggest unique database on disclosure of research uses a new database of contractual pro- that improving the legal environment and and invention by faculty at eleven major visions for investments by 646 private- availability of credit information would U.S. universities over a period of 17 years partnership VCs in 1,266 startup compa- be disproportionately beneficial for pro- to explore whether university licensing nies over 1,534 investment rounds. The moting access to formal finance by young has compromised basic research. They authors discover that more experienced firms. study the relationships among disclo- VCs, who likely have better monitoring Recently collected data show that, sures to industry and federally sponsored abilities and whose withdrawal as active, within any manufacturing industry, verti- research, publications, citations, “expected value-added investors is more costly to cally integrated firms tend to have larger, citations,” and basic publications. They entrepreneurs, are less likely to use con- higher productivity plants, to account for find that recent disclosure activity has tracts that give them greater cash flow the bulk of sales, and to sell externally a positive effect on research funding by rights if company performance is poor. most of the inputs that they produce. In a industry and the federal government. But, This result survives a battery of controls weak contracting environment that char- if faculty disclose their results multiple for company characteristics, including acterizes less developed countries (LDCs), times, then the positive effect on federal valuation, as well as specifications that vertically integrated firms are vulnerable funding can disappear and become nega- control for endogenous selection effects. to “spinouts” by employees who make spe- tive. Both recent and repeated disclosures The relation between VC experience and cialized inputs (formerly provided inter- increase the faculty member’s publication downside protections is weaker when nally) subject to hold-up and who capture count, as well as the importance of these agency problems are less severe. the profits formerly made from external publications in terms of citation. There is Hochberg, Ljungqvist, and Vissing- sales of generic inputs. Rauch shows that also weak evidence that disclosure activ- Jorgenson first propose and then test this vulnerability leads to inefficiently ity is associated with increases in other a theory of learning and informational low entry and helps to explain the “miss- measures of “basic” research. Finally, the hold-up in the VC market. Their model ing middle” in the size distribution of authors examine life-cycle effects and find predicts that higher returns on the cur- LDC firms and the limited local con- that the ability to attract funding and the rent fund increase: the probability that tent of LDC exports. Vertically integrated rate of publication increase as the faculty a VC will raise a follow-on fund, the size firms can fight back by hiring “cronies” to member ages, but at a decreasing rate. of the follow-on fund, and the perfor- manage their input divisions: members of Research also tends to be less basic as fac- mance fee investors are charged in the networks that informally sanction hold- ulty members age. The authors further follow-on fund. If learning is asymmet- ups, or children who keep profits “in the find that post tenure, both types of fund- ric, such that incumbent investors learn family” even if they spin out. This paper ing decrease and work becomes less basic. more about fund managers’ skill than predicts the association of co-ethnic net- Bengtsson and Sensoy ask how potential new investors do, the model works with high rates of entrepreneurship investors’ abilities to mitigate agency also predicts persistence in returns, poor and the prominence of family-owned busi- problems in non-contractual ways will performance among first-time funds, per- ness groups in LDC manufacturing. The affect contract design. Their empirical sistence in investors from fund to fund, government can achieve a higher level of setting is the venture capital (VC) indus- and over-subscription in follow-on funds entry at minimum cost by directing sub- try, in which there are substantial agency raised by successful fund managers. The sidies to vertically integrated family firms, problems, considerable flexibility in con- empirical evidence is consistent with but only if families with competent chil- tract design, and wide variation in the these predictions. dren can be identified ex ante. abilities of VCs to monitor and add value

44 NBER Reporter • 2008 Number 4 Economic Fluctuations and Growth NBER’s Program on Economic and NBER and Ivan Werning, MIT and Javier Miranda, U.S. Census Fluctuations and Growth met on and NBER, “The Political Economy of Bureau, “Business Volatility, Job October 17 at the Federal Reserve Bank Nonlinear Capital Taxation” Destruction, and Unemployment” of Chicago. Organizers Janice C. Eberly, Critic: Narayana R. Kocherlakota, Critic: Ruediger Bachmann, University NBER and Northwestern University, and University of Minnesota and NBER of Michigan Giuseppe Moscarini, NBER and Yale University, chose the following papers to Charles I. Jones, University of Arvind Krishnamurthy and Annette discuss: California, Berkeley and NBER, Vissing-Jorgensen, Northwestern “Intermediate Goods and Weak Links: University and NBER, “The Aggregate Robert B. Barsky, University of A Theory of Economic Development” Demand for Treasury Debt” Michigan and NBER, and Eric Sims, Critic: Richard Rogerson, Arizona State Critic: Henning Bohn, University of University of Michigan, “Information, University and NBER California, Santa Barbara Animal Spirits, and the Meaning of Innovations in Consumer Confidence” Steven J. Davis, University of Chicago Michael Woodford, Columbia Critic: John H. Cochrane, University of and NBER; R. Jason Faberman, University and NBER, “Information- Chicago and NBER Federal Reserve Bank of Philadelphia; Constrained State-Dependent Pricing” John Haltiwanger, University of Critic: Ariel Burstein, University of Emmanuel Farhi, Harvard University Maryland and NBER; and Ron Jarmin California, Los Angeles and NBER

Innovations to measures of consumer ence? In his paper, Jones returns to two ment. According to these models, a lower confidence convey incremental informa- old ideas in development economics; he intensity of idiosyncratic shocks produces tion about economic activity far into the proposes that linkages and complementa- less job destruction, fewer workers flow- future. Barsky and Sims compare the shape rity are at the heart of the explanation. ing through the unemployment pool, and of impulse responses to confidence inno- First, linkages between firms through less frictional unemployment. To evalu- vations in the data with the predictions of intermediate goods deliver a multiplier ate the importance of this theoretical a calibrated New Keynesian model. They similar to the one associated with capi- mechanism, the authors relate industry- find little evidence of a strong causal chan- tal accumulation in a neoclassical growth level unemployment flows from 1977 to nel from autonomous movements in sen- model. Because the intermediate goods’ 2005 to industry-level indicators for the timent to economic outcomes (the “ani- share of revenue is about one half, this intensity of idiosyncratic shocks. They mal spirits” interpretation). Rather, these multiplier is substantial. Second, just as find strong evidence that declines in the impulse responses support an alternative a chain is only as strong as its weakest intensity of idiosyncratic labor demand hypothesis: that the surprise movements link, problems at any point in a produc- shocks drove big declines in the incidence in confidence reflect information about tion chain can reduce output substan- and rate of unemployment. This evidence future economic prospects (the “informa- tially if inputs enter production in a com- implies that the unemployment rate has tion” view). The authors conclude that plementary fashion. Jones builds a model become much less sensitive to cyclical confidence innovations are best charac- to quantify these forces and shows that it movements in the job-finding rate. terized as noisy measures of changes in can easily generate 50-fold aggregate dif- Krishnamurthy and Vissing- expected productivity growth over a rela- ferences in income. Jorgensen show that the U.S. debt-to- tively long horizon. Unemployment inflows fell from GDP ratio is negatively correlated with Farhi and Werning study efficient 4 percent of employment per month the spread between corporate bond yields nonlinear taxation of labor and capital in in the early 1980s to 2 percent or less and Treasury bond yields. This result holds a dynamic Mirrleesian model incorporat- by the mid-1990s and thereafter. U.S. even after they control for the default risk ing political economy constraints. Their data also show a secular decline in the on corporate bonds. The authors argue main result is that the marginal tax on job destruction rate and the volatility that the corporate bond spread reflects a capital income is progressive, in the sense of firm-level employment growth rates. convenience yield that investors attribute that richer agents face higher marginal Davis, Faberman, Haltiwanger, Jarmin, to Treasury debt. They show that regu- tax rates. Per capita income in the richest and Miranda interpret this decline as a latory demanders of Treasuries, includ- countries of the world exceeds that in the decrease in the intensity of idiosyncratic ing foreign central banks, have inelastic poorest countries by a factor of more than labor demand shocks, a key parameter in demand curves, and they estimate the 50. What explains this enormous differ- search and matching models of unemploy- effect that these buyers have on Treasury

NBER Reporter • 2008 Number 4 45 yields. They also discuss the implications He endogenizes the imperfect infor- mediate levels of the information cost, the for the aggregate value of Treasury con- mation using a variant of the theory of model is equivalent to one proposed by venience, the financing of the U.S. deficit, “rational inattention” proposed by Sims. Caballero and Engel, but it provides an the behavior of interest rate swap spreads, This results in a one-parameter family of economic motivation for the hazard func- and investors’ portfolio choices. models, indexed by the cost of informa- tion and very specific predictions about Woodford generalizes the standard tion, which nests both the standard state- its form. For moderate levels of the infor- (full-information) model of state-depen- dependent pricing model and the Calvo mation cost, the Calvo model of price-set- dent pricing in which decisions about model of price adjustment as limiting ting is a fairly accurate approximation to when to review a firm’s existing price cases (corresponding to an information the exact equilibrium dynamics, except in must be made on the basis of imprecise cost of zero and an unboundedly large the case of (infrequent) large shocks. awareness of current market conditions. information cost, respectively). For inter-

Market Microstructure

The NBER’s Working Group on Reserve Board, “Strategic Cross-Trading Discussant: Ronnie Sadka, Boston Market Microstructure met on October in the U.S. Stock Market” College 24 in Cambridge. Group Director Bruce Discussant: Bruce Mizrach, Rutgers Lehmann of University of California, University Anna Obizhaeva, University of San Diego, Eugene Kandel, Hebrew Maryland, “Price Impact and Spread: University, Jerusalem, and Avanidhar Chiraphol Chiyachantana, Singapore Application of Bias-Free Estimation Subrahmanyam, University of California, Management University, and Pankaj Methodology to Portfolio Transitions” Los Angeles, jointly organized the meet- Jain, University of Memphis, “The Discussant: Charles Jones, Columbia ing. These papers were discussed: Opportunity Cost of Inaction in University Financial Markets: An Analysis of Ioanid Rosu, University of Chicago, Institutional Decisions and Trades” Alex Boulatov and Thomas George, “Liquidity and Information in Order Discussant: Sunil Wahals, Arizona State University of Houston, “Securities Driven Markets” University Trading when Liquidity Providers are Discussant: Uday Rajan, University of Informed” Michigan Azi Ben-Rephael and Avi Wohl, Tel Discussant: Ronald Goettler, University Aviv University, and Ohad Kadan, of Chicago Paolo Pasquariello, University of Washington University in St. Louis, Michigan, and Clara Vega, Federal “The Diminishing Liquidity Premium”

order can be seen as an unambiguously percentage of informed traders surpris- Using a dynamic model of an order- positive signal; by contrast, a limit buy ingly generate smaller spreads. driven market, Rosu analyzes the inter- order is typically a weaker positive, and in Pasquariello and Vega provide a the- action between liquidity traders and some cases even a negative, signal. Rosu’s ory and new empirical evidence of cross- informed traders. Agents choose freely model generates a rich set of relationships price impact — that is, the permanent between limit and market orders by trad- among prices, spreads, trading activity, impact of informed trades in one asset ing off execution price and waiting costs. and volatility. In particular, the order flow on the prices of other (either related or In equilibrium, informed and patient trad- is autocorrelated if and only if there are fundamentally unrelated) assets — in the ers generally submit limit orders, except informed traders in the market, and the U.S. stock market. They develop a model when the fundamental value of the asset autocorrelation increases with the per- of multi-asset trading in the presence of that they privately observe is far from the centage of informed traders. Higher vola- two realistic market frictions: information current market-inferred value, in which tility and lower trading activity generate heterogeneity and imperfect competition case they become impatient and submit larger spreads while, after controlling for among informed traders. In that setting, a market order. As a result, a market buy volatility and trading activity, a higher they show cross-price impact to be the

46 NBER Reporter • 2008 Number 4 equilibrium outcome of strategic trading and Wohl demonstrate that both the potential explanations that might under- activity among risk-neutral speculators sensitivity of returns to liquidity and the lie these patterns. across many assets in order to mask their liquidity premium have declined signifi- Boulatov and George study securi- information advantage about some other cantly over the past four decades to levels ties trading when informed agents pro- assets. The authors find strong, robust evi- that cannot be statistically distinguished vide liquidity, either because dealers have dence of cross-asset informational effects from zero. Furthermore, the profitabil- superior access to market information or in a comprehensive sample of the trading ity of trading strategies based on buying because informed traders exploit strate- activity in NYSE and NASDAQ stocks illiquid stocks and selling liquid stocks gies involving limit orders. The authors between 1993 and 2004. has declined significantly over the past show that both informed dealers and Chiyachantana and Jain present four decades. The authors offer possible informed traders can profit more at the the first comprehensive analysis of a fre- explanations for these results related to expense of uninformed liquidity trad- quently ignored component of imple- the proliferation of hedge funds, index ers when markets are transparent than mentation shortfall: the opportunity cost funds, and exchange-traded funds. when markets are opaque. This is because of institutional decisions that are not Obizhaeva develops a bias-free transparency serves as a coordination executed. Of total decisions made, over methodology for estimating the param- device that reduces competition among 8 percent are partly or wholly unfilled. eters of trading costs and applies it to a informed traders. When the informed are The opportunity costs of this failure to unique dataset of portfolio transition allowed to choose whether to trade via trade are 24 basis points, or $20 billion trades. She estimates price impact and market orders or price-contingent (limit) in the sample period, much higher than effective spread in both traditional mar- orders, informed traders will gravitate the price impact and five times that of kets and crossing networks for the period toward limit orders. The endogenous commissions. There is a significant asym- 2001 through 2005. These estimates vary allocation of traders to order types maxi- metry in the opportunity cost of buy- across securities: price impact increases mizes competition among the informed, versus-sell decisions based on whether with the stocks’ overall trading volume thereby minimizing expected losses to market conditions are bullish or bearish. and volatility; in contrast, effective spread liquidity traders. The predictions here are Opportunity costs decrease with firm decreases with these characteristics. For consistent with recent empirical evidence: size, speed of transaction, number of thinly traded securities, trading costs are that the price impact of limit-order arriv- brokers, and exchange listing, whereas almost invariant with respect to trade size, als is greater than that of market-order they increase with market volatility. and spread-related payments account for executions. This suggests that the usual Opportunity loss from non-execution their largest fraction. For actively traded approach in empirical microstructure persists and increases slightly with the securities, trading costs are very sensitive research, of measuring the impact of pri- passage of time. Transaction-cost risks to trade size, and spread-related payments vate information as the trade-correlated are higher for small stocks and volatile are less significant. The positive associa- permanent component of price changes, stocks. tion between price impact and trading significantly understates the true impor- Previous evidence suggests that less volume is counterintuitive and not easily tance of private information securities liquid stocks yield higher average returns. explained within existing market micro- markets. Using NYSE data, Ben-Rephael, Kadan, structure models. Obizhaeva outlines

NBER Reporter • 2008 Number 4 47 Labor Studies NBER’s Program on Labor Studies and Carmit Segal, Universitat Pompeu Working Paper No. 14308) met in Cambridge on October 31. Fabra, “Does Temporary Affirmative Program Director Richard B. Freeman Action Produce Persistent Effects? A Douglas Almond, Columbia University and NBER Research Associate Lawrence Study of Black and Female Employment and NBER; Joseph J. Doyle, MIT and F. Katz, both of Harvard University, orga- in Law Enforcement” NBER; Amanda Kowalski, MIT; and nized the meeting and chose the follow- Heidi Williams, Harvard University, ing papers to discuss: William Kerr, Harvard University, “Estimating the Marginal Returns to and William Lincoln, University Medical Care: Evidence from At-Risk Jesse Rothstein, Princeton University of Michigan, “The Supply Side of Newborns” and NBER; Stephanie Cellinis, George Innovation: H-1B Visa Reforms and Washington University; and Fernando U.S. Ethnic Invention” Janet Currie, Columbia University and Ferreira, University of Pennsylvania and NBER; Mark Stabile, University of NBER, “The Value of School Facilities: Morris M. Kleiner, University of Toronto; and Phongsack Manivong Evidence from a Dynamic Regression Minnesota and NBER, and Alan B. and Leslie L. Roos, University of Discontinuity Design.” Krueger, Princeton University and Manitoba: “Child Health and Young NBER, “The Prevalence and Effects Adult Outcomes” Amalia Miller, University of Virginia, of Occupational Licensing”(NBER

Rothstein, Cellinis, and Ferreira spanning a period of 33 years, they show license from a government agency; that is a draw on the unique characteristics of that AA plans increase black employ- higher percentage than was found in stud- California’s system of school finance, ment for all ranks of police, averaging ies that rely on state-level data. Workers comparing districts in which school bond between 4.2 and 6.5 percentage points who have higher levels of education are referenda passed or failed by narrow mar- over and above any prevailing trends in more likely to work in jobs that require a gins, to estimate the impact of invest- the country. They find no erosion of black license. Union workers and government ments in school facilities. They find that employment gains from AA in the decade employees are more likely to have a license passing a referendum causes immediate, and a half following AA termination. requirement than nonunion or private sizable increases in home prices, imply- Nevertheless, in departments whose plans sector employees. Estimates here suggest ing a willingness on the part of marginal are terminated, they find a significant that licensing has about the same quanti- homebuyers to pay $1.50 or more for decrease in black employment growth tative impact on wages as unions — that each dollar per pupil of facility spending. relative to departments whose plans con- is about 15 percent — but unlike unions, These effects do not appear to be driven tinue. In contrast to their findings for which reduce variance in wages, licensing by changes in the income or racial compo- blacks, the authors find only marginal does not significantly reduce wage disper- sition of homeowners. While the authors employment gains for women and none sion for individuals in licensed jobs. find suggestive evidence that bond pas- at higher ranks. Almond, Doyle, Kowalski, and sage leads to increases in student test Kerr and Lincoln exploit large Williams estimate the marginal returns scores, this effect cannot explain more changes in the H-1B visa program. They to medical care for at-risk newborns by than a small portion of the housing price find that fluctuations in levels of H-1B comparing health outcomes and provi- effect, indicating that bond passage leads admissions significantly influence the rate sion of medical treatment on either side of to improvements in other dimensions of of Indian and Chinese patenting in cities common risk classifications, most notably school output (for example, safety) that and firms dependent upon the program. the “very low birth weight” threshold of may be not captured by test scores. They also find weak crowding-in effects, 1500 grams. First, using data on the cen- Miller and Segal exploit the rich such that total invention increases with sus of U.S. births in available years from variation in timing and outcomes of higher admission levels. 1983–2002, they find that newborns with 140 employment discrimination law- Using data from a specially designed birth weights just below that threshold suits brought against U.S. law enforce- Gallup survey, Kleiner and Krueger pro- have lower one-year mortality rates than ment agencies to estimate the cumulative vide the first nation-wide analysis of the newborns with birth weights just above employment effects of temporary, exter- labor market implications of occupational this cutoff — even though mortality risk nally imposed affirmative action (AA). licensing for the United States. They find in general tends to decrease with birth Using confidential administrative data on that in 2006, 29 percent of the workforce weight. Second, using hospital discharge 479 of the largest state and local agencies was required to hold an occupational records for births in five states in available

48 NBER Reporter • 2008 Number 4 years from 1991–2006, the researchers expected newborn deaths by one— for and 1987 in the Canadian province of find that newborns with birth weights just newborns with birth weight near 1500 Manitoba. These children were followed below 1500 grams have discontinuously grams, is on the order of $500,000 to until 2006, and their records were linked higher charges and frequencies of specific $650,000 in 2006 dollars. to provincial registries with data on health medical inputs. They estimate a $4,000 Previous research has shown a strong outcomes. The authors compare children increase in hospital costs as birth weight connection between birth weight and a with health conditions to their own sib- approaches 1500 grams from above, rela- child’s future health outcome. To address lings born at most nine years apart, and tive to mean hospital costs of $40,000 just some important questions not considered they control for health at birth. They find above 1500 grams. Taken together, these in that work, Currie, Stabile, Manivong, that health problems in early childhood estimates suggest that the cost of sav- and Roos use a unique dataset based are significant determinants of outcomes ing one “newborn statistical life” — that on public health insurance records for linked to adult socioeconomic status. is, the cost of reducing the number of 50,000 children born between 1979

Political Economy NBER’s Program on Political and Edward Miguel, University of Countries” (NBER Working Paper No. Economy met in Cambridge on October California , Berkeley and NBER, 14316) 31. NBER Research Associate Romain “Government Transfers and Political Discussant: Benjamin Olken, MIT and Wacziarg of University of California, Los Support” NBER Angeles organized the meeting. These Discussant: Rohini Pande, Harvard papers were discussed: University Nathan Nunn, Harvard University and NBER, and Leonard Wantchekon, Pedro Dal Bó, Brown University Pol Antràs, Harvard University and New York University: “The Trans- and NBER, and Andrew Foster NBER, and Gerard Padró i Miquel, Atlantic Slave Trade and the Origins of and Louis Putterman, Brown London School of Economics and Mistrust within Africa” University, “Institutions and Behavior: NBER, “Foreign Influence and Welfare” Discussant: William Easterly, New York Experimental Evidence on the Effects of (NBER Working Paper No. 14129) University and NBER Democracy” Discussant: Arnaud Costinot, MIT and Discussant: Alan Gerber, Yale University NBER Adam Meirowitz, Princeton University, and NBER and Kenneth Shotts, Stanford Alberto Alesina, Harvard University University, “Pivots Versus Signals in Verónica Amarante and Andrea and NBER, and Ekaterina Elections” Vigorito, Universidad de la Republica, Zhuravskaya, New Economic School, Discussant: Steven Callander, Montevideo; Marco Manacorda, “Segregation and the Quality of Northwestern University London School of Economics; Government in a Cross-Section of

Bo, Foster, and Putterman describe affect behavior, in addition to having are 25 to 31 percentage points more likely a novel experiment on the effect of a pol- an effect through the choice of poli- to favor the current government (rela- icy designed to encourage cooperation in cies. More generally, their findings have tive to the previous government). The a prisoner’s dilemma game. The effect of implications for empirical studies of treat- program’s impacts on political support this policy on the level of cooperation is ment effects in other contexts: the effect are larger among poorer households and greater when it is chosen democratically of a treatment can differ depending on for those near the center of the political by the subjects rather than being imposed whether it is endogenous or exogenous. spectrum, as predicted by the probabi- exogenously. In contrast to earlier stud- Amarante, Vigorito, Manacorda, listic voting model in political economy. ies, their experimental design allows them and Miguel estimate the impact of a large The researchers estimate that the annual to control for selection effects (for exam- anti-poverty program — the Uruguayan cost to the government of increasing their ple, those who choose the policy may be PANES — on political support for the political support by 1 percent is on the affected differently by it). Their results government that puts it in place. They order of US$89 million, or 1.7 percent of imply that democratic institutions may find that program-beneficiary households annual government expenditures.

NBER Reporter • 2008 Number 4 49 Antras and Miquel develop a model same country. They also attempt to cor- the respondent’s current distance from of foreign influence and apply it to the relate measures of segregation with mea- the coast. They show that the relationship study of optimal tariffs. They show that sures of quality of the polity and policy- between the slave trade and an individu- policies end up maximizing a weighted making, and they construct an instrument al’s level of trust cannot be explained by sum of domestic and foreign welfare; for- to overcome the endogeneity problem the slave trade’s effect on factors external eign influence may enhance welfare, from (that arises because groups move within to the individual, such as domestic insti- the point of view of aggregate world wel- country borders, partly in response to tutions, or the legal environment. Instead, fare, because it helps to alleviate externali- policies). Their results suggest that more the evidence shows that the effects of the ties arising from the cross-border effects ethnically and linguistically segregated slave trade work primarily through verti- of policies. However, foreign influence countries, that is, those where groups live cally transmitted factors that are internal can prove harmful in the presence of large more spatially separately, have a substan- to the individual, such as cultural norms imbalances in power across countries. The tially lower quality of government. In con- of behavior. researchers apply their model of foreign trast, they find no relationship between Meirowitz and Shotts consider a influence to the study of optimal trade religious segregation and the quality of two-period model of elections in which policy, deriving a modified formula for government. voters have private information about the optimal import tariff. They show that Nunn and Wantchekon investigate their policy preferences. A first-period a country’s import tariff is more distorted the historical origins of mistrust within vote can have two types of consequences: when the influenced country is small rela- Africa. Combining contemporary house- it may be pivotal in the first election, tive to the influencing country, and when hold survey data with historical data on and it provides a signal that affects can- natural trade barriers between the two slave shipments by ethnic group, they didates’ positions in the second election. countries are small. show that individuals whose ancestors Pivot events are exceedingly unlikely, but Alesina and Zhuravskaya produce were heavily threatened by the slave trade when they occur the effect of a single a new compilation of data on ethnic, lin- exhibit less trust in others and less trust vote is enormous. In contrast, vote totals guistic, and religious composition at the in the government today. The research- always have some signaling effect, but sub-national level for a large number of ers confirm that this relationship is causal the effect of a single vote is small. The countries. They then use the data to mea- by instrumenting the historic intensity of authors investigate which effect— pivot sure segregation of different ethnic, reli- the slave trade by the ancestor’s historic or signaling — drives equilibrium voting gious, and linguistic groups within the distance from the coast, controlling for behavior in large electorates.

50 NBER Reporter • 2008 Number 4 Public Economics

NBER’s Program on Public Working Paper No. 14301) University; and Jesse Rothstein, Economics met at Stanford University Discussant: B. Douglas Bernheim, Princeton University and NBER, “The on November 6–7. Organizers Raj Stanford University and NBER Value of School Facilities: Evidence from Chetty and Emmanuel Saez, both of the a Dynamic Regression Discontinuity University of California, Berkeley and Hanming Fang and Peter Arcidiacono, Design” NBER, chose the following papers to Duke University and NBER, and Discussant: Michael Lovenheim, discuss: Esteban Aucejo, Duke University, Stanford University “Does Affirmative Action Lead to Amy Finkelstein, MIT and NBER; Mismatch? A New Test and Evidence” Raj Chetty and Emmanuel Saez, Liran Einav, Stanford University and Discussant: Caroline M. Hoxby, “Information and Behavioral Response NBER; and Mark R. Cullen, Yale Stanford University and NBER to Taxation: Evidence from an University, “Estimating Welfare in Experiment with EITC Clients at H&R Insurance Markets Using Variation in David Albouy, University of Michigan Block” Prices” (NBER Working Paper No. and NBER, “Are Big Cities Really Bad Discussant: J. Karl Scholz, University of 14436) Places to Live? Improving Quality-of- Wisconsin and NBER Discussant: Dan Silverman, University Life Estimates Across Cities” (NBER of Michigan and NBER Working Paper No. 14472) Wojciech Kopczuk, and Roger H. Discussant: Patrick Kline, University of Gordon, University of California, San Henrik Kleven, London School of California, Berkeley Diego and NBER, “The Choice of the Economics, and Wojciech Kopczuk, Personal Income Tax Base” Columbia University and NBER, Fernando Ferreira, University of Discussant: Matt Weinzerl, Harvard “Transfer Program Complexity and Pennsylvania and NBER; Stephanie University the Take-up of Social Benefits” (NBER Riegg Cellini, George Washington

Einav, Finkelstein, and Cullen and may induce incomplete take up. The mismatch (in the sense that the intended show how standard consumer and pro- authors’ model assumes a government beneficiary of affirmative action admis- ducer theory can be used to estimate wel- interested in ensuring a minimum income sion policies are made worse off ) could fare in insurance markets with selection. level for as many deserving individuals occur only if selective universities possess Their key observation is that the same as possible. It characterizes optimal pro- private information about students’ post- price variation needed to identify the grams as when policymakers can choose enrollment treatment effects. This neces- demand curve also identifies how costs the rigor of screening (and associated sary condition for mismatch provides the vary as market participants endogenously complexity) along with a benefit level and basis for a new test. The authors propose respond to price. With estimates of both an eligibility criterion. Optimal programs an empirical methodology to test for pri- the demand and cost curves, welfare anal- that are not universal always feature a vate information in such a setting. They ysis is straightforward. The authors illus- high degree of complexity. Although the implement it using data from Campus trate their approach by applying it to government is interested only in ensur- Life and Learning Project (CLL) at Duke. the employee health insurance choices at ing a minimum benefit level, the opti- Preliminary evidence shows that Duke Alcoa, Inc. They detect adverse selection mal policy may feature benefits higher does possess private information that is a in this setting, but estimate that its quan- than this target minimum. This is because statistically significant predictor of a stu- titative welfare implications are small, and benefits generically screen better than dent’s post-enrollment academic perfor- not obviously remediable by standard either eligibility criteria or complexity. mance, but Duke’s private information public policy tools. The authors present numerical simula- only explains a very small percentage of Kleven and Kopczuk model com- tions with respect to budget size, ability the variation in student performance. The plexity in social programs as a byprod- distribution, complexity costs, and stigma authors also propose strategies to evaluate uct of efforts to screen between deserving and discuss their results in light of empiri- more conclusively whether the evidence and undeserving applicants. More rigor- cal findings for public programs in the of Duke private information has gener- ous screening technology may have desir- United States. ated mismatch. able effects on targeting efficiency, but the Arcidiacono, Aucejo, and Fang Hedonic estimates of quality of life associated complexity introduces trans- argue that once they take into account the across cities correspond to the cost-of- action costs into the application process rational enrollment decisions of students, living in a city relative to its local wage

NBER Reporter • 2008 Number 4 51 level. Albouy adjusts this standard model an elasticity of home prices with respect ers who report self-employment income. to account for federal taxes, non-hous- to school spending of 0.5. They also find However, there is increased bunching ing costs, and non-labor income, pro- suggestive evidence that passage causes near the peak even among wage earners, ducing quality-of-life estimates different higher test scores several years later, once suggesting that the information induced from those in the existing literature. This the bond-financed projects are complete. a real labor supply response. Third, for tax adjusted model produces city rankings These effects do not appear to be driven filers initially in the phase-out range, earn- that are positively correlated with those in by changes in the income and racial com- ings are essentially unaffected by the pro- the popular literature, and predicts how position of local households. Finally, the vision of information, perhaps because tax housing costs rise with wage levels, after magnitude of the price effect suggests professionals framed the work disincen- controlling for amenities. Mild seasons, that bond passage leads to improvements tive created by the EITC as being small. sunshine, and coastal location account in other dimensions of school output (for Overall, the changes in behavior induced for most quality-of-life differences; once example, safety) that may be not captured by information are substantial: EITC sub- these amenities are accounted for, quality by test scores. sidy rates would have to be increased by at of life does not depend on city size, con- Chetty and Saez test whether pro- least 20 percent ($10 billion) to generate trary to previous findings. viding information about the work incen- responses of the same size. Cellini, Ferreira, and Rothstein tives created by the Earned Income Tax Gordon and Kopczuk lay out the- use the housing market to estimate the Credit (EITC) amplifies its effects on oretically and estimate empirically how impact of investments in school facilities. labor supply. Conducting a random- best to use available information about Drawing on California’s unique system ized field experiment with 43,000 EITC each individual, in addition to earnings, of school finance, they compare districts claimants at H&R Block in which half when solving for the optimal tax base. In in which school bond referenda passed, the clients were provided simple, per- contrast to most of the literature on this or failed by narrow margins. The authors sonalized information about the EITC topic, their paper shows how equity con- account for the dynamic nature of bond schedule, they obtain three results. First, siderations may be incorporated quanti- referenda: the probability of future bond tax filers initially in the increasing and tatively in the analysis. They find that the proposals depends on the outcomes of past peak ranges of the EITC in the base year optimal tax base should include capital elections. They show first that bond funds are more likely to locate near the peak income, at least to some degree. In con- stick exclusively in the capital account, of the EITC schedule after receiving the trast to current practice, though, prop- with no effect on current expenditures information. Provision of information erty tax payments and mortgage inter- or other revenues. This suggests that the reduces the rate of extreme poverty (earn- est payments should not be deductible, effect of referendum passage reflects the ings below $7,000) by 15 percent and also because these deductions are costly on impact of improvements in the quality reduces the probability of moving into equity grounds, and presumably on effi- of school facilities. The authors find that the phase-out range. Second, the bunch- ciency grounds as well. passing a referendum causes immediate, ing around the peak caused by infor- sizable increases in home prices, implying mation provision is stronger for tax fil-

52 NBER Reporter • 2008 Number 4 International Finance and Macroeconomics The NBER’s Program on International View” University and NBER Finance and Macroeconomics met in Discussant: Gita Gopinath, Harvard Cambridge on November 7. Organizers University and NBER Hanno Lustig, University of California, Charles Engel, NBER and University of Los Angeles and NBER; Nick Wisconsin, and Linda Tesar, NBER and Lukasz A. Drozd, University of Roussanov, University of Pennsylvania; University of Michigan, chose the fol- Wisconsin, and Jaromir B. Nosal, and Adrien Verdelhan, Boston lowing papers to discuss: Columbia University, University, “Common Risk Factors in “Understanding International Prices: Currency Markets” Cristina Arellano, University Customers as Capital” Discussant: Craig Burnside, Duke of Minnesota, and Ananth Discussant: Paolo Pesenti, Federal University and NBER Ramanarayanan, Federal Reserve Reserve Bank of New York and NBER Bank of Dallas, “Default and Maturity Yan Bai, Arizona State University, and Structure in Foreign Bonds” Ariel Burstein, University of California, Jing Zhang, University of Michigan, Discussant: Olivier Jeanne, John Los Angeles and NBER, and Nir “Financial Integration and International Hopkins University Jaimovich, Stanford University and Risk Sharing” NBER, “Understanding Movements Discussant: Vivian Yue, New York Anton Korinek, University of in Aggregate and Product-Level Real University Maryland, “Regulating Capital Flows Exchange Rates” to Emerging Markets: An Externality Discussant: David Weinstein, Columbia

Arellano and Ramanarayanan study dent constraints. He demonstrates that of capital flows using historical data from the maturity composition and the term most forms of capital flows into such Indonesia, and describes a pecking order structure of interest rate spreads on gov- economies impose a macroeconomic of financial flows that reflects the differ- ernment debt in emerging markets. When externality, leading decentralized agents ent magnitudes of the resulting exter- interest rate spreads rise, debt maturity to take on too much systemic risk, and nalities. Furthermore, he defines a social shortens and the spread on short-term making the recipient country more vul- pricing kernel that describes the optimal bonds is higher than on long-term bonds. nerable to financial instability and crises. magnitude of policy measures to restore To account for this pattern, the authors Every capital inflow entails future out- social efficiency. build a dynamic model of international flows in the form of repayments, divi- Drozd and Nosal develop a new the- borrowing, with endogenous default and dends, or profit distributions. In states ory of pricing-to-market driven by slug- multiple maturities of debt. Short-term of the world when financing constraints gish market shares. Their key innovation debt can deliver higher immediate con- in an economy become binding, capi- is a capital theoretic model of marketing sumption than long-term debt; large long- tal outflows necessitate an increase in in which relations with the customers are term loans are not available because the the current account and a reduction in valuable. They discipline the introduced borrower cannot commit to save in the aggregate demand. This puts pressure on friction using a unique prediction of the near future towards repayment in the dis- the exchange rate and triggers a financial model about the low short-run and high tant future. However, issuing long-term accelerator mechanism, that is a mutual long-run price elasticity of international debt can insure against the need to roll- feed-back cycle of depreciating exchange trade flows, consistent with the data. The over short-term debt at high interest rate rates, deteriorating balance sheets, tight- model accounts for several pricing impli- spreads. The trade-off between these two ening financing constraints, and declining cations that are puzzling for a large class benefits is quantitatively important for aggregate demand. Decentralized agents of theories. The good performance on the understanding the maturity composition take prices as given and do not internal- quantities side is maintained. in emerging markets. When calibrated ize that the capital outflows associated Using wholesale price data for com- to data from Brazil, the model matches with their repayments contribute to the mon products sold in Canada and United the dynamics in the maturity of debt issu- financial accelerator. As a result, they do States over the period 2004–6 and infor- ances and its co-movement with the level not internalize the full social cost of such mation on the country of production of spreads across maturities. payments and take on too much sys- for individual products, Burstein and Korinek analyzes the external financ- temic risk in their financing decisions. Jaimovich document new facts on inter- ing decisions of emerging market econo- Korinek illustrates how these externalities national relative price movements. They mies that are prone to collateral-depen- can be quantified for different categories find that international relative prices

NBER Reporter • 2008 Number 4 53 at the level of individual products are est rate currencies are about 5 percent per tionless international financial markets, roughly three to four times as volatile as annum smaller than those on high interest while actual international financial mar- the Canada-U.S. nominal exchange rate rate currencies after accounting for trans- kets are far from frictionless. In particu- at quarterly frequencies. Aggregate real action costs. A single return-based factor, lar, financial contracts are incomplete and exchange rates, constructed by averag- the return on the highest minus the return enforceability of debt repayment is lim- ing movements in international relative on the lowest interest rate currency port- ited. Default risk of debt contracts con- prices for individual goods, closely fol- folios, explains the cross-sectional varia- strains borrowing, and more importantly, low the appreciation of the Canadian dol- tion in average currency excess returns it makes borrowing more difficult in bad lar over this period. The large movements from low- to high-interest-rate currencies. times, precisely when countries need insur- in international relative prices for traded Lustig, Roussanov, and Verdelhan show ance the most. Thus, default risk of debt goods are in conflict with the hypothesis that the high-minus-low currency return contracts hinders international risk shar- of relative purchasing power parity, but measures the component of the stochastic ing. When countries remove their official point instead to the practice of pricing- discount factor innovation that is com- capital controls, default risk is still present to-market by exporters. In light of these mon across countries. as an implicit barrier to capital flow; the findings, the authors construct a model of Conventional wisdom suggests that observed increase in capital flow under international trade and pricing to market financial liberalization can help countries financial liberalization is in fact too lim- that can account for the observed move- insure against idiosyncratic risk. There ited to improve risk sharing. If default risk ments in product- and aggregate real- is little evidence, however, that coun- of debt contracts were eliminated, capi- exchange rates for both traded and non- tries have increased risk sharing despite tal flow would be six times greater, and traded products. recent widespread financial liberalization. international risk sharing would increase Currency excess returns are highly According to Bai and Zhang, the key to substantially. predictable and strongly counter-cyclical. understanding this puzzling observation The average excess returns on low inter- is that conventional wisdom assumes fric-

Health Care

NBER’s Program on Health Care Taylor, Carnegie Mellon University, Adjusted Price Index for Colorectal met in Cambridge on November 12. “Unhealthy Insurance Markets: Search Cancer Drugs” Organizer Richard Frank, of Harvard Frictions and the Cost and Quality of Medical School and NBER, chose the Health Insurance” (NBER Working Amy Finkelstein and Daron following papers to discuss: Paper No. 14455) Acemoglu, MIT and NBER, and Matt Notowidigdo, MIT, “Income and Ming Tai-Seale, Texas A&M Jonathan Clark, Harvard University, Health Spending: Evidence from Oil University, and Tom McGuire, Harvard and Robert Huckman, Harvard Price Shocks” University, “Time is Up: The Increasing University and NBER, “Broadening Shadow Price of Time in Primary Care Focus: Complementarities and the Guy David and Seth Richards, Office Visits” Benefits of Specialization in the University of Pennsylvania, and Sara Hospital Industry” Markowitz, Emory University and R.D. Cebul and Mark Votruba, Case NBER, “The Effects of Pharmaceutical Western Reserve University; James Claudio Lucarelli, Cornell University, Marketing and Promotion on Adverse B. Rebitzer, Case Western Reserve and Sean Nicholson, Cornell Drug Events and Regulation” University and NBER; and Lowell University and NBER, “A Quality

A physician’s time is a precious Seale and McGuire formulate the physi- below a shadow price. Following the sec- resource in primary care, and the physi- cian’s decision problem and characterize ond “behavioral” rule, the physician ter- cian must constantly evaluate the gain two rules for deciding about when to end minates a visit when “time is up” and a tar- from spending more time with the cur- a visit. The first rule, which is labeled “effi- get number of minutes have expired. The rent patient versus moving on to address cient,” has the physician end a visit when authors test for the behavioral rule against the health care needs of the next one. Tai- the estimated value of more time falls the alternative using video recordings

54 NBER Reporter • 2008 Number 4 of 385 visits by elderly patients to their lar care, and the degree to which these equilibrium and the local general equi- primary care physician. The researchers returns depend on a hospital’s intensity librium effects of an increase in income structure the data at the “topic” level and of services that relate to cardiovascular on health expenditures. Their central esti- find evidence consistent with the behav- care. The researchers find suggestive evi- mate is an income elasticity of 0.7, with ioral rule. Specifically, time elapsed within dence of positive returns, on average, to an elasticity of 1.1 as the upper end of the a visit is a very strong determinant of focused experience in cardiovascular care. 95 percent confidence interval. Point esti- whether the physician decides this is the Moreover, these returns are contingent on mates from alternative specifications fall “last topic” to be discussed, thereby effec- the intensity with which hospitals provide on both sides of this central estimate, but tively ending the visit. The authors con- clinical services that are closely related to are almost always less than 1. Consistent sider whether dislodging a target-time cardiovascular care. with their finding that health spending mentality from physicians (and patients) The average price of providing a does not appear to be a luxury good, the might contribute to more productive pri- colorectal cancer patient with a 24-week authors do not find a significant effect of mary care practice. chemotherapy regimen increased from increased income on hospital technol- Health insurance is a complex prod- $127 in 1993 to $36,300 in 2005, largely ogy adoption; this suggests that there uct, which creates search frictions that because of the approval and widespread are unlikely to be substantial global gen- can distort market outcomes. Cebul, use of five new drugs between 1996 eral equilibrium effects (which would not Rebitzer, Taylor, and Votruba study and 2004. Lucarelli and Nicholson ask be estimated by their empirical strategy) the effect of frictions in the market for whether the substantial increase in spend- of rising income on health spending via employer-based health insurance. They ing has been worthwhile. They construct induced innovation. The overall reading find that frictions are most severe in the a price index for colorectal cancer drugs of the evidence is that rising income is “fully insured” part of the group health for each quarter between 1993 and 2005 unlikely to be a major driver of the rising insurance market. They estimate that fric- that takes into account the quality (that health share of GDP. tions in this market segment cause a quar- is, the efficacy and side effects, as reported David, Markowitz, and Richards ter of the consumer surplus to shift from in clinical trials) of each drug on the mar- analyze the relationship between post- policy-holders to insurers (a transfer of ket and the value that oncologists place marketing promotional activity and the $32.5 billion in 1997). Their analysis also on drug quality. They find that the naive reporting of adverse drug events by mod- suggests that frictions in insurance mar- price index, which does not adjust for eling the interaction between a welfare kets reduce incentives to invest in future the changing attributes of drugs on the maximizing regulator (the FDA) and a health. market, greatly overstates the true price profit maximizing firm. In their anal- The literature on related diversifica- increase. Both the hedonic price index ysis, demand is sensitive to both pro- tion suggests that multi-unit firms with and a quality-adjusted price index show motion and regulatory interventions. a portfolio of related businesses outper- that prices have actually remained fairly Promotion-driven market expansions form both single-unit firms and multi- constant over this 13-year period, with enhance profitability, but may involve unit firms composed of unrelated busi- slight increases or decreases depending on the risk that the drug would be prescribed nesses. Previous explanations for this the model’s assumptions. inappropriately, leading to adverse reg- relationship have centered on economies Health expenditures as a share of ulatory actions against the firm. This of scope achieved by sharing common GDP have more than tripled over the model exposes the effects of the current resources, such as advertising, or pro- last half century. A common conjec- regulatory system on consumer and pro- duction capacity. Clark and Huckman ture is that this is primarily a conse- ducer welfare. Particularly, the emphasis consider another potential explanation: quence of rising real per capita income, on safety over benefits distorts the mar- complementarities, or the extent to which which more than doubled over the same ket allocation of drugs away from some the marginal returns to the intensity of a period. Acemoglu, Finkelstein, and of the most appropriate users. Using an focal activity increase with the intensity Notowidigdo investigate this empiri- innovative combination of commercial of related activities. Using patient-level cally by using the time-series variation in data on pharmaceutical promotion and data from the hospital industry, they con- global oil prices between 1970 and 1990, FDA data on regulatory interventions sider the existence of complementarities interacted with cross-sectional variation and adverse drug reactions, the authors with respect to focused organizational in the oil reserves across different areas provide some evidence that increased lev- experience. Specifically, they investigate of the Southern United States, as instru- els of promotion and advertising lead to the extent to which there are returns ments for local area income. This strategy increased reporting of adverse medical to focused experience in cardiovascu- enables them to capture both the partial events for certain conditions.

NBER Reporter • 2008 Number 4 55 Higher Education The NBER’s Working Group on Duke University and NBER, “Does Bruce A. Weinberg, Ohio State Higher Education met in Cambridge Affirmative Action Lead to Mismatch? University and NBER, “Scientific on November 13. The group’s Director, A New Test and Evidence” Leadership” Charles T. Clotfelter of NBER and Duke (For discussion of this paper, see p. 51) University, chose these papers to discuss: Ginger Zhe Jin, University of Maryland Todd R. Stinebrickner, University of and NBER, and Alexander Whalley, Judith Scott-Clayton, Harvard Western Ontario and NBER, and Ralph University of California, Mercer, “The University, “On Money and Motivation: Stinebrickner, Berea College, “Learning Power of Attention: Do Rankings A Quasi-Experimental Analysis of about Academic Ability and the College Affect the Financial Resources of Public Financial Incentives for College Drop-out Decision” Colleges?” (NBER Working Paper No. Achievement” 12941) Thomas Dee, Swarthmore College and Esteban Aucejo, Duke University; and NBER, “Stereotype Threat and the Peter Arcidiacono and Hanming Fang, Student-Athlete”

Programs that link substantial mation play a central role in models of concentrated among males and students amounts of college financial aid to stu- decisionmaking, there is little empirical with below-median SAT scores. dent achievement have proved increas- evidence to inform these assumptions. At the beginning of the twentieth ingly popular in recent years. These pro- Stinebrickner and Stinebrickner use century, Europe as a whole and Germany grams could work either by relaxing unique data to examine how college stu- in particular dominated science. Today, financial constraints or by inducing addi- dents from low-income families update the United States does. Using rich data- tional student effort (or both). Scott- their beliefs about academic ability and sets on Nobel laureates in Chemistry, Clayton examines the PROMISE schol- to examine the role that learning about Medicine, and Physics, and highly cited arship in West Virginia, which provides ability, and a variety of other factors, play publications, Weinberg shows that free tuition and fees to college students in the college drop-out decision. investments in science and a competi- who maintain a minimum GPA and The academic underperformance of tive market for science in the United course load. Using an unusually com- student-athletes at selective colleges and States were the most important factors prehensive administrative database, she universities has motivated ongoing con- in the emergence of the United States as exploits discontinuities in both the eli- cern about the role of athletics in these a scientific leader. In contrast to previous gibility formula and the timing of imple- institutions. The growing literature on work, he finds here that scientists flee- mentation to identify program effects. the effects of social identity suggests one ing the Nazis played a relatively modest She finds robust and statistically signifi- possible explanation for this phenom- role. The evidence suggests that scientific cant effects on key academic outcomes, enon. Student-athletes may experience leadership is more fluid than generally including a 6.7 percentage point increase a reduction in cognitive performance in believed, highlighting the importance in four-year BA completion rates among evaluative settings like classrooms where of investments in science and a competi- PROMISE recipients. These impacts are they expect to be viewed through the lens tive environment— both as developing concentrated at the precise thresholds of a negative stereotype. Dee presents an countries seek to build cutting-edge sci- for annual scholarship renewal—partic- economic model of such a “stereotype entific communities, and as the United ularly the minimum course load require- threat” that reconciles prior evidence States and Europe seek to maintain their ment—and disappear in the fourth year on how student effort and performance positions. of college when students are still receiv- are influenced by this social-identity Jin and Whalley investigate whether ing the scholarship but no longer have phenomenon; he also presents empiri- and how college quality rankings affect the opportunity to renew. The findings cal evidence from a laboratory experi- a key factor in the ranks measure of suggest that the program works by estab- ment in which student-participants were quality — financial resources per stu- lishing clear academic goals and incen- randomly assigned to a treatment that dent — for public colleges. They show tives to meet them, rather than simply primed their awareness of an athletic that when a public college is exogenously reducing the cost of college. identity. This treatment reduced the test- included in the U.S. News & World Although assumptions about how score performance of athletes relative to Report rankings, educational and gen- agents update subjective beliefs in non-athletes by 14 percent (effect size eral expenditures per student increase response to the arrival of new infor- = -1.0). However, this effect was largely by 3.2 percent. To fund the additional

56 NBER Reporter • 2008 Number 4 expenditure, state appropriations per cate additional citizen attention to the lege guides. As the college quality beliefs student increase by 3.4 to 6.8 percent, issue of public college quality, and the of citizens are updated state govern- while tuition is not responsive at all. The increased attention steers more funding ments may adjust funding accordingly. state appropriation response may be real- towards public colleges. On the other The authors find evidence in support of ized in two potential channels: on the hand, college rankings may provide new the first explanation. one hand, U.S. News rankings may allo- information in addition to existing col-

Education Program Meets The NBER’s Program on Education Kirabo Jackson, Cornell University, Gauri Kartini Shastry, University of met in Cambridge on November 14. “Student Demographics, Teacher Virginia, “Human Capital Response Program Director Caroline M. Hoxby Sorting, and Teacher Quality: Evidence to Globalization: Education and of NBER and Stanford University orga- from the End of School Desegregation” Information Technology in India” nized the meeting. These papers were discussed: Jonah E. Rockoff, Columbia University Maria Fitzpatrick, Stanford University, and NBER, and Lesley Turner, “Preschoolers Enrolled and Mothers Richard J. Murnane, Harvard Columbia University, “Short-Run at Work? The Effects of Universal University and NBER, and Jennifer Impacts of Accountability on School Pre-Kindergarten” Steele and John Willett, Harvard Quality” University, “Do Financial Incentives Fernando B. Botelho and Ricardo Help Low-Performing Schools Attract Hanley Chiang, Mathematica, Inc., A. Madeira, University of Sao Paulo, and Keep Academically Talented “How Accountability Pressure on and Marcos A. Rangel, University Teachers? New Evidence from Failing Schools Affects Student of Chicago, “Discrimination Goes California” Achievement” to School? Understanding Racial Differences in Non-Blind Grading”

Murnane, Steele, and Willett cap- points. However, on average, GTF recip- ous measures of teacher quality (including italize on a natural experiment that ients left low performing schools at a estimated value-added). Jackson shows occurred in California between 2000 and higher rate in their first year of teaching that this was primarily because of a labor 2002: in those years, the state offered a than academically talented teachers with- supply response. competitively allocated $20,000 incen- out GTFs who chose to work in low-per- In November 2007, the New York tive, called the Governor’s Teaching forming schools. City Department of Education assigned Fellowship (GTF), aimed at attracting Jackson uses the reshuffling of - stu each elementary and middle school a let- academically talented, novice teachers dents caused by the end of student bus- ter grade (A to F) as part of a new account- to low-performing schools and retain- ing in Charlotte-Mecklenburg to investi- ability system. The grades were based on ing them in those schools for at least gate the relationship between changes in continuous numeric scores derived from four years. Taking advantage of data on student attributes and changes in teacher levels and changes in student achieve- the career histories of 14,045 individuals quality that are not confounded with ment and other school environmental fac- who pursued California teaching licenses changes in school or neighborhood char- tors such as attendance, and were linked between 1998 and 2003, the authors acteristics. His results suggest that the to a system of rewards and consequences are able to estimate the unbiased impact spatial correlation between teachers’ resi- for schools and principals. Rockoff and of the GTF on the decisions of recipi- dences, students’ residences, and schools Turner use the discontinuities in the ents to begin and continue working in could lead to spurious correlation between assignment of school grades to estimate low-performing schools. They find that student attributes and teacher character- the short-run impact of accountability. acquiring a GTF increased the probabil- istics. Re-shuffling the students led to Specifically, they examine student achieve- ity that recipients entered low-performing teacher resorting, so that schools that ment in English Language Arts and math- schools within two years after licensure experienced a repatriation of black stu- ematics (measured in January and March program enrollment by 34 percentage dents also experienced a decrease in vari- of 2008, respectively) using school-level

NBER Reporter • 2008 Number 4 57 aggregate data. Although schools had persistence of reading improvements is capital response, these districts experi- only a few months to respond to the less consistent. Second, Chiang analyzes enced smaller increases in skilled wage release of accountability grades, receipt the effects of sanction threats on various premiums. of a low grade significantly increased stu- features of educational production and Three states recently introduced dent achievement in both subjects, with finds that they increase school spending Universal Pre-Kindergarten programs larger effects in math. The authors find on instructional technology, curricular offering free preschool to all age-eligi- no evidence that accountability grades development, and teacher training. Both ble children; policymakers in many other were related to the percentage of students strands of evidence are consistent with a states are promoting similar programs. tested, implying that accountability sys- predominant role for educational reforms Using restricted-access data from the tems can cause real changes in school qual- in generating test score gains by threat- Census, together with year and birth- ity that increase student achievement over ened schools. day-based eligibility cutoffs, Fitzpatrick even a short time horizon. Parental evalu- Recent studies have shown that estimates the effects of Universal Pre-K ations of educational quality improved trade liberalization increases the skilled availability on overall preschool enroll- considerably for schools receiving low wage premium in developing countries. ment and maternal labor supply. She finds accountability grades; however, these Thus globalization may benefit elite that Universal Pre-K availability increases schools also experienced a larger increase skilled workers relatively more than poor statewide preschool enrollment by at least in survey response rates, holding open the unskilled workers, increasing inequality. 14 percent but has little effect on the labor possibility of selection bias in these com- However, that effect may be mitigated supply of most women. The exception plementary results. if human capital investment responds is women residing in rural areas, whose Although an emerging body of evi- to new global opportunities. One key probability of being employed increases dence has shown that the threat of sanc- question is whether a country with a by 20 percent. tions on low-performing schools can raise more elastic human capital supply is bet- Recent literature suggest that student test scores in the short run, the ter positioned to benefit from globaliza- observed racial differentials in labor mar- extent to which these test score improve- tion. Shastry examines how the impact kets are the result of lower investment ments are attributable to schools’ manip- of globalization varies across Indian dis- in the accumulation of skills, or of pre- ulation of the accountability system has tricts with different costs of skill acqui- market factors, by individuals of African remained uncertain. Chiang provides sition, focusing on the cost of learning descent. If parents and children update two new strands of evidence to evaluate English, a relevant qualification for high- investment decisions after extracting sig- the relative importance of educational skilled export jobs. Linguistic diversity in nals about scholastic abilities from school reforms and gaming behavior in generat- India compels individuals to learn either reports, then differential errors in -per ing test score gains by threatened schools. English or Hindi as a lingua franca. Some ceived ability could reinforce racial gaps First, using a methodology that exploits districts have lower relative costs of learn- in the accumulation of human capital. Florida’s system of imposing sanction ing English because of linguistic predis- Botelho, Madeira, and Rangel present threats on the basis of a cutoff level of per- positions and psychic costs associated evidence drawn from a unique dataset of formance, the author estimates medium- with past nationalistic pressure to adopt Brazilian elementary, middle, and high- run effects on student test scores from Hindi. Shastry demonstrates that dis- schools suggesting that teachers’ grad- having attended a threatened elementary tricts with a more elastic supply of English ing (when compared to blindly scored school. Threat-induced math improve- skills benefited more from globalization: tests of proficiency) suffers from cardi- ments from elementary school largely per- they experienced greater growth in both nal and ordinal biases associated with a sist at least through the first one to two information technology jobs and school child’s race . years of middle school, while evidence for enrollment. Consistent with this human

58 NBER Reporter • 2008 Number 4 Monetary Economics The NBER’s Program on Monetary Surico, Bank of England, “Monetary California, Los Angeles and NBER Economics met in Cambridge on Policies and Low-Frequency November 14. Organizers Ricardo Reis Manifestations of the Quantity Theory” George-Marios Angeletos, MIT and Michael Woodford of NBER and Discussant: James H. Stock, Harvard and NBER, and Jennifer La’O, MIT, Columbia University chose these papers University and NBER “Dispersed Information over the to discuss: Business Cycle: Optimal Fiscal and Olivier Coibion, College of William Monetary Policy” Mark Bils, University of Rochester & Mary, and Yuriy Gorodnichenko, Discussant: Robert King, Boston and NBER; Peter J. Klenow, Stanford University of California, Berkeley and University and NBER University and NBER; and Benjamin NBER, “What Can Survey Forecasts A. Malin, Federal Reserve Board, Tell Us about Informational Rigidities?” Alejandro Justiniano, Federal Reserve “Reset Price Inflation and the Impact of Discussant: Justin Wolfers, University of Bank of Chicago, and Giorgio E. Monetary Policy Shocks” Pennsylvania and NBER Primiceri, Northwestern University and Discussant: Jon Steinsson, Columbia NBER, “Potential and Natural Output” University and NBER Bernardo Guimaraes and Kevin D. Discussant: Christopher Erceg, Federal Sheedy, London School of Economics, Reserve Board Thomas J. Sargent, New York “Sales and Monetary Policy” University and NBER, and Paolo Discussant: Ariel Burstein, University of

A standard state-dependent pric- data from 1900–2005. They interpret a year. Importantly, the dynamic pro- ing model generates little monetary non- outcomes in terms of population values of cess followed by forecast errors following neutrality. Two ways of generating more those sums of coefficients implied by two structural shocks is consistent with the meaningful real effects are time-depen- dynamic stochastic general equilibrium predictions of models of informational dent pricing and strategic complementari- (DSGE) models. The DSGE models make rigidities. The authors interpret this find- ties. These mechanisms have telltale impli- the sums of coefficients depend on the ing as providing support for the recent cations for the persistence and volatility monetary policy rule via cross-equation expansion of research into models of of “reset price inflation,” which is the rate restrictions of a type that earlier authors, informational rigidities. In addition, they of change of all desired prices (including including Lucas and Sargent, emphasized document several stylized facts about the for goods that have not changed price in in the context of testing the natural unem- conditional responses of forecast errors the current period). Using the micro data ployment rate hypothesis. When the U.S. and disagreement among agents that can underpinning the CPI, Bils, Klenow, and data are extended beyond Lucas’s 1955– be used to differentiate between some Malin construct an empirical measure 75 period, the patterns revealed by scat- of the models of informational rigidities of reset price inflation. They find that ter plots mutate in ways that the authors recently proposed. time-dependent models imply unrealisti- want to attribute to prevailing monetary A striking fact about prices is the cally high persistence and stability of reset policy rules. prevalence of “sales” — large temporary price inflation. This discrepancy is exacer- Coibion and Gorodnichenko use price cuts followed by prices returning bated by adding strategic complementari- three different surveys of economic fore- exactly to their former levels. Guimaraes ties, even under state-dependent pricing. casts to assess both the support for and and Sheedy build a macroeconomic A state-dependent model with no strate- the properties of informational rigidities model with a rationale for sales based gic complementarities aligns most closely faced by agents. Specifically, they track on firms facing consumers with different with the data. the impulse responses of mean forecast price sensitivities. Even if firms can adjust To detect the quantity theory of errors and disagreement among agents sales without cost, monetary policy has money, Sargent and Surico look at scat- after exogenous structural shocks. Their large real effects owing to sales being stra- ter plots of filtered time series of infla- key contribution is to document that, in tegic substitutes: a firm’s incentive to have tion and money growth rates and interest response to structural shocks, mean fore- a sale is decreasing in the number of other rates and money growth rates. They relate casts fail to completely adjust on impact, firms having sales. Thus the flexibility seen those scatter plots to sums of two-sided leading to statistically and economically in individual prices attributable to sales distributed lag coefficients constructed significant deviations from the null of does not translate into flexibility of the from fixed-coefficient and time-varying full information: the half life of forecast aggregate price level. vector auto-regressions (VARs) for U.S. errors is roughly between six months and Angeletos and La’O study how the

NBER Reporter • 2008 Number 4 59 heterogeneity of information affects the gent taxes and monetary policy can boost potential output — that is, the level of efficiency of the business cycle and the learning over the business cycle; 4) typi- output that would prevail under perfect design of optimal fiscal and monetary cally, this implies that the optimal tax is competition; and natural output — that policy. They do so within a model that countercyclical, while the optimal mon- is, the level of output that would prevail features a standard Dixit-Stiglitz demand etary policy is less accommodative than with flexible prices and wages. They find structure, introduces dispersed private what is consistent with replicating the that potential output is smooth, which information about the underlying aggre- flexible-price equilibrium; and 5) even results in an output gap that closely gate productivity shock, and allows this if monetary policy were to replicate the resembles traditional measures of de- information to be imperfectly aggregated flexible-price equilibrium, this would not trended output. Meanwhile natural out- through certain prices and macroeco- mean targeting price stability. Rather, the put is extremely volatile, because of the nomic indicators. Their key findings are: optimal monetary policy has the nominal very high variability of markup shocks. 1) when information is exogenous to the interest rate increase, and the price level These disturbances, however, are very sim- agents’ actions, the response of the econ- fall, in response to a positive innovation ilar to measurement errors because they omy to either fundamentals or noise is in productivity. only explain price and wage inflation at efficient along the flexible-price equilib- Justiniano and Primiceri estimate a very high frequencies. Under this alterna- rium; 2) the endogeneity of learning ren- DSGE model with imperfectly competi- tive interpretation, potential and natural ders the business cycle inefficient: there is tive products and labor markets and sticky output move one-to-one. too little learning and too much noise in prices and wages. They use the model the business cycle; 3) both state-contin- to back out two counterfactual objects:

Asset Pricing NBER’s Program on Asset Pricing University of Pennsylvania and NBER, met on November 21 in Cambridge. “What’s Vol Got to Do With It?” Martijn Cremers and Antti Petajisto, Organizers Monika Piazzesi, Stanford Discussant: Lars P. Hansen, University Yale University, and Eric Zitzewitz, University and NBER, and Stijn Van of Chicago and NBER Dartmouth College, “Should Nieuwerburgh, New York University and Benchmark Indices Have Alpha? NBER, chose the following papers to Akub Jurek, Princeton University, Revisiting Performance Evaluation” discuss: “Crash-Neutral Currency Carry Trades” Discussant: Lubos Pastor, University of Discussant: Adrian Verdelhan, Boston Chicago and NBER Eric Swanson and Glenn Rudebusch, University Federal Reserve Bank of San Francisco, John Y. Campbell and Luis M. Viceira, “The Bond Premium in a DSGE Model Santiago Bazdrech and Frederico Belo, Harvard University and NBER, and with Long-Run Real and Nominal University of Minnesota, and Xiaoji Adi Sunderam, Harvard University, Risks” Lin, London School of Economics, “Inflation Bets or Deflation Hedges? Discussant: Stanley E. Zin, Carnegie “Labor Hiring, Investment, and Stock The Changing Risks of Nominal Bonds” Mellon University and NBER Return Predictability in the Cross Discussant: George Tauchen, Duke Section” University Itamar Drechsler, University of Discussant: Francois Gourio, Boston Pennsylvania, and Amir Yaron, University

The term premium on nomi- premium puzzle.” However, in models introducing Epstein-Zin preferences nal long-term bonds in the standard of endowment economies, researchers into a canonical DSGE model can also dynamic stochastic general equilibrium have been able to generate reasonable produce a large and variable term pre- (DSGE) model used in macroeconom- term premiums by assuming that inves- mium without compromising the mod- ics is far too small and stable relative tors have recursive Epstein-Zin prefer- el’s ability to fit key macroeconomic to empirical measures obtained from ences and face long-run economic risks. variables. Long-run real and nominal the data — an example of the “bond Rudebusch and Swanson show that risks further improve the model’s abil-

60 NBER Reporter • 2008 Number 4 ity to fit the data with a lower level of such that insurance against currency that these alphas primarily arise from household risk aversion. crashes is cheapest precisely when it is the disproportionate weight the Fama- Uncertainty plays a key role in needed most. Although excess returns French factors place on small value economics, finance, and decision sci- to crash-neutral strategies decline rela- stocks that have performed well, and ences. Financial markets, and in partic- tive to their unhedged counterparts, from the CRSP value-weighted mar- ular derivative markets, provide fertile they remain positive and highly statisti- ket index that is a downward-biased ground for understanding how per- cally significant. These results indicate benchmark for U.S. stocks. The authors ceptions of economic uncertainty and that crash risk premiums can explain explore alternative ways to construct cash-flow risk manifest themselves in 30-40 percent of the total excess return these factors and propose alternative asset prices. Drechsler and Yaron dem- to currency carry trades. Rationalizing models constructed from common and onstrate that the variance premium, all of the excess return via a crash risk easily tradable benchmark indexes. Such defined as the difference between the premium would require implied vola- index-based models outperform the squared VIX index and expected real- tilities of out-of-the-money currency standard models, in terms of both asset ized variance, captures attitudes toward options to be roughly four times greater pricing tests and performance evalua- uncertainty. The authors present con- than those observed in the data. tion of mutual fund managers. ditions under which the variance pre- Bazdrech, Belo, and Lin show that The covariance between U.S. mium displays significant time variation the firm-level hiring rate predicts stock Treasury bond returns and stock returns and return predictability. A calibrated, returns in the cross-section of U.S. pub- has moved considerably over time. generalized Long-Run Risks model gen- licly traded firms, even after control- While it was slightly positive on aver- erates a variance premium with time ling for investment, size, book-to-mar- age in the period 1953–2005, it was variation and return predictability that ket, and momentum, as well as other particularly high in the early 1980s and is consistent with the data, while simul- known predictors of stock returns. The negative in the early 2000s. Campbell, taneously matching the levels and vola- predictability shows up in both Fama- Sunderam, and Viciera specify and tilities of the market return and risk free MacBeth cross-sectional regressions estimate a model in which the nominal rate. The evidence indicates an impor- and in portfolio sorts and it is robust term structure of interest rates is driven tant role for transient non-Gaussian to the exclusion of microcap firms from by five state variables: the real interest shocks to fundamentals that affect the sample. The authors propose a pro- rate, risk aversion, temporary and per- agents’ views of economic uncertainty duction-based asset pricing model with manent components of expected infla- and prices. adjustment costs in labor and capital tion, and the covariance between nomi- Currency carry trades implemented that replicates well the main empiri- nal variables and the real economy. The within G10 currencies have historically cal findings. Labor adjustment costs last of these state variables enables the delivered significant excess returns with make hiring decisions forward looking model to fit the changing covariance of annualized Sharpe ratios in excess of one. in nature and thus informative about bond and stock returns. Log nominal Jurek investigates whether these excess the firms’ expectations about future bond yields and term premiums are qua- returns reflect compensation for expo- cash-flows and risk-adjusted discount dratic in these state variables, with term sure to crash risk by analyzing the time- rates. The model implies that the invest- premiums determined mainly by the series dynamics of the moments of the ment rate and the hiring rate predicts product of risk aversion and the nom- risk-neutral distribution extracted from stock returns because these variables inal-real covariance. The concavity of currency options, and by examining proxy for the firm’s time-varying condi- the yield curve — the level of interme- returns to crash-neutral currency carry tional beta. diate-term bond yields, relative to the trades in which exposure to crashes has Standard Fama-French and Carhart average of short- and long-term bond been hedged by combining positions in models produce economically and sta- yields — is a good proxy for the level currencies and currency options. Risk- tistically significant nonzero alphas even of term premiums. The nominal-real neutral and realized skewness move for passive benchmark indexes, such covariance has declined since the early in opposite directions in response to as the S&P 500 and the Russell 2000. 1980s, driving down term premiums. realized currency returns, Jurek shows, Cremers, Petajisto, and Zitzewitz find

NBER Reporter • 2008 Number 4 61 Behavioral Finance

The NBER’s Behavioral Finance California, Berkeley, “Asset Pricing in Discussant: Lori Walsh, Securities and group met on November 22 in Cambridge. Large Information Networks” Exchange Commission Brad Barber, University of California, Discussant: Xavier Gabaix, New York Davis, and Terrance Odean, Hass School University and NBER Randy Cohen, Harvard University; and of Business at Berkeley, organized the Christopher Polk and Bernard Silli, meeting. These papers were discussed: Nicola Gennaioli, CREI, and Andrei London School of Economics, “Best Shleifer, Harvard University and Ideas” Camelia M. Kuhnen and Agnieszka NBER, “Local Thinking” Discussant: Clemens Sialm, University Tymula, Northwestern University, Discussant: Drazen Prelec, MIT of Texas, Austin and NBER “Rank Expectations, Feedback, and Social Hierarchies” Roger M. Edelen, University of Harrison Hong, Princeton University, Discussant: Shimon Kogan, University California, Davis; Richard Evans, and Leonard Kostovetsky, University of Texas, Austin University of Virginia; and Gregory of Rochester, “Red and Blue Investing: B. Kadlec, Virginia Tech, “What Do Values and Finance” Han Ozsoylev, University of Oxford, Soft-Dollars Buy? Performance, Expense Discussant: Raymond Fisman, and Johan Walden, University of Shifting, Agency Costs” Columbia University and NBER

Kuhnen and Tymula develop and average output, which is mainly attrib- no agent has an incentive to change the test a theoretical model of the role utable to the fight for dominance at the network structure. They show that if of self-esteem — generated by pri- top of the hierarchy. These results sug- agents are ex ante identical, then strong vate feedback regarding relative per- gest that in environments where mon- conditions are needed to allow for non- formance — on the behavior of agents etary incentives are weak, moral hazard degenerate network structures, includ- working on a simple task for a flat can be mitigated by providing feedback ing power-law distributed networks. If, wage. The authors isolate the impact to agents regarding their relative per- on the other hand, agents face different on one’s output of learning one’s rank formance, and by optimally choosing costs of forming links, which are inter- in the group as opposed to the effect the reference peer group. Therefore, preted broadly as differences in social of any reputation, strategy-updating, social hierarchy effects on productivity skills, then power-law distributed net- or peer-monitoring effects. Feedback may influence optimal team formation. works arise quite naturally. has both ex-ante and ex-post effects on Kogan, Ozsoylev, and Walden Gennaioli and Shleifer present the productivity of workers and on the study asset pricing in economies with a theory of judgment under uncer- dynamics of social hierarchies. Agents large information networks. They derive tainty based on the idea that indi- work harder and expect to rank better closed-form expressions for price, vola- viduals evaluate hypotheses by filling when they are told they may learn their tility, profitability, and several other in missing data (called frames) from ranking, relative to cases when they are key variables, as a function of the topo- their memory. The researchers assume told that feedback will not be provided. logical structure of the network. The that the frames that come to mind After receiving feedback, individuals authors focus on networks that are first are those most predictive of the who learn that they have ranked bet- sparse and have power-law degree dis- hypothesis being evaluated relative to ter than expected decrease their out- tributions, in line with empirical stud- its alternatives. Combining this view put, but they expect an even better rank ies of large scale human networks. This of frames with the assumption of lim- in the future; those told they ranked makes it possible to rank information ited memory yields a theory of deci- worse than expected will increase their networks along several dimensions sionmaking that accounts for many output and at the same time lower their and to derive several novel results. For representativeness-related findings of expectations of rank going forward. example, price volatility is a non-mono- Kahneman and Tversky, including the These effects are stronger in earlier tonic function of network connect- Linda experiment. rounds of the task, while subjects learn edness, as is average expected prof- Edelen, Evans, and Kadlec examine how they compare to their peers. This its. Moreover, the profit distribution several hypotheses regarding mutual- rank hierarchy is established early on, among investors is intimately linked to fund commission payments. Consistent and remains relatively stable afterwards. the properties of the information net- with an information motive, they find, Private relative rank information helps work. The authors also examine which relatively active funds pay higher excess create a ratcheting effect in the group’s networks are stable, in the sense that commissions and these “soft dollar”

62 NBER Reporter • 2008 Number 4 payments are associated with improved approximately 39 to 127 basis points ble (for example tobacco, guns, defense, return performance. However, excess per month depending on the bench- and natural resources). Although commissions are also related to an mark employed. This suggests two con- a higher fraction of Democrat-run expense-shifting motive and these pay- clusions. First, the U.S. stock market mutual funds are socially responsi- ments are associated with lower return does not appear to be efficiently priced, ble (SRI), this result holds for non- performance — suggesting that agency because even the typical active mutual SRI funds and after controlling for costs arise from soft-dollar payments. fund manager is able to identify a stock other fund and manager characteris- The strongest evidence for expense that outperforms. Second, the orga- tics. The effect is more than one-half shifting occurs with relatively contro- nization of the money management of the under-weighting observed for versial distribution expenses, and these industry appears to make it optimal for SRI funds. Furthermore, the authors payments exhibit the most severe per- managers to introduce stocks into their find that Democrat managers also tilt formance degradation (agency costs). portfolio that are not outperformers, towards firms with positive social fea- Overall, the impact of soft dollar pay- even though they are able to pick good tures, such as excellent employee rela- ments on performance is negative. stocks. tions and clean environmental records. Cohen, Polk, and Silli provide Do political values influence invest- They document similar results among powerful evidence that mutual fund ing? Hong and Kostovetsky answer a smaller sample of hedge fund manag- managers can pick stocks that outper- this question using data on the politi- ers, suggesting that lax corporate gov- form the market. They examine the cal contributions and stock holdings of ernance in the mutual fund industry performance of stocks that represent U.S. investment managers. They find is not the main driver of their results. managers’ “Best Ideas” and find that that mutual fund managers who make Finally, they discuss how political val- the stock that active managers display campaign donations to Democrats hold ues influence investing and the implica- the most conviction towards ex-ante less of their portfolios (relative to non- tions of their findings for the growing outperforms the market, and the other donors or Republican donors) in indus- SRI movement and stock prices. stocks in those managers’ portfolios, by tries that are deemed socially irresponsi-

✴

NBER Reporter • 2008 Number 4 63 Corporate Finance The NBER’s Corporate Finance of Illinois and NBER and Luigi Zingales, University of Program met on November 21 in Chicago and NBER, “Innovation and Cambridge. Organizers Holger Mueller, Viral Acharya, New York University; Institutional Ownership” NBER and New York University, and Raghuram Rajan, University of Discussant: David S. Scharfstein, Joshua Rauh, NBER and University of Chicago and NBER, and Stewart C. Harvard University and NBER Chicago, chose these papers to discuss: Myers, MIT and NBER, “The Internal Governance of Firms” Todd Gormley, Washington University Martijn Cremers, Yale University, and Discussant: Jeremy C. Stein, Harvard in St. Louis, and David Matsa, Yaniv Grinstein, Cornell University, University and NBER Northwestern University, “Growing “The Market for CEO Talent: Implica­ Out of Trouble? Legal Liability and tions for CEO Compensation” Josh Lerner, Harvard University and Corporate Responses to Adversity” Discussant: Xavier Gabaix, New York NBER, and Ulrike Malmendier, Discussant: Antoinette Schoar, MIT University and NBER University of California, Berkeley and NBER and NBER, “With a Little Help C. Fritz Foley, Harvard University from My (Random) Friends: Success Gerard Hoberg, University of and NBER; Dhammika Dharmapala, and Failure in Post-Business School Maryland, and Gordon M. Phillips, University of Connecticut; and Entrepreneurship” University of Maryland and NBER, Kristin Forbes, MIT and NBER, Discussant: Simon Johnson, MIT and “Product Market Synergies and “The Unintended Consequences of the NBER Competition in Mergers and Homeland Investment Act: Implications Acquisitions” for Financial Constraints, Governance, Philippe Aghion, Harvard University Discussant: Paul Tetlock, Columbia and International Tax Policy” and NBER; John Van Reenen, London University Discussant: Heitor Almeida, University School of Economics and NBER;

Cremers and Grinstein study the significantly increasing repatriations from problems and ensure firms have substan- market for CEO talent in public U.S. foreign affiliates.Dharmapala , Foley, and tial value, even without any external gov- corporations from 1993–2005. They Forbes analyze the impact of the tax ernance. Internal governance seems to find large fragmentation of CEO talent holiday on firm behavior. They find that work best when both top management pools. In particular, about 68 percent repatriations did not lead to an increase and subordinates are important to value of new CEOs are former employees of in investment, employment, or R and creation. The authors then allow for gov- their own firm (“insider CEOs”) and 86 D — even for the firms that lobbied for ernance provided by external financiers percent are former employees in firms the tax holiday stating these intentions. and find situations where external - gov belonging to the same industry (includ- Instead, a $1 increase in repatriations was ernance, even if crude and uninformed, ing insider CEOs). Talent pool structure associated with an increase of approxi- complements internal governance in explains several compensation practices: mately $1 in payouts to shareholders. improving efficiency. Interestingly, this CEO compensation is only benchmarked These responses are consistent with the allows the development of a theory of against other firms, and pay-for-luck is view that the domestic operations of U.S. dividend policy, where dividends are paid prevalent only when the industry has a multinationals were not financially con- by self-interested CEOs to maintain a small percentage of insider CEOs. Finally, strained and that U.S. multinationals are balance between internal and external the authors study the importance of talent reasonably well-governed. The results also control. Finally, the authors explore how pools. Finding little support when incor- have significant implications for under- the internal organization of firms may be porating the fragmentation of CEO tal- standing the impact of the U.S. corporate structured to enhance the role of inter- ent pools, they offer a reinterpretation of tax system on the behavior of multina- nal governance. Young firms with limited the Gabaix-Landier results. tional firms. external oversight, and firms in countries The Homeland Investment Act of Acharya, Myers, and Rajan develop with poor external governance, can have 2004 provided for a one-time tax holi- a model of internal governance where substantial value, and improving external day on the repatriation of foreign earn- the self-serving actions of top manage- governance may not be a panacea for all ings, thereby allowing U.S. multination- ment are limited by the potential reaction governance problems. als to access earnings retained abroad at a of subordinates. The authors find that A central question in the entrepre- lower cost. Firms responded to this act by internal governance can mitigate agency neurship literature is how to encour-

64 NBER Reporter • 2008 Number 4 age entrepreneurship and whether peers build a model that nests managerial lazi- have career concerns. In support of the affect the decision to become an entrepre- ness with career-concern considerations, agency model, the authors find that total neur. Lerner and Malmendier exploit the where institutional ownership increases assets grow most among firms with weak fact that Harvard Business School assigns the incentives managers have to innovate external governance, high management students into sections, which have varying by reducing the career risk of innovative ownership, or low institutional owner- representation of former entrepreneurs. projects. While the lazy-manager hypoth- ship, whereas firms with strong external They find that the presence of entre- esis predicts a substitution effect between governance, low management ownership, preneurial peers strongly predicts sub- institutional ownership and product or high institutional ownership instead sequent entrepreneurship rates of their market competition, the career-concern increase their payouts to shareholders. peers who did not have an entrepreneur- hypothesis allows for complementarity. These results suggest that agency conflicts ial background, but in a more complex The effect of institutional investors on may be exacerbated when firms are closer way than the literature has previously sug- innovation increases with product market to financial distress. gested. A higher share of students with competition, supporting the career-con- Hoberg and Phillips examine how an entrepreneurial background in a given cern model. This model is also supported product similarity and competition influ- section leads their peers to lower rather by the authors’ finding that that CEOs are ences mergers and acquisitions and the than higher subsequent rates of entre- less likely to be fired in the face of profit ability of firms to exploit product market preneurship. However, the decrease in downturns when institutional ownership synergies. Using novel text-based analy- entrepreneurship is entirely driven by a is higher. sis of firm 10K product descriptions, they reduction in unsuccessful entrepreneur- Gormley and Matsa measure how find three key results: 1) Firms are more ial ventures. The relationship between the a typical firm responds when a chemical likely to enter restructuring transactions shares of pre-HBS and successful post- to which its workers are exposed is newly when the language describing their assets HBS peer entrepreneurs is insignificantly identified to be a carcinogen. While there is similar to all other firms, consistent with positive. Sections with few prior entrepre- is no evidence of a pre-existing trend, the their assets being more re-deployable. 2) neurs have a considerably higher variance authors find that firms, particularly those Targets earn lower announcement returns in their rates of unsuccessful entrepre- more vulnerable to the realization of an when similar alternative target firms exist. neurs. The authors argue that these results adverse shock, tend to undertake aggres- 3) Acquiring firms in competitive prod- are consistent with intra-section learn- sive growth and increased acquisitions uct markets experience increased profit- ing, where the close ties between section- after experiencing the liability shock. ability, higher sales growth, and increased mates lead to insights about the merits of The acquisitions appear to be targeted at changes in their product descriptions business plans. diversifying the firms’ assets by acquir- when they buy target firms that are simi- Aghion, Van Reenen, and Zingales ing large businesses with relatively high lar to them and different from rival firms. find that institutional ownership in pub- operating cash flows, recent growth, and The authors’ findings are consistent with licly traded companies is associated with total payouts. These deals are associated similar merging firms exploiting synergies more innovation (measured as cited- with high takeover premiums and neg- to create new products and increase their weighted patents), even after they control ative abnormal returns. These findings product differentiation relative to ex-ante for possible endogeneity of institutional are broadly inconsistent with the per- rivals. ownership. To explore the mechanism fect capital markets model, but fit well through which this link arises, the authors with an agency model where managers

NBER Reporter • 2008 Number 4 65 Bureau Books

The following volumes may be ordered directly from the University of Chicago Press, Chicago Distribution Center, at Telephone: 1-800-621-2736 Email: [email protected]

For more information on ordering and electronic distribution, see http://www.press.uchicago.edu/Misc/Chicago/infopage.html

Health at Older Ages: The Causes and Consequences of Declining Disability among the Elderly Health at Older Ages: The Causes of economists analyzes the foundations of fits that include an increased proportion and Consequences of Declining Disability disability decline, quantifies this phenom- of seniors in the workplace, relief for the among the Elderly, edited by David M. enon in economic terms, and discusses a healthcare system and care-giving fami- Cutler and David A. Wise, is available number of options for accelerating future lies, and reduced medical expenses for the from the University of Chicago Press this improvements in the health of our most elderly themselves. Health at Older Ages January. elderly populations. will be an essential contribution to the Americans are living longer — and This breakthrough volume argues debate about meeting the medical needs staying healthier longer — than ever that educational attainment, high socio- of an aging nation. before. Despite the rapid disappearance of economic status, an older retirement age, Both Cutler and Wise are Professors pensions and health care benefits for retir- and accessible medical care together have of Economics at Harvard University and ees, older people are healthier and better contributed to improved health and qual- members of the NBER’s Programs of off than they were twenty years ago. In ity of life for seniors. A decline in disabil- Research on Aging and Health Care. Health at Older Ages, a distinguished team ity, for example, yields economic bene-

Producer Dynamics: New Evidence from Micro Data

Producer Dynamics: New Evidence to address issues in industrial organiza- Dunne is a senior economic advi- from Micro Data, one of NBER’s Studies tion, labor, growth, macroeconomics, and sor in the Research Department of the in Income and Wealth edited by Timothy international trade. Federal Reserve Bank of Cleveland. Jensen Dunne, J. Bradford Jensen, and Mark J. This innovative volume brings is a Research Associate in the NBER’s Roberts, is available from the University together a group of renowned economists Program on International Trade and of Chicago Press this January. to probe topics such as: firm dynamics Investment and an Associate Professor The Census Bureau recently has across countries; patterns of employment in Georgetown University’s McDonough begun releasing official statistics that mea- dynamics; firm dynamics in nonmanufac- School of Business. Roberts is also an sure the movements of firms in and out turing industries such as retail, health ser- NBER Research Associate and is a of business and workers in and out of vices, and agriculture; employer-employee Professor of Economics at Pennsylvania jobs. The economic analyses in Producer turnover from matched worker/firm State University. Dynamics exploit this newly available data datasets; and turnover in international on establishments, firms, and workers, markets.

66 NBER Reporter • 2008 Number 4 The Structure of Wages: An International Comparison The Structure of Wages: An resents a first effort to examine linked tive perspective that examines wage and International Comparison, part of the employer-employee data across countries income differences within and between NBER’s Comparative Labor Markets and to analyze labor trends and their insti- countries, this volume will be required Series edited by Edward P. Lazear and tutional background in the United States reading for economists and those working Kathryn L. Shaw, is available from the and eight European countries: Belgium, in industrial organization. University of Chicago Press this January. Denmark, France, Italy, The Netherlands, Lazear is the Jack Steele Parker The distribution of income, the rate Norway, Sweden, and West Germany. A Professor of Human Resources Manage­ of pay raises, and the mobility of employ- distinguished team of contributors reveal ment and Economics at Stanford Uni­ ees are all crucial to our understanding how a rising wage variance rewards star versity’s Graduate School of Business; of labor economics. But though there employees at a higher rate than ever Shaw is the Ernest C. Arbuckle Professor is much research on the distribution of before, how talent has become concen- of Economics at the same institution. wages across individuals in the economy, trated in a few firms over time, and how Both are members of the NBER’s Program wage differentials within firms remain a outside market conditions affect wages in of Research in Labor Studies. mystery to economists. This volume rep- the twenty-first century. From a compara-

NBER Reporter • 2008 Number 4 67 Nonprofit Org. NBER U.S. Postage Reporter PAID NATIONAL BUREAU OF ECONOMIC RESEARCH National Bureau of Economic Research 1050 Massachusetts Avenue Cambridge, Massachusetts 02138-5398 (617) 868-3900

Change Service Requested