A Service of

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

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

Periodical Part NBER Reporter Online, Volume 2006

NBER Reporter Online

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

Suggested Citation: National Bureau of Economic Research (NBER) (Ed.) (2006) : NBER Reporter Online, Volume 2006, NBER Reporter Online, National Bureau of Economic Research (NBER), Cambridge, MA

This Version is available at: http://hdl.handle.net/10419/61989

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 WINTER 2006/7

Program Report

IN THIS ISSUE

Program Report Productivity  Productivity

Research Summaries Exchange Rate Regimes … 0 Microeconomic Evidence on Contagion … 3 Ernst R. Berndt* Tax Policy Towards Energy … 6

NBER Profiles 9 Conferences 21 NBER News 35 Program and Working Group Meetings 39 It is now seven years since Zvi Griliches, the NBER’s Productivity Bureau Books 63 Program Director for its first twenty years, passed away in October 1999. Due in large part to Zvi’s enormous intellectual legacy and the extraordi- nary efforts he expended in nurturing and mentoring young scholars, I am pleased to report that the NBER’s Productivity Program today is vibrant and robust, and that its researchers are carrying on the tradition of examin- ing sources and consequences of innovation and productivity growth, and in the process developing and empirically exploiting new datasets. Efficiency and productivity are essential concepts in almost every ’s tool kit, and thus it is not surprising that many of the NBER’s Productivity Program members are affiliated with other NBER Programs where these concepts are important as well, including Labor Studies, Industrial Organization, Corporate Finance, Economic Fluctuations and Growth, International Trade and Investment, , and NBER Service Brings Health Care. What distinguishes the Productivity Program’s research You New Data for Free focus from these other Programs at the NBER is its strong emphasis on the roles of research and development, patents, incentive systems, regula- A new, free NBER email ser- tions, knowledge spillovers, technological progress, organizational form, vice gives you daily email links to and market structure in influencing the extent and nature of produc- all U.S. government data releases, tivity growth and innovation. In addition, a disproportionate share of including unemployment, trade, Productivity Program researchers have traditionally focused considerable interest rates, GDP, etc. We keep track of your preferences and email attention on issues involving economic measurement, such as measures you the requested links when they of inputs, outputs, prices, quality change, and multifactor productivity, a are released. To sign up for any or focus that reflects Zvi Griliches’s enduring bequest. all of the government releases, visit www.nber.org/releases and register * Berndt directs the NBER’s Productivity Program and is the Louis B. Seley your choices. IT’S — FREE!! Professor of Applied Economics at the MIT Sloan School of Management. In this article, the numbers in parentheses refer to NBER Working Papers.

NBER Reporter Winter 2006/7  Rather than attempting to summarize the full scope of program activity, much of which NBER Reporter overlaps with other NBER programs, I will high- light in this report research in six broad areas, domains particularly prominent in Productivity Program research over the last four to five years. The sequence I follow will begin with research The National Bureau of Economic Research is a private, nonprofit research orga- on individual inventors, followed by research nization founded in 1920 and devoted to objective quantitative analysis of the on knowledge flows within and across firms American economy. Its officers and board of directors are: and other institutions, on patents and intellec- President and Chief Executive Officer — Martin Feldstein tual property protection, on market structure, Vice President for Administration and Budget — Susan Colligan international trade and investment, and recent Controller — Kelly Horak research on macroeconomics and productivity BOARD OF DIRECTORS growth, particularly on the role of information Chairman — Elizabeth E. Bailey and communications technology investments. Vice Chairman — John S. Clarkeson Treasurer — Robert Mednick Innovation at the Level of DIRECTORS AT LARGE the Individual Inventor Peter Aldrich Jessica P. Einhorn John Lipsky Elizabeth E. Bailey Martin Feldstein Laurence H. Meyer Does technological progress, by expand- John Herron Biggs Jacob A. Frenkel Michael H. Moskow Andrew Brimmer Judith M. Gueron Alicia Munnell ing knowledge, place an increased educational John S. Clarkeson Robert S. Hamada Rudolph A. Oswald burden on successive generations of innova- Don R. Conlan George Hatsopoulos Robert T. Parry tors? Do today’s innovators spend longer time Kathleen B. Cooper Karen N. Horn Marina v. N. Whitman George Eads Judy Lewent Martin B. Zimmerman in learning, and/or do they become more nar- rowly expert? Benjamin Jones (11359) shows DIRECTORS BY UNIVERSITY APPOINTMENT that the age at which Nobel Prize winners and other great inventors produce great ideas has , California, Berkeley Joel Mokyr, Northwestern Jagdish W. Bhagwati, Columbia Andrew Postlewaite, Pennsylvania increased substantially over the twentieth cen- Michael J. Brennan, California, Los Angeles Craig Swan, Minnesota tury, specifically because of a large drop in pro- Glen G. Cain, Wisconsin Uwe Reinhardt, Princeton ductivity at young ages, and is closely related to Ray C. Fair, Yale Nathan Rosenberg, Stanford Franklin Fisher, MIT David B. Yoffie, Harvard an increasing age at completion of formal educa- Saul H. Hymans, Michigan Arnold Zellner, Chicago tion. Focusing on more ordinary inventors, Jones Marjorie B. McElroy, Duke (11360) shows that the age at first patent, team- DIRECTORS BY APPOINTMENT OF OTHER ORGANIZATIONS work, and specialization are all increasing over time. These papers suggest dramatic changes Richard B. Berner, National Association for Business Economics Gail Fosler, The Conference Board in the nature of innovation, with a decline in Richard C. Green, American Finance Association output by the very young and a ubiquitous Dr. Arthur Kennickell, American Statistical Association move towards greater teamwork in the imple- Thea Lee, American Federation of Labor and Congress of Industrial Organizations mentation of ideas. Related research by David Robert Mednick, American Institute of Certified Public Accountants Galenson (12185, 12058) on artistic innovation Angelo Melino, Canadian Economics Association finds that artists who innovate early in their lives Jeffrey M. Perloff, American Agricultural Economics Association John J. Siegfried, American Economic Association do so suddenly, while those who innovate late do William W. Lewis, Committee for Economic Development so more gradually. Gavin Wright, Economic History Association In a series of papers (9017, 10923, 11654) The NBER depends on funding from individuals, corporations, and private foun- Kenneth Sokoloff and collaborators have used dations to maintain its independence and its flexibility in choosing its research new micro data sets on patents, inventors, and activities. Inquiries concerning contributions may be addressed to Martin Feldstein, President & CEO, NBER 1050 Massachusetts Avenue, Cambridge, patent assignment contracts in the MA 02138-5398. All contributions to the NBER are tax deductible. beginning in the nineteenth century, and exam- ined the changing division of labor between The Reporter is issued for informational purposes and has not been reviewed by the Board of Directors of the NBER. It is not copyrighted and can be freely repro- those who invented new technologies and those duced with appropriate attribution of source. Please provide the NBER’s Public who exploited them commercially. Soon after Information Department with copies of anything reproduced. the major patent reform of 1836, intermediar- Requests for subscriptions, changes of address, and cancellations should be sent ies — such as patent lawyers, agents, and agen- to Reporter, National Bureau of Economic Research, Inc., 1050 Massachusetts cies — emerged, facilitating transactions between Avenue, Cambridge, MA 02138-5398. Please include the current mailing label. buyers and sellers of patents. However, the move-

2 NBER Reporter Winter 2006/7 ment of inventors into firms — at least as Managing R and D and innovation ence of nearby industrial facilities helped employees — did not proceed very far raises a number of issues regarding incen- shape the direction of university research before the 1920s, at which time inventive tives. Beginning in the late 1980s, U.S. programs, there was a significant, positive activity shifted to R and D laboratories corporations increasingly linked compen- and causal effect running from university housed in large corporations. In between, sation of central research personnel to the research to the growth of pharmaceutical inventors developed long-term attach- economic incentives of the corporation. research laboratories in the first half of the ments with a firm in which the inventor Joshua Lerner and Julie Wulf (11944) twentieth century. was a principal, often by bringing inves- examine the impact of the shifting com- An obvious way in which knowl- tors with them. In related research com- pensation of the heads of corporate R and edge flows can occur between academia paring inventors in the United States and D. They report that among firms with and industry is via collaborative publica- Britain between 1790 and 1930 (10966), centralized Rand D operations, more tions. Does the success of these collabo- Sokoloff and collaborators report that long-term incentives (for example, stock rations depend on the research status of the relatively low patent application fee options and restricted stock) are associ- the faculty involved? Focusing initially in the United States (about 2.5 percent ated with more heavily cited patents, with on biotechnology, Lynne Zucker, Michael of that in Britain until late in the nine- more patent filings, and with patents of Darby, and Jeff Armstrong (8499) report teenth century), combined with the U.S. greater generality. While they cannot dis- that counts of collaborative publications administrative examination rather than tinguish between the roles of better proj- by top research university faculty and the British registration and prize proce- ect selection or better people selection, firm employees are an empirically use- dures, resulted in U.S. patentees typically they interpret these findings as being con- ful indicator of knowledge transfer lead- having relatively humbler origins than sistent with the view that performance ing to firm success, but that collabora- their British counterparts.1 pay of corporate R and D heads is associ- tive articles of the very top star academic Another research stream, more theo- ated with more innovative firms. scientists with firm employees predicted retical, focuses on incentives faced by aca- significantly more firm success than - col demic researchers. It examines whether Knowledge Flows and Innovation laborations with other faculty. Based on returns to university licensing divert fac- Across Organizations a census of biotechnology firms that did ulty from basic to applied research and to and did not go public, Darby and Zucker less leisure as they age. Marie Thursby and R and D (8954) find that among other factors, the co-authors (11197, 10758) also intro- strength of the firm’s science base (use of duce complications from tenure and from Geographic proximity between aca- recombinant DNA technology, number the fact that academic researchers may demia and industry R and D laboratories of articles by star academic scientists as — earn license income and enhanced pres- has long been hypothesized to facilitate or with — firm employees, number of bio- tige both inside and outside the univer- knowledge spillovers. Using program- tech patents) reduced the time to initial sity. These authors conclude that it is far level data on pharmaceutical drug dis- public offering (IPO) and increased the from obvious that licensing damages basic covery expenditures during the 1980s expected proceeds raised from the IPO. research and education. and 1990s and location-program-level Zucker and Darby (9825, 11181) While patent data has long been used data on relevant academic science, Jeffrey have extended this line of research to in empirical research, information on the Furman, Margaret Kyle, Iain Cockburn, study nanoscale science and technology identity (name and location) of the inven- and Rebecca Henderson (12509) study developments. An almost completed out- tor has seldom been employed, because of how proximity of pharmaceutical research put of this research is NanoBank.org, a what Manuel Trajtenberg (12479) calls laboratories to universities and the labora- public digital library matching and link- the “who is who” problem: the name of a tories of competing pharmaceutical com- ing individuals and organizations within given inventor may be spelled differently panies influenced the number of pat- and across the nanotechnology subsets across his/her patents, and the same name ents generated by the laboratories. They of the Institute for Scientific Information may correspond to different inventors find that “public” science — generated (ISI) Web of Science, U.S. patent data, (the “John Smith” problem). To address by universities, academic medical cen- and firm financial records; the NBER this problem involving over four million ters, and government laboratories — gen- will host a conference in 2007 where pre- patent records and 1.6 million inven- erated positive spillovers to geographi- sented papers will use data drawn from tors from across the world, Trajtenberg cally proximate private pharmaceutical the beta version of NanoBank. The con- developed an elaborate computerized research laboratories, but that no such cept of star scientist in numerous other data mining methodology, resulting in spillovers occurred among private labora- areas of science and technology is exam- detailed data on individuals’ patenting tories.2 Focusing on an earlier phenome- ined empirically in Zucker and Darby history, their employers, and co-inven- non — the growth of U.S. industrial phar- (12172), in which they follow the 1981– tors. He finds that 40 percent of patentees maceutical laboratories between 1927 2004 careers of 5,401 star scientists, as have more than one patent, and 70,000 and 1946, Megan MacGarvie and Jeffrey measured by ISIHighlyCited.com. They have more than ten patents. Furman (11470) find that while the pres- find that the number of stars in a U.S.

NBER Reporter Winter 2006/7 3 geographical region, or in one of the Bibliometric research has long quan- verse of industry patent citations from a top-25 science and technology countries, tified knowledge flows and influence by set of California-based research universi- generally has a significant and quantita- measuring citations in published articles ties, and assess changes over time in the tively large positive impact on the prob- and patents. The Institute for Scientific propensity of U.S. industry patents to cite ability of firm entry in the same area of Information collects data from the top these papers, controlling for a variety of science and technology and that other 110 U.S. research universities in twelve other factors. Branstetter and Ogura find measures of academic knowledge stocks main fields that cover nearly all of science. patterns in their data consistent with the have weaker and less consistent effects. Based on 1981–99 data on 2.4 million notion that there has been an increase in Hence it is the stars themselves, more papers and 18.8 million citations, James knowledge spillovers from academic sci- than their discoveries, which play a key Adams, J. Roger Clemmons, and Paula ence to commercial invention, but that role in the formation or transformation E. Stephan (10875) compute citation this increase is highly concentrated in a of high-tech industries. In terms of migra- probabilities as actual citations divided small number of technical fields. tion, Zucker and Darby report that in the by potential citations. The mean citation Diffusion of Superior Management United States stars become more concen- probability within fields is on the order Practices within Multinational Firms trated over time, moving from areas with of 10-5, whereas cross-field citations are relatively few peers to those with many in one tenth to one-hundredth as large, or While patent and professional jour- their discipline. On the other hand, these 10-6 to 10-7. Scientific influence is asym- nal citations are channels through which authors also document the tendency of metric within fields, and occurs primar- knowledge flows can be observed and foreign-born American stars to return to ily from top institutions to those less quantified, there are many other ways in their homeland when it develops suffi- highly ranked. Using the same database, which knowledge transfer can occur, par- cient strength in their area of science and Adams (10640) reports that team size ticularly within firms. NBER research- technology. (as measured by the number of authors ers are beginning to search for evidence Based on co-authorship counts on an article) has increased by about 50 of knowledge transfer within multina- among condensed matter physicists in the percent between 1981 and 1999, hold- tional firms. Productivity growth in sec- French “Centre National de la Recherche ing a number of other factors constant. tors intensively using information tech- Scientifique”, Jacques Mairesse and Laure Team size data are supplemented by mea- nologies (IT) has been greater in the Turner (11172) report that co-authorship sures of domestic and foreign institu- United States than in Europe since 1995. intensity is about 40 times higher within tional collaborations, capturing the geo- Using U.K. panel data on U.S. and non- a given laboratory than across laboratories graphic dispersion of team workers. The U.S. multinational-owned establishments, within the same city and about 100 times time-series evidence suggests that the Nick Bloom, Raffaella Sadun, and John greater than in laboratories in other cities. trend toward larger and more dispersed VanReenen find that U.S. owned estab- Immediate proximity is therefore critical. teams accelerates at the start of the 1990s, lishments have a stronger relationship Jeffrey Furman and Scott Stern which Adams conjectures may reflect sud- between IT and productivity capital than (12523) examine the impact of a different den declines in the cost of collabora- either non-U.S. multinationals or domes- institution — biological resource centers tion attributable to improvements in tele- tic establishments.3 This finding is robust (BRCs) — on the growth of the cumula- communications. Private universities and to inclusion of fixed effects and holds tive “knowledge stock” in molecular biol- departments whose scientists have earned when a sample of establishments taken ogy. BRCs authenticate, preserve, and offer prestigious awards participate in larger over by U.S. multinationals is examined. independent access to biological research teams, as do departments that have larger Moreover, this U.S. multinational effect materials such as cells, cultures, and speci- amounts of federal funding. Placement of of IT is particularly strong in sectors such mens, thereby reducing the marginal costs former graduate students is a key determi- as retail and wholesale that use IT inten- to researchers of building on prior research nant of institutional collaborations, espe- sively; notably, as discussed below, it is efforts. Employing a difference-in-differ- cially collaborations with firms and for- these very same industries that account ences estimator linking specific material eign scientific institutions. Adams finds for much of the U.S.-European produc- deposits to journal articles, these research- that scientific influence increases with tivity growth differential since the mid- ers report not only a selection effect (dis- team size and institutional collabora- 1990s. In related research, Lee Branstetter proportionately important materials are tions. He interprets increased team size (8015) finds that Japanese multinational deposited and preserved in BRCs), but also as reflecting an increase in the division of firms’ network of affiliates in the United that materials being deposited in BRCs labor, and concludes that scientific pro- States are a significant channel of knowl- result in a significant marginal “boost” in ductivity increases with the scientific divi- edge spillovers from Japan to the United the diffusion of knowledge, as measured sion of labor. States and vice-versa. These results are by journal article citations. The latter effect Since relatively little academic research consistent with those of Wolfgang Keller increases with time and varies with the is patented, and only a fraction of the pat- and Stephen Yeaple (9504) who, using economic and institutional conditions in ents are ever licensed, Lee Branstetter and firm-level Compustat data and detailed which deposits occur. Yoshiaki Ogura (11561) examine the uni- unpublished data from the U.S. Bureau

 NBER Reporter Winter 2006/7 of Economic Analysis on the industry entific research, the increased use of for- their existence does. The authors also classification of foreign-owned affiliates, mal IP appears to be significantly shaping report that the extension of patentability find evidence for major FDI spillovers the structure, conduct, and performance to software was initially negative for soft- from affiliates of foreign-owned firms in of both university and industry research- ware firms, especially for those producing the United States to U.S.-owned firms ers. Additional implications of IP protec- application software or services. between 1987 and 1996; such spillovers tion for the generation and accumulation The worldwide expansion of patent- explain about 11 percent of the U.S. of scientific and commercial knowledge ing activity by firms in many sectors has manufacturing multifactor productivity are considered by Murray and Stern in led to an increase in the uncertainty and growth during this time period. another NBER publication.4 Related costs associated with enforcing one’s own research by Bronwyn Hall (7643) and patents and defending against the pat- Patents, R&D, Innovation and Hall and Alfonso Gambardella (11120) ents of others. Building on earlier work Firms’ Valuations has centered on IP issues arising from uni- by Hall and Rosemarie Ziedonis (7062) versity-industry interactions, and docu- that found that increases in patenting in While the existence of intellectual ments the tensions that have arisen. the semiconductor industry were driven property (IP) protection through pat- Patents on software and business largely by a need to amass large defensive ents has long been thought to provide methods have become highly contro- patent portfolios because of technologi- positive incentives for R and D and inno- versial, with critics claiming that pat- cal complexity and threat of holdup, Hall vation, a growing “anti-commons” per- ents stifle innovation by holding up the (10605) reports that patent growth in the spective highlights the negative role of development of technology that builds United States since 1984 has taken place patents in facilitating knowledge accu- on patented prior art and by swamp- in all technologies, but not in all indus- mulation relative to publication in profes- ing inventors with patent-infringement tries, being concentrated in the electri- sional journals. Fiona Murray and Scott suits; see, for example, Josh Lerner and cal, electronics, computing, and scientific Stern (11465) note that a given discovery Feng Zhu (11168). Iain Cockburn and instruments industries. Although the 5–8 may contribute both to scientific research Megan MacGarvie (12563) examine the percent annual growth may reflect in part (journal publication) and to useful com- effects of software patents on entry and accelerated innovation, this growth has mercial applications (patents); they exam- exit in narrowly defined classes of - soft seriously affected patent offices worldwide ine patent-paper pairs in biotechnology, ware products, using a dataset with com- and has led to increasing concern over exploiting the fact that patents are granted prehensive coverage of both mature pub- patent quality and timeliness of issuance. with a substantial lag, often years after the lic firms and small privately held firms Research by Cockburn, Sam Kortum, and knowledge is initially disclosed through between 1994 and 2004; they find both Scott Stern (8980) on the relationship journal publication. Diffusion of citations stifling and stimulating effects of patents between poor examination and subse- occurs in both the pre-grant period and on entry. All else equal, greater numbers quent costly patent litigation, however, after formal IP rights are granted. Relative of patents held by incumbents have a was unable to uncover any relationship. to the expected citation pattern for pub- negative impact on entry rates into nar- Hall, Stuart Graham, Dietmar Harhoff, lications with a given quality level, the rowly defined software product markets, and David Mowery (8807, 9731) inves- “anti-commons” hypothesis predicts that while greater numbers of patents held tigate the workings and outcomes of the the citation rate to a scientific publication by entrants increase the rate of entry and patent opposition system, a procedure should fall after formal IP rights associ- decrease the rate of exit, all else equal. not available in the United States, but ated with that publication are granted. Related research by Lerner (7918, 10223) used in Europe, and assess how such a sys- Using a difference-in-differences esti- examines the impact of a notable judicial tem might function in the United States. mator for 169 patent-paper pairs (and decision involving State Street Bank on In related research, Lerner (7477, 7478 including a control group of other publi- financial patenting behavior. and 8977) examines patent office prac- cations from the same prestigious Nature In Hall and MacGarvie (12195), the tice, patent protection, and innovation journal for which no patent was granted), authors examine valuation effects of soft- over a 150-year time period, for a number Murray and Stern find evidence for a mod- ware patenting. Major changes in soft- of countries.5 est anti-commons effect, with the post-pat- ware patentability occurred in the U.S. David Popp, Ted Juhl, and Daniel ent grant decline in citations of about 10– Patent and Trademark Office in 1995. Johnson (9518) examine grant lags for 20 percent. This decline becomes more After1 995, software patents became more U.S. patent applications and find con- pronounced with the number of years valuable than ordinary patents, but Hall siderable differences across technology. elapsed since the date of patent grant, and and MacGarvie find that if the patents are Patents in biotechnology and and soft- is particularly salient for articles written held by hardware firms, then it does not ware experience the longest delays, but by researchers with public sector affilia- matter whether these patents are cited. for different reasons: biotechnology pat- tions. Hence, while this evidence suggests That is, the “importance” of software pat- ents are most likely to go through several that formal IP rights do not seem to have ents held by non-software firms has no revisions during the examination pro- a devastating impact on subsequent sci- impact on firms’ valuations, although cess, reflecting their complexity, whereas

NBER Reporter Winter 2006/7 5 software patents do not undergo revi- tially higher, which the authors interpret that the relationship between innova- sions more frequently, but rather sit in as suggesting that there may be underin- tion and competition appears to be an the queue longer. Adding more software vestment in innovative activities in these inverted U-shape. At low levels of com- examiners therefore might reduce grant two countries. petition, increases in competition spur lags, but lags in biotechnology approv- Conventional accounting practices on more innovation, but at higher lev- als are unlikely to be greatly affected by traditionally exclude from gross domestic els of innovation, additional competition increasing the number of biotechnology product investments in intangible capital, appears eventually to reduce any further examiners. such as R and D, patents, brand equity innovation. In their U.K. dataset, most The energy sector has been the focus and advertising, and human competency. industries were on the increasing part of of R and D by both government and Carol Corrado, Charles Hulten, and the slope, consistent with earlier findings industry. Around 1981, according to Popp Daniel Sichel estimate that this practice that the net impact of competition is to (11415), U.S. government R and D shifted implicitly ignores approximately $1 tril- increase innovation. They conclude that its focus from applied (for example, syn- lion of the output of the non-farm busi- more competition is, on average, likely to thetic fuels) to more basic in nature. Using ness sector in the United States by the late boost competition, but that at extremely patent citations as a measure of energy R 1990s, an amount approximately equal to high levels of competition there may be a and D quality, Popp finds that the likeli- the amount of investment spending on trade-off between the positive efficiency hood of a patent receiving a citation from tangible capital goods, and about 10 per- and pricing effects of additional compe- a future patent in the same field has fallen cent of gross domestic product.6 The same tition with the potentially negative inno- over time, which he interprets as evidence authors (11948) extend the time dimen- vation impacts. In contrast, arguing that of diminishing returns to R and D over sion and carry out a formal sources-of- growth in any country at any time is typi- time. Distinguishing government patents growth analysis. Among the more impor- cally uneven and instead concentrated from before and after the 1981 change in tant findings, the authors report that in a few firms in a few industries achiev- focus, Popp reports that government pat- intangible inputs have grown more rap- ing metamorphic technological progress ents filed after 1981 are more likely to be idly than other inputs over the last four as a result of highly uncertain but break- cited, and moreover, that descendants of decades, that most of this expansion is through innovations, Darby and Zucker these patents — private patents that cite not attributable to the growth in scientific (12094) argue (a la Schumpeter) that these government patents — are 30 per- R and D, but instead reflects growth in despite lagged diffusion, consumers’ wel- cent more likely to be cited by subsequent non-traditional intangibles such as non- fare is greater with dominant firms in con- patents. Popp concludes that government scientific R and D, and management and centrated industries because of enhanced and industry R and D have distinct roles human competencies. When intangibles innovation. to play in the innovation process. are included in the analysis, capital deep- Bee Yan Aw, Sukkyun Chung, and A lengthy NBER collaboration ening replaces multifactor productivity as Mark Roberts (8629) compare linkages among Bronwyn Hall, Adam Jaffe, and the principal source of economic growth among firm-level productivity, R and D Manuel Trajtenberg has put into the after 1995; notably, including intangibles investment, and survival for firms in the public domain a database that has been has little effect on the acceleration of mul- same industry in Korea and in Taiwan, used subsequently by many researchers tifactor productivity in the mid-1990s. They find that Taiwanese industries are (the NBER’s Patent Citation Data File). These, and a number of related issues characterized by less concentrated mar- An early contribution by these authors involving the construction of R and D ket structure, more producer turnover, (7741) investigated whether the citations satellite accounts as a supplement to the a smaller percentage of plants operating received by U.S. firm’s patents conveyed National Income and Product Accounts, at low productivity levels, and smaller information about their private stock are considered by Barbara Fraumeni and productivity differentials between surviv- market valuation; they found that pat- Sumiye Okubo.7 ing and failing producers. They interpret ent citations were more informative about these results as reflecting strong competi- value than the patents themselves, with Innovation, Organizational tive pressures in Taiwan that lead to mar- interesting variation across sectors. Form, and Market Structure ket selection based on productivity differ- Hall has also pursued the closely ences, and the presence of impediments to related area of the market value of R and Causality between innovation and entry or exit that insulate low productiv- D spending, both for U.S. firms (with Jaffe market structure generally has been ity producers in Korea. and Trajtenberg in 7741) and for firms in viewed as being bidirectional, reflecting In a different line of research, Chad a number of major European countries both Schumpeterian and network exter- Syverson (10501, 12231) examines how (with Raffaele Oriani in 10408). While R nality influences. In (9269), Philippe demand-side product substitutability and D undertaken in France, the United Aghion, Nick Bloom, Richard Blundell, affects industry structure and performance States, and Germany is valued similarly, Rachel Griffith, and Peter Howitt develop in the U.S. ready-mixed concrete industry, the value of R and D performed in the a theoretical framework and then demon- an industry where product substitutabil- United Kingdom and Italy is substan- strate empirically with U.K. firm data ity is determined largely by the density of

 NBER Reporter Winter 2006/7 concrete producers in the market. With uct technologies (8251, 9175, 9680, 9816, from trade. An early contribution was by high density, consumers’ ability to substi- 10165, 10956, 11136, and 11292). Jonathan Eaton and Sam Kortum (6253), tute implies that relatively inefficient pro- Interest in factors affecting an orga- for which they were awarded the Frisch ducers are unable to be profitable, resulting nization’s ability to innovate has evolved Medal in 2004. In this paper, Eaton and in higher minimum and average produc- to the development of a new form of sur- Kortum modeled observed aggregate tivity levels, less productivity dispersion, vey, called innovation surveys, in a number trade volumes as a tug of war between and lower prices because of spatial compe- of European countries. Based on a com- technology differences that led to more tition. Transport costs play a very critical mon core questionnaire, the country sur- trade and trade costs leading to less trade. role in the ready-made concrete industry, veys assemble information on innovators In subsequent research with Andrew so in additional work (10049) Syverson and non-innovators, where “innovators” Bernard and J. Bradford Jensen, Eaton explores the substitutability-productivity are defined as firms that have introduced and Kortum (7688) examined U.S. plant- link across a number of different industries. a new product or process over the last level export data, and empirically exam- He finds that product substitutability mea- three years, “new” is defined as substan- ined facts about how few firms export, sured in several ways — transport costs, tially improved or completely new, and how small a fraction of exporters are, and physical product differentiation, and adver- a distinction is made between products how much greater is their productivity. tising-driven brand differentiation — is new to the firm but not necessarily new to In Eaton, Kortum, and negatively related to within-industry pro- the market and products new to the firm (10344), this framework is extended ductivity dispersion and positively related and the market. Jacques Mairesse, along to study the detailed export behavior to industries’ median productivity levels. with co-authors Pierre Mohnen, Elizabeth of French firms, while in Johannes Van In related work using plant-specific physi- Kremp, and others (8644, 10237, 10897, Biesebroeck (10020) related research cal output and unit price measures, Lucia 12280, and 12320), has developed a three- focuses on the effects of exports on the Foster, John Haltiwanger, and Syverson tier framework that examines firms’ R productivity performance of sub-Saharan (11555) also report that physical produc- and D investment function, a knowledge manufacturing plants. Eaton and Kortum tivity is inversely correlated with plant- function with R and D as an input, and (12385) summarize this line of research level prices. an innovation output function. They use linking innovation, diffusion, and inter- In yet another strand of literature, this framework to account for differences national trade. as an alternative framework to “racing” across firms, industries, and countries in A most interesting case analysis of models that assume R and D competition the propensity for “innovativity”, analo- the productivity and price impacts of between a potential entrant and an incum- gous to multifactor productivity in tradi- entry by a large retailer is the study by bent, Joshua Gans, David Hsu, and Scott tional growth accounting analyses. They Beata Javorcik, Wolfgang Keller, and Stern (7851) endogenize the choice of interpret innovativity as reflecting the abil- James Tybout (12457), based on inter- product market competition versus cooper- ity to transform R and D and other innova- views of Mexican firms, on the response ation with established firms (via licensing, tion factors into innovation output, along of the Mexican soaps, detergents, and alliances, or acquisition). They hypothe- with other unobserved and unmeasured surfactant (SDS) producers to entry by size that the relative returns to cooperation factors. Wal-Mart. The authors argue that the increase with control over IP rights, low NBER researchers also have examined most fundamental effect of the North transaction costs, and greater sunk costs the relationship between market struc- American Free Trade Agreement and associated with product market entry. They ture and diffusion. Envisaging takeovers as the General Agreements on Tariff and find empirical support for all three factors playing roles similar to the entry and exit of Trade on Mexico’s SDS industry was to in determining commercialization strate- firms, Boyan Jovanovic and Peter Rousseau induce Wal-Mart to enter Mexico. Once gies, and conclude that the pro-competi- (9279) argue that from 1890–1930, when there, Wal-Mex fundamentally changed tive impact of start-up innovation — the electricity and the internal combustion the retail sector, forcing SDS firms to “gale of ” — depends engine spread through the U.S. economy, cut their prices and/or innovate. Those on imperfections in the market for ideas. and more recently from 1971–2001 (the unable to respond to this new environ- More generally, this research strand estab- “Information Age”), takeovers played a ment tended to lose market share and, in lishes that the industrial organization con- major role in accelerating the diffusion of some cases, to disappear altogether. As a sequences of start-ups are endogenous new technologies. result, many Mexican producers achieved to the commercialization environment, impressive efficiency gains, both from including factors such as the strength of Productivity and International labor shedding and innovation, which in IP rights, the availability of venture capita, Trade turn was fueled by innovative input sup- and sunk costs. Josh Lerner and various co- pliers and by multinationals bringing new authors consider related research on issues NBER Productivity researchers products and processes into Mexico. involving R and D and marketing-perfor- have contributed extensively to litera- Advocates of stronger IP protection mance impacts from various forms of alli- tures explaining bilateral patterns of trade in developing countries have suggested ances and financing for a variety of prod- between countries, and quantifying gains that stronger IP rights would induce mul-

NBER Reporter Winter 2006/7 7 tinationals to transfer more and better Gordon has addressed both of these issues in western Europe (the EU-15 countries) technology to IPR-reforming countries, in three papers. Separating actual produc- slowed down about as much as it acceler- and to do so at a more rapid rate than tivity growth from its underlying trend, ated in the United States, implying that would obtain in a weak IPR environ- Gordon calculates trend acceleration half of the divergence was attributable to ment, thereby benefiting the IPR-reform- from around 1.5 percent annually in the better American performance and half to ing countries. Using Bureau of Economic early 1990s to over 3 percent per year in worse European performance. He then Analysis data on foreign direct invest- 2002–3.8 Actual growth was even faster argues that since Europe uses much of the ment, Lee Branstetter, Raymond Fisman, in 2002–3, reflecting a standard cyclical same IT software and hardware as does and Fritz Foley (11516) report that recent phenomenon he calls the “early recovery the United States, it is implausible to cite IPR reforms have in fact resulted in a bubble”, in which during the early stage IT investment as playing an important quantitatively significant increase in tech- of an economic recovery, firms persist in role on either side of the Atlantic (recall nology transfer by U.S.-based multina- cutting costs and shedding labor, reacting my earlier discussion of different produc- tional parents to their affiliates in IPR- to the previous recession even as output tivity in U.S.- and European-owned mul- reforming countries. Bee Yan Aw, Mark begins to recover. Gordon attributes the tinational plants in Europe, by Bloom and Roberts, and Tor Winston (11174) focus labor productivity “explosion” in the early co-authors). Rather, he attributes the dif- on the complementary role of export mar- part of this decade primarily to unusu- ference to variations in specific industries, ket participation and R and D invest- ally deep corporate cost cutting result- most notably wholesale and retail trade, ment as a source of knowledge acquisition ing from the sharp drop in profits and and in finance. Regarding retailing, the and productivity growth for Taiwanese stock prices in 2000–2, from the after- key development that Gordon cites is the electronics producers, based on “learning math of the accounting scandals, and the development of the “big box” format, led by exporting”. They find that firms that increasing reliance of executive pay on by Wal-Mart, Home Depot, Best Buy, and export but do not invest in R and D have stock options. An alternative hypothesis, others. In turn, the success of this formula significantly higher future productivity offered by Susanto Basu, John Fernald, relies on largely deregulated use of land than firms that do not engage in either Nicholas Oulton, and Sylaja Srinivasan in the United States, where it is relatively activity, but that firms that both export (10010) is that much of the productivity easy to build a Wal-Mart at expressway and perform their own R and D have the payoff of the heavy IT investments of the interchanges. Gordon notes that much highest average future productivity lev- late 1990s were delayed because of lags in of European retailing still takes place in els among all groups. This pattern is con- adopting software and business practices small stores in central cities with little sistent with the hypothesis that R and D to the rapid improvements in IT hard- physical space to take advantage of mod- and exports are complementary activi- ware capability of the previous decade. ern technology. ties contributing to firm-level productiv- Gordon and Ian Dew-Becker (11842) A related but different set of issues are ity growth. examine the consequences of this labor examined by William Nordhaus (11354), More generally, the very substantial productivity growth on nominal wage who focuses on the productivity rebound literature appearing between 1993 and growth and inflation. While increased in the last decade in U.S. manufacturing 2003 on international technology diffu- productivity growth since 1995 has mod- industries, where manufacturing employed sion via the mechanisms of international erated inflation, Gordon and Dew-Becker has declined sharply. He finds that the trade and foreign direct investment, as obtain the provocative finding that over productivity rebound since 1995 has been well as the geographical localization and the entire 1966–2001 time period, only widespread, with approximately 40 per- productivity growth impacts of this dif- the top 10 percent of the income distri- cent of it occurring in “New Economy” fusion, is identified and summarized in bution achieved gains in wage and salary industries. Interestingly, Nordhaus finds Wolfgang Keller (8573). income equal to the growth rate of labor that the relevant productivity-employ- productivity, while the bottom 90 percent ment elasticities indicate that more rapid Aggregate Productivity fell behind. The authors attribute this productivity growth leads to increased, Growth and the Role of increased skewness of the income distri- rather than decreased, employment in Information Technology bution to disproportionate income gains manufacturing. This leads him to - con Investments at the top — primarily of entertainment clude that productivity growth is not to and sports “superstars” and to chief execu- be feared, at least not in U.S. manufactur- Two questions that have motivated tive officers of large corporations — along ing, where the largest recent employment macroeconomic productivity over the last with downward pressure for most work- declines have occurred. decade have been what caused the revival ers coming from shrinking unionization, Focusing on an earlier era, and using in U.S. productivity growth after 1995 rising imports, and job competition from pooled cross-section, time series data for and its further jump in 2001–4, and why immigration. 44 industries over the decades of the did productivity growth in Europe slow Turning to U.S.-European compari- 1960s, 1970s, and 1980s in the United down just as that in the United States was sons, Gordon (10661) documents that States, Edward Wolff (8743) finds no accelerating? NBER researcher Robert J. after1 995 the growth rate of productivity econometric evidence that computer

 NBER Reporter Winter 2006/7 investment is positively linked to multi- cult sectors such as finance and banking, forthcoming. factor (not labor) productivity growth. and computing consumers’ valuations of 5 An overview of patent policy develop- However, computerization is positively new goods. I also have omitted discussion ments and their consequences in the last associated with occupational restructur- of research on the productivity impacts two decades in the United States is found ing and changes in the composition of of various federal and state regulatory in A. Jaffe and J. Lerner, Innovation and intermediate inputs and capital coeffi- policies, including environmental regu- its Discontents: How Our Broken Patent cients. He also finds very modest evi- lations in the context of climate change. System is Endangering Innovation and dence that the growth of worker skills is Discussion of these issues is deferred to a Progress, and What To Do About It, positively related to industry productiv- subsequent issue of the NBER Reporter. Princeton: Press, ity growth. In other research in which he 2005. incorporates the age structure of capital 1 B. Zorina Khan and K. L. Sokoloff, 6 C. Corrado, C. Hulten, and D. Sichel, into the measurement of productivity, “Of Patents and Prizes: Great Inventors “Measuring Capital and Technology: An Wolff (9768) finds that once variations and the Evolution of Useful Knowledge Expanded Framework”, in C. Corrado, in the vintage composition of capital are in Britain and the United States, J. Haltiwanger, and D. Sichel, eds., taken into account and capital stock is 1750–1930”, Department of Economics, Measuring Capital in the New Economy, measured in efficiency units, multifactor University of California-Los Angeles. Press for the productivity growth is smoothed consid- Available from [email protected]. National Bureau of Economic Research, erably, particularly during the 1970s slow- 2 A less technical summary of this Conference on Research in Income and down, relative to ignoring the effects of research is given in M. Kyle, “Does Wealth, 2005, pp. 11–41. changing vintage capital composition. Locale Affect R&D Activity? The Case of 7 B. M. Fraumeni and S. Okubo, “R&D Pharmaceuticals”, Federal Reserve Bank in the National Income and Product Concluding Remarks of San Francisco Economic Letter, Nov. Accounts: A First Look at Its Effect on 13, 2004. GDP”, in C. Corrado, J. Haltiwanger, This summary of recent contributions 3 N. Bloom, R. Sadun, and J. and D. Sichel, eds., Measuring Capital in by NBER Productivity Program research- VanReenen, “It Ain’t What You Do but the New Economy, University of Chicago ers documents that the program con- the Way That You Do I.T.: Investigating Press for the National Bureau of Economic tinues to be wide-ranging, vibrant, and the Productivity Miracle Using Research, Conference on Research in robust. It is also worth noting that while Multinationals”, presented at 2005 NBER Income and Wealth, 2005, pp. 275–316. lengthy, this summary is incomplete in Summer Institute Productivity Program. 8 R. J. Gordon, “Exploding Productivity that I have omitted detailed discussion of Available from [email protected]. Growth: Context, Causes, and much Productivity Program research that 4 F. Murray and S. Stern, “When Ideas Implications”, Brookings Papers on focuses on measurement issues, such as are Not Free: The Impact of Patents on Economic Activity, Vol. 34 (2003, No. 2), that involving price measurement incor- Scientific Research,” in A. Jaffe, J. Lerner, pp. 207–98. Available as NBER Reprint porating quality changes, standard errors and S. Stern, eds., Innovation Policy and No. 2300. for price indexes, alternative measures of the Economy, Vol. 7, MIT Press for the innovation, measuring output in diffi- National Bureau of Economic Research,

NBER Reporter Winter 2006/7 9 Research Summaries

Exchange Rate Regimes, Globalization, Financial Crises, and Monetary Policy

Michael D. Bordo*

My research in the past decade has con- Today while advanced countries can market that mitigated these effects involved centrated largely on four related themes successfully float, emergers who are less significant start-up costs, while the impor- that I discuss in this article: Exchange Rate financially mature and must borrow abroad tance of scale suggests that network exter- Regimes, Globalization, Financial Crises, in advanced country currencies are afraid to nalities and liquidity were pivotal in the and Monetary Policy. float, for the same reason as their nineteenth emergence of overseas markets in domestic century forbearers were. To obtain access to currency debt.5 Exchange Rate Regimes foreign capital, they may need a hard peg to The limiting case of a fixed exchange the core country currencies. In my paper with rate regime is a monetary union. My study As discussed in the Fall 1999 NBER Marc Flandreau the key distinction between of the history of monetary unions (MUs) Reporter, much of my earlier work focused core and periphery countries, both then with Lars Jonung 6 — based on the exam- on the gold standard and related mone- and now, is financial maturity, evidenced in ples of the United States, Germany, and tary regimes. A series of papers with Finn the ability to issue international securities Italy — suggests that the success of MUs Kydland, Ronald MacDonald, and Hugh denominated in domestic currency3 (or the of the past has been intimately linked with Rockoff emphasized the importance of absence of “original sin”, a phrase coined by both fiscal and political unification. The credible commitment mechanisms in the Eichengreen and Haussman (1999)4). implementation of EMU was largely driven design of monetary regimes, focusing on the However, a case study by Chris Meissner, by the political will of elites and its ultimate gold standard.1 My recent work extends this Angela Redish, and myself of the debt history success may also depend upon the political approach. of several former colonies of Great Britain will of the citizenship. The choice of exchange rate regimes, (the United States, , Australia, New between fixed and floating exchange rates, Zealand, and ), who had largely Globalization evolved considerably in the past hundred overcome the problem of original sin by the years.2 Before 1914, advanced countries third quarter of the twentieth century, finds Globalization — the integration of adhered to gold while periphery countries that sound fiscal institutions, high credibility goods, labor, and capital markets — has been either emulated the advanced countries or of the monetary regimes, and good financial one of the dominant issues in the past sev- floated. Some peripheral countries were development are not sufficient to completely eral decades. The present era of globaliza- especially vulnerable to financial crises and break free from original sin. Conversely, tion was preceded by an earlier era in the late debt default, in large part because of their poor performance in these policy realms is nineteenth century — from 1870 to World extensive external debt obligations denomi- not, for the most part, a necessary condition War I. Globalization in historical perspec- nated in core country currencies. This left for Original Sin. The factor we emphasize tive was the subject of a recent NBER con- them with the difficult choice of floating for the common progress toward borrowing ference volume, edited with Alan Taylor but restricting external borrowing or devot- in domestic currencies across the five coun- and Jeffrey Williamson. The articles in the ing considerable resources to maintaining an tries is the presence of shocks, such as wars, book covered many aspects of the globaliza- extra hard peg. massive economic disruption, and the emer- tion experience, including the integration gence of global markets. The differences in of markets, growth convergence, inequal- *Bordo is a Research Associate in the NBER’s evolution between the United States and ity, financial development, the transmission Programs on Monetary Economics and the the Dominions we attribute to differences in of shocks, and the political economy of the Development of the American Economy. size, the role of a key currency, which charac- backlash in the interwar period that ended He is also a Professor of Economics at terized the United States and not the others, the first era of globalization. Rutgers University. During 2006–7, he and to membership in the British Empire. My research with Barry Eichengreen is the Pitt Professor of Economic History The importance of major shocks suggests and Douglas Irwin7 focused on a compari- at Cambridge University and a Fellow at that the establishment of a domestic bond son of the record of financial and commer- Kings College Cambridge.

10 NBER Reporter Winter 2006/7 cial globalization in the two eras of global- policies. One possible determinant of cri- 1973–98, with countries in the same region ization.8 The empirical evidence we survey ses in emerging countries is the presence of not receiving assistance, suggests that the suggests that, while in some respects the original sin. real performance of the former group was financial integration of the pre-1914 era My work with Chris Meissner finds, in possibly worse than the latter. Similar results remains unsurpassed, in other respects both eras of globalization, an increased prob- are obtained after adjusting for self-selection today’s financial markets are even more ability that emerging countries with original bias and counterfactual policies17. closely integrated than those in the past. sin experienced debt, currency, and bank- The difference today is that new informa- ing crises. Furthermore, crises were more Monetary Policy tion-generating-and-processing technolo- likely to be a problem for middle-income gies have reduced the market-segmenting emerging countries that were less financially Economic history has long provided effects of asymmetric information. In con- developed. In the first era of globalization, a useful laboratory for the practitioners of sequence, the range of financial claims that countries like Italy, Portugal, Argentina and monetary policy. My research in the area has are traded internationally has broadened. Brazil were more crisis-prone than Australia, focused on deflation and monetary policy While in the past entities (governments, Canada, Denmark, and Sweden. and asset prices. railroads, and mining companies) with tan- Another part of this research program Deflation in historical perspective gible and therefore relatively transparent focuses on the transmission of financial cri- assets were predominant, now international ses.11 Antu P. Murshid and I (2000)12 pres- The return in the 1990s to an environ- investors transact freely in a much broader ent evidence from weekly data on sovereign ment of low inflation has raised the spec- range of securities. bond prices and interest rates for episodes of ter in the United States of deflation and the We also find that the commercial inte- financial turbulence from1 880 to 1997. We collapse of prices such as occurred in the gration before World War I was more lim- find little evidence for contagion, defined 1930s. My work with Angela Redish focuses ited. Given that integration today is even as an incidence in bilateral cross-market on the deflationary experience of 1870–96 more pervasive than a hundred years ago, it correlations, adjusted for heteroscedastic- during the pre-1914 classical gold standard is surprising that trade tensions and finan- ity. However, we do find evidence sugges- period. That episode has resonance for today cial instability have not been worse in recent tive of transmission via fundamentals-based because the gold standard regime, anchored years. Institutional innovations that have trade channels. A related paper (Bordo and by a credible commitment to maintaining taken place in the past century, such as the Murshid, 2002)13, using principal compo- long-run price stability, conveys similarities evolution of sound monetary and fiscal poli- nents analysis on monthly spreads on long- to today’s central bank commitment to low cies, the establishment of the Bretton Woods term bond yields, finds that financial market inflation, and in both eras globalization was Institutions, GATT and the WTO, may be shocks were more globalized before 1914 present with major technology shocks. We part of the explanation. than they are now.14 use a structural VAR methodology to dis- Financial crises represent an extreme tinguish between good deflation, reflecting Financial Crises and Financial form of financial instability. Michael Dueker, productivity-driven increases in aggregate Instability David C. Wheelock, and I (2000, 2002)15 supply, and bad deflation, driven by col- construct indexes of financial stability for lapses in aggregate demand. Our findings The recent era of globalization has been the United States and the United Kingdom for the United Kingdom and Germany are characterized by a large number of finan- for 1790–1997. We find that aggregate price that deflation was primarily of the good vari- cial crises, both currency and banking cri- level and inflation shocks contributed to ety; for the United States, these results gen- ses, especially in the emerging markets.9 My financial instability in the 1790-1933 and erally prevail with the exception of a bank- research with Barry Eichengreen and others 1986-97 periods. ing-panic-induced demand-driven deflation considers whether this is a new phenom- Finally, my research has examined the episode in the mid-1890s.18 Andrew Filardo enon. Using a database for 21 countries role of policy in preventing, managing, and and I (2004)19 generalized this finding to a from 1880–1997, we compiled evidence on resolving crises. Anna J. Schwartz and I panel of over twenty countries for the past the incidence, duration, and costs of vari- (1998) examine the historical record over two centuries. With the exception of the ous types of crises across four exchange rate the past two centuries on international res- interwar period we find that deflation was regimes.10 Like the evidence on financial cues. We contrast the experiences before generally benign. integration, ours came with a strong feeling 1973 of rescues of monetary authorities of Monetary Policy and Asset Prices of déjà vu. The incidence and duration of advanced countries that were temporarily currency crises have increased since earlier short of liquidity with the experience in The link between monetary policy and eras of globalization, while those of bank- the 1990s of bailouts of insolvent emerg- asset price movements has been of peren- ing crises are similar. The output losses of cri- ing countries. In a later paper, we assess the nial interest to policymakers. Should the ses are quite similar across eras of globaliza- impact of IMF loans on the macro perfor- monetary authorities intervene to offset an tion except for twin crises, which have been mance of the recipients.16 A simple with- asset boom before it turns into a bubble that worse. Moreover, we find that crises seem without comparison of countries receiving bursts, with severe consequences for the real to reflect both economic structure and bad IMF assistance during crises in the period economy? Olivier Jeanne and I (2002)20

NBER Reporter Winter 2006/7  present some stylized facts on boom-bust Historical Perspective, M. D. Bordo, A. torical experience on banking crises. dynamics in stock and property prices in Taylor, and J. Williamson, eds., Chicago: 10 M. D. Bordo and B. Eichengreen, “Crises twenty OECD countries in the past thirty University of Chicago Press (2003); and M. Now and Then: What Lessons from the years, as well as for the United States for the D. Bordo, “Market Discipline and Financial Last Era of Financial Globalization”, NBER past 150 years. We find that asset booms Crisis Policy: An Historical Perspective”, Working Paper No. 8716, January 2002; turning into busts are infrequent events, but Research in Financial Services: Private and M. D. Bordo, B. Eichengreen, D.Klingebiel, when they happen they are sometimes asso- Public Policy, Vol. 18, (2004), pp 154–82. and S.Martinez-Peria, “Is the Crisis ciated with considerable economic distress. 4 B. Eichengreen and R. Haussman, Problem Growing More Severe?” Economic We develop a theoretical framework that “Exchange Rates and Financial Fragility”, Policy, (2001); and M. D. Bordo and B. outlines the circumstances under which a NBER Working Paper No. 7418, November Eichengreen, “Is Our Current International central bank may consider following a pre- 1999. Economic Environment Unusually Crisis emptive policy to prevent the consequences 5 M. Flandreau and N. Sussman, Prone?”, in International Financial System: of a bust. “Old Sins: Exchange Rate Clauses and Conference Proceedings, (1999), D. Wheelock and I (2004)21 examine the European Foreign Lending in the 19th Gruen and L. Gower, eds., Reserve Bank of economic environment in which past U.S. Century”, in Other People’s Money: Debt Australia, Sydney. stock market booms occurred as a first step Denomination and Financial Instability 11 Earlier evidence on the transmission of toward understanding how these booms in Emerging Markets Economies, B. banking crises is in M. D. Bordo, “ Some come about and whether monetary pol- Eichengreen and R. Hausmann, eds., Historical Evidence 1870–1933 on the icy should be used to defuse them. In gen- Chicago and London: University of Chicago Impact and Transmission of Financial Crises”, eral, we find that booms in the past century Press (2005). NBER Working Paper No. 1606, September and a half occurred in periods of rapid real 6 M. D. Bordo and L. Jonung, “ The 1986; and M. D. Bordo, B. Mizrach, growth and productivity advance, suggest- Future of EMU: What Does the History of and A. J. Schwartz, “Real Versus Pseudo ing that booms are driven at least partly by Monetary Unions Tell Us?” NBER Working International Systemic Risk: Some Lessons fundamentals. We find no consistent rela- Paper No. 7365, September 1999. Also From History”, Review of Pacific Basin tionship between inflation and stock mar- see M. D. Bordo, “The United States as a Financial Markets and Policies, (1998). ket booms, though; booms have typically Monetary Union and the Euro: A Historical 12 M. D. Bordo and A. P. Murshid, “Are occurred when money and credit growth Perspective”, Cato Journal, Vol. 24, No. 1–2, Financial Crises Becoming Increasingly More were above average. Spring/ Summer 2004, pp.163–70. Contagious? What is the Historical Evidence 7 M. D. Bordo, B. Eichengreen, and J. on Contagion?”, NBER Working Paper No. 1 M. D. Bordo and F. Kydland, “The Kim, “Was There Really an Earlier Period 7900, September 2000. Gold Standard as a Rule: An Essay in of International Financial Integration 13 M. D. Bordo and A. P. Murshid, Exploration”, Explorations in Economic Comparable to Today?”, NBER Working “Globalization and Changing Patterns in History (1995); M. D. Bordo and H. Paper No. 6738, September 1998; and M. the International Transmission of Shocks in Rockoff, “The Gold Standard as a `Good D. Bordo, B. Eichengreen, and D. Irwin, “Is Financial Markets”, NBER Working Paper Housekeeping Seal of Approval’”, Journal of Globalization Today Really Different than No. 9019, June 2002. Economic History (1996); and M.D.Bordo Globalization a Hundred Years Ago?”, in 14 This differs from Paulo Mauro, Nathan and R. MacDonald, “Violations of the Brookings Trade Policy Forum, S. Collins Sussman, and Yishay Yafeh, “Emerging ‘Rules of the Game’ and the Credibility of and R. Lawrence, eds, Brookings Institution, Market Spreads: Then Versus Now”, the Classical Gold Standard, 1880–1904”, Washington, D.C. (1999). Quarterly Journal of Economics 117, NBER Working Paper No. 6115, July 1997. 8 Also see M. Obstfeld and A. Taylor, (2002), pp.695–733, which may be attribut- 2 J. A. Frankel, “Experience of and Lessons Global Capital Markets: Integration, Crisis, ed to the use of different country samples. Our from Exchange Rate Regime in Emerging and Growth, Cambridge University Press, sample contains both advanced and emerg- Economies”, NBER Working Paper No. (2004). ing countries while theirs was based solely 10032, October 2003. Also M. D. Bordo and 9 In “Why Clashes Between Internal and on emerging countries. In “Have National H. James, “One World Money, Then and External Stability Goals End in Currency Business Cycles Become More Synchronized?”, Now”, NBER Working Paper No. 12189, Crises, 1797–1994”, NBER Working Paper NBER Working Paper No. 10130, December May 2006, posits that in today’s world of No. 5710, June 1997, Michael D. Bordo and 2003, Bordo and Thomas Helbling also find globalization and large shifts in relative Anna J. Schwartz present a narrative of the evidence that global shocks are the key deter- prices, the Einsteinian or relative conception history of currency crises from 1797–1994, minant of the synchronization of business of money implicit in floating exchange rates is which reflects a clash between internal and cycle movements observed across exchange rate more suitable than the Newtonian view that external stability goals. In “Some Historical regimes from 1880 to the present. underlies the case for fixed exchange rates. Evidence 1870-1933 on the Impact and 15 M. D. Bordo, M. Dueker, and D. C. 3 M. D. Bordo and M. Flandreau, International Transmission of Financial Wheelock, “Aggregate Price Shocks and “Core Periphery, Exchange Rate Regimes, Crises”, NBER Working Paper No. 1606, Financial Instability: An Historical Analysis”, and Globalization”, in Globalization in September 1986, Bordo examines some his- NBER Working Paper No. 7652, April

12 NBER Reporter Winter 2006/7 2000, and “Aggregate Price Shocks and October 2004, Michael D. Bordo, Ashoka and A. Redish, “Good versus Bad Deflation: Financial Stability: The United Kingdom Mody, and Nienke Oomes find evidence that Lessons from the Gold Standard Era”, NBER 1796–1999”, NBER Working Paper No. following the adoption of an IMF program, Working Paper No. 10329, February 2004. 8583, November 2001. both macroeconomic aggregates and capital 19 M. D. Bordo and A. Filardo, “Deflation 16 In “The International Monetary Fund: flows improve, although they may initially and Monetary Policy in a Historical Its Present Role in Historical Perspective”, deteriorate somewhat. Consistent with theo- Perspective: Remembering the Past or Being NBER Working Paper No. 7724, June 2000, retical predictions and earlier empirical find- Condemned to Repeat It?”, NBER Working Michael D. Bordo and Harold James describe ings, they also find that IMF programs are Paper No. 10833, October 2004. the evolution of the IMF from its origins as most successful in improving capital flows to 20 M. D. Bordo and O. Jeanne, “Boom- the guardian of the Bretton Woods adjustable countries with bad, but not very bad, fun- Busts in Asset Prices, Economic Instability, peg exchange rate system and financier of tem- damentals. In such countries, IMF programs and Monetary Policy”, NBER Working Paper porary current account deficits for advanced are also associated with improvements in the No. 8966, May 2002. countries to its present primary roles as devel- fundamental themselves. 21 M. D. Bordo and D. C. Wheelock, opment financier and crisis manager for the 18 M. D. Bordo and A. Redish, “Is Deflation “Monetary Policy and Asset Prices: A Look emerging world. Depressing? Evidence from the Classical Gold Back at Past U.S. Stock Market Booms”, 17 In “Keeping Capital Flowing: The Role of Standard”, NBER Working Paper No. 9520, NBER Working Paper No. 10704, August the IMF”, NBER Working Paper No. 10834, March 2003; and M. D. Bordo, J. L. Lane, 2004.

The Microeconomic Evidence on Contagion, Capital Controls, and Capital Flows

Kristin J. Forbes*

On December 19, 2006, exchange rates, stock indices, investment, and the definition has evolved over time.4 The implemented restrictive capital controls on growth. This “microeconomic approach” to last decade has clearly shown that crises that foreign investment and the Thai stock index answering questions in international finance originate in relatively small economies (such fell by over 15 percent before trading was has the benefit of using the wealth of infor- as Thailand) can quickly affect markets of halted. Investors quickly began to draw com- mation incorporated in firm behavior — in- very different sizes and structures located parisons to the summer of 1997, asking if formation that is lost in the aggregation pro- around the world, including markets that events in Thailand would spark contagion cess used to create macroeconomic statistics. appear unrelated to the country where the and be the first in a series of crises around the Heterogeneity across firms can be an impor- crisis originated. globe. My research over the past few years tant tool in identifying the impact of vari- Early analyses of contagion tested for has attempted to provide the framework ous macroeconomic events. This strategy of increased comovement between countries and analysis to understand these types of using disaggregated and firm-level data in after a crisis (in variables such as their stock events — covering topics from financial con- international economics recently has become returns, bond spreads, exchange rates, or tagion and global linkages to capital controls popular in the international trade literature1 capital inflows). One complication with this and capital flows. as well as in international finance.2 approach, however, is that the correlation Many researchers — including other coefficients underlying this analysis depend NBER affiliates — have made important Contagion, Currency Crises, on market volatility and can be biased. My contributions to these topics. Much of my and the Cross-Country work with Roberto Rigobon shows how this own research has tried to take a different Transmission of Shocks bias can significantly affect estimates of con- approach by using firm-level data to under- tagion.5 We develop a correction for this bias stand the macroeconomic movements in A series of financial crises — Mexico, and show that most recent crises were trans- Asia, Russia, Argentina — motivated an mitted to other countries through linkages * Forbes is a Research Associate in the academic literature on “contagion” and the that exist in all states of the world — and not NBER’s Program on International Finance 3 and an Associate Professor of Economics at international transmission of crises. The through special transmission channels that MIT’s Sloan School of Management. Her term contagion is generally used to refer to only occur during crises. Other authors have profile appears later in this issue. the spread of negative shocks — although since used different identifying assumptions

NBER Reporter Winter 2006/7 3 in order to adjust for this bias in tests for its, especially in the less developed finan- linking output growth for all countries in contagion.6 cial markets common in emerging markets. the world. Estimates and impulse-response This approach of testing how crises affect Moreover, even if stock returns accurately matrices show that these multiplier effects cross-country co-movements, however, has a capture expected future profitability, the can transmit shocks in very different pat- major shortcoming: it cannot explain why actual impact of a crisis could significantly terns than predicted by standard bilateral- these macroeconomic variables co-move or diverge from the expected impact. Therefore, linkage models. exactly how shocks are transmitted interna- in my next set of papers I shifted from focus- tionally. For investors interested in how a cri- ing on the financial market impact of crises to Capital Controls sis spreads, and especially for governments their broader impact on other firm-level vari- and policymakers that would like to contain ables, such as sales, profits, and investment. Governments and policymakers have and prevent crises, understanding exactly In one of these papers, I develop a sim- struggled continually with how to protect how shocks spread is of critical importance. ple model to clarify exactly how exchange their economies from contagion and the Do crises spread mainly through “real” link- rate movements affect firms’ profitability, negative effects of global linkages. One pre- ages, such as trade and banking flows? Or, focusing on the impact on the relative costs scription for emerging markets and devel- do they spread through investor behavior, of labor and capital for firms that produce oping countries that has recently gained driven by portfolio balancing or informa- undifferentiated products.9 The empirical support is capital controls — and especially tional asymmetries or herding or irrational- analysis shows that in the longer term, which controls on capital inflows. Several emerg- ity? Sorting out these various explanations firms gain and lose from devaluations is ing markets — from to Russia to is further complicated by the fact that many determined by the relative intensity of cap- Thailand — recently have implemented new cross-country linkages are highly correlated, ital and labor in production, as well as by controls on capital inflows. Although capital so it is difficult to identify these various how the devaluations affect interest rates. In controls may yield benefits in certain situa- mechanisms in empirical work. a closely related paper, I expand the analy- tions, my research suggests that capital con- In order to differentiate between these sis to include firms that produce differenti- trols come at a substantial cost and are not various mechanisms, my research moved ated as well as undifferentiated products.10 the panacea that many believe. from the macroeconomic to the microeco- The results show that firms with greater for- The most cited “success” story of capi- nomic level. Within each country there is a eign sales exposure have significantly better tal controls is the market-based tax on capi- large variation in how different companies performance (according to a range of indi- tal inflows enacted by Chile from 1991–8 are affected by shocks. By using this firm cators) after depreciations, and firms with (the encaje). Several papers have analyzed the heterogeneity it is possible to identify the higher debt ratios tend to have lower net macroeconomic impact of the encaje. The mechanisms by which crises spread. To put income growth. The results in these papers general conclusions in this literature are that this approach into practice, I created a data- can help to explain why devaluations boost the Chilean capital controls generated some set with financial statistics, product infor- economic growth in some cases, but cause small economic benefits (such as shifting mation, geographic data, and stock returns economic contractions in others. the composition of capital inflows to a lon- for over 10,000 companies in 46 countries.7 In a final series of papers on the cross- ger maturity), but no significant economic My results suggest that direct trade flows country transmission of shocks, I broaden my costs. In an attempt to better assess the and trade competition in third markets were focus to non-crisis as well as crisis periods.11 impact of these controls, my research shifted important determinants of country vulner- Work with Menzie Chinn tests whether to the microeconomic effects of the encaje ability to crises.8 I also find some evidence of cross-country linkages through direct trade, by studying its effect on Chilean firms. I a “portfolio recomposition” effect, in which trade competition in third markets, bank document that the capital controls distorted investors were forced to sell assets in liquid lending, or FDI can explain why movements firm’s financing decisions and significantly markets in response to a crisis elsewhere. in the world’s largest financial markets often increased financial constraints for smaller, Finally, I find little evidence of a generalized have such large, yet diverse, effects on other publicly traded firms.14 The Chilean capital “credit crunch” driving contagion. financial markets and how these cross-mar- controls appear to have led to a misallocation Most of the empirical research on the ket linkages have changed over time.12 The of resources away from smaller firms. Since transmission of crises — whether at the mac- paper uses a factor model and shows that smaller firms are valuable sources of job cre- roeconomic or microeconomic level — has movements in the U.S. market, as well as sec- ation and economic growth in many emerg- focused on how crises affect financial vari- toral factors, were important determinants ing markets, these results suggest that capital ables (such as their impact on stock prices). of market returns in the later half of the controls could have more widespread nega- This approach is not only timely, since the 1990s. Other work with Tilak Abeysinghe tive effects on investment and growth than financial information is immediately avail- looks at how spillover effects across markets previously believed. able, but also has the benefit that stock are amplified by having “multiplier effects” This work on Chilean firms is part of a returns should incorporate all known infor- on other countries, which in turn have their new literature on the microeconomic effects mation about the future impact of the crisis. own spillover effects.13 To track these mul- of capital controls.15 Most of the earlier lit- However, stock returns may be influenced tiplier effects, we develop a structural VAR erature on capital controls focused on cross- by factors other than expected future prof- model using a new specification strategy country, macroeconomic analysis and found

14 NBER Reporter Winter 2006/7 mixed evidence of the costs of controls and ses, these results suggest that multinationals Although this literature using microeco- the benefits of capital account liberaliza- often send capital in the opposite direction, nomic data to investigate these questions in tion.16 More recent papers in very different allowing their affiliates to expand economic international finance is only in its infancy, it fields have used microeconomic data and activity during currency crises when local already has yielded important insights and documented several ways in which capital firms are financially constrained. will continue to be a valuable approach in controls reduce the efficiency of financial Capital flows to U.S. multinational affil- the future. markets and significantly distort decision iates, however, are only a small proportion making by firms, individuals and govern- of global capital flows. A much bigger por- 1 See A. Bernard, “Firms in International ments. Although this literature is only in tion of global capital moves in the reverse Trade”, NBER Reporter, Fall 2006, for a its infancy, the initial results present a com- direction — into the United States. Why summary of recent work using this approach pelling case that capital controls yield wide- were foreigners willing to invest over $1.2 in international trade. spread and often unexpected costs. Countries trillion in the United States in 2006 — capi- 2 Due to space constraints, this research that have recently become enamored with tal inflows necessary to fund the U.S. current summary is not able to describe all of the work capital controls as a means of reducing the account deficit? Even more surprising, why in the new literature using microeconomic appreciations of their currencies and insu- are private sector firms and investors will- data in international finance. Key contribu- lating their economies from negative shocks ing to invest hundreds of billions of dollars tions include NBER working papers by Mark should carefully weigh these costs against the in the United States each year, even though Aguiar, Anusha Chari, Kathryn Dominguez, perceived benefits of the controls. they have recently earned substantially lower Gita Gopinath, Peter Henry, Sergio returns than if they had invested in similar Schmukler, Linda Tesar, Frank Warnock, Determinants of Global Capital assets in their own countries? and Shang-Jin Wei. Flows My recent work attempts to understand 3 S. Claessens and K. J. Forbes, eds., the microeconomic factors driving the deci- International Financial Contagion, Boston, Closely related to this work on cap- sions by firms and individuals to invest in the MA: Kluwer Academic Publishers, 2001. ital controls — a means of stopping the United States.18 Certain characteristics of 4 See K. J. Forbes and R. Rigobon, movement of capital across borders — is my U.S. financial markets (such as their liquid- “Contagion in Latin America: Definitions, more recent work on capital flows. In this ity, strong , and reserve Measurement and Policy Implications”, research, I continue to take a microeconomic status of the dollar) may make the United NBER Working Paper No. 7885, September approach to understanding macroeconomic States a major source of investment in an 2000, and Economia, 2001, 1(2), pp.1–46. trends, attempting to explain what drives optimal, portfolio maximization model. 5 K. J. Forbes and R. Rigobon, “No individual firms, investors and governments Foreigners, however, may also consistently Contagion, Only Interdependence: to send capital abroad. underestimate the returns from holding U.S. Measuring Stock Market Co-movements”, Research with Mihir A. Desai and Fritz assets. Resolving exactly what drives the indi- NBER Working Paper No. 7267, July 1999, Foley focuses on one specific type of capital vidual decisions behind capital flows into and Journal of Finance, 2002, 57(5), pp. flow: how multinational firms adjust capital the United States is critically important to 2223–61. flows to their foreign affiliates during crises.17 understanding if the current system of global 6 For example, see B. Boyer, M. Gibson, More specifically, we examine how local imbalances can be expected to continue, and and M. Loretan, “Pitfalls in Tests for firms and multinational affiliates respond to if not, how quickly it could unwind. Changes in Correlations,” Federal Reserve large currency movements. Although eco- Board International Finance Division, nomic theory suggests that depreciations Looking Forward Working Paper 597R, 1999; G.Corsetti, M. can reduce relative production costs and Pericoli, and M. Sbracia, “Some Contagion, therefore give firms a competitive advantage, Although there have not been any Some Interdependence: More Pitfalls in many firms claim that a “credit crunch” dur- major financial market crises over the past Tests of Financial Contagion,” Journal of ing depreciations makes it difficult to obtain few years, there will inevitably be more cri- International Money and Finance, 2005, 8, the financing necessary to increase produc- ses in the future. Moreover, as global capital pp. 1177–99; and R. Rigobon, “Contagion: tion. Our research shows that multinational flows continue to increase, and especially as How to Measure It?”, NBER Working Paper affiliates increase sales, assets, and invest- more complicated financial instruments gain No. 8118, February 2001. ment significantly more than local firms importance, new risks and vulnerabilities 7 K. J. Forbes, “The Asian Flu and Russian during, and subsequent to, depreciations. will emerge. Understanding the factors driv- Virus: Firm-level Evidence on How Crises Multinational affiliates outperform local ing these capital flows and the impact of dif- are Transmitted Internationally”, NBER firms because they receive credit from their ferent policies to reduce vulnerabilities will Working Paper No. 7807, July 2000, and parent companies and therefore avoid the be an important research topic. Using micro- Journal of International Economics, 2004, “credit crunch” affecting local firms. While economic data to explain exactly what drives 63(1), pp. 59–92. much of the previous literature focused on the related decisions by firms and individu- 8 This role of trade linkages (and espe- how foreign investors tend to withdraw als and how they are affected by different cially competition in third markets) in capital from emerging markets during cri- policies will be a powerful tool of analysis. the transmission of crises is also supported

NBER Reporter Winter 2006/7 5 by sectoral analysis. See K. J. Forbes, “Are Financial Markets Over Time”, NBER Mud in the Wheels of Market Discipline”, Trade Linkages Important Determinants of Working Paper No. 9555, March 2003, and NBER Working Paper No. 10284, February Country Vulnerability to Crises?”, NBER Review of Economics and Statistics, (2004), 2004, and The Cato Journal, 2005, 1, pp. Working Paper No. 8194, March 2001, and 86(3), pp.705–22. 153–66. S. Edwards and J.A. Frankel, eds. Preventing 13 T. Abeysinghe and K. J. Forbes, “Trade 16 For excellent surveys of the literature on Currency Crises in Emerging Markets, Linkages and Output-Multiplier Effects: capital account liberalization, see P. Henry, (2002), pp. 77–124. A Structural VAR Approach with a Focus “Capital Account Liberalization: Theory, 9 K. J. Forbes, “Cheap Labor Meets on Asia”, NBER Working Paper No. Evidence, and Speculation”, NBER Working Costly Capital: The Impact of Devaluations 8600, November 2001, and Review of Paper No. 12698, November 2006, and M. on Commodity Firms”, NBER Working International Economics, 2005, 13(2), pp. A. Kose, E. Prasad, K. Rogoff, and S. J. Wei, Paper No. 9053, July 2002, and Journal of 356–75. “Financial Globalization: A Reappraisal”, (2002), 69(2), pp. 14 K. J. Forbes, “One Cost of the Chilean NBER Working Paper No. 12484, August 335–65. Capital Controls: Increased Financial 2006. 10 K. J. Forbes, “How Do Large Constraints for Smaller Traded Firms”, 17 M. Desai, C. F. Foley, and K. J. Forbes, Depreciations Affect Firm Performance?”, NBER Working Paper No. 9777, June 2003, “Financial Constraints and Growth: NBER Working Paper No. 9095, August and Journal of International Economics, Multinational and Local Firm Responses to 2002, and IMF Staff Papers (2002), 49, pp. 2007, forthcoming. Currency Crises,” NBER Working Paper No. 214–38. 15 K. J. Forbes, “The Microeconomic 10545, June 2004, and Review of Financial 11 See the symposium on Global Linkages Evidence on Capital Controls: No Free Studies, 2007, forthcoming. published in the Review of Economics and Lunch”, NBER Working Paper No. 18 K. J. Forbes, “Why Do Foreigners Really Statistics, 2004, 86(3). 11372, May 2005, and S. Edwards, ed., Invest in the United States?”, forthcoming as 12 K. J. Forbes and M. Chinn, “A International Capital Flows, 2007, forthcom- an NBER Working Paper. Decomposition of Global Linkages in ing. Also see K. J. Forbes, “Capital Controls:

Tax Policy Towards Energy and the Environment

Gilbert E. Metcalf *

My research over the past several years enue from pollution taxes could be used result, researchers began to identify the gains has focused on the role of taxes and other to reduce other distortionary taxes, thereby from raising revenue via environmental pol- instruments in environmental and energy generating a second “dividend” with a pol- icy instruments (pollution taxes or auction policy. I have focused mainly on instrument lution tax. Some analysts concluded that revenues from cap and trade systems). With design issues in a general equilibrium frame- the existence of this second dividend argued Don Fullerton, I showed that the popu- work, as well as on the distributional implica- for a higher tax on pollution than the first- larly held view that revenue-raising instru- tions of energy and environmental taxation. best Pigouvian prescription, where the tax is ments were preferred to non-revenue-raising set equal to the social marginal damages of instruments focused on the wrong point.2 Environmental Policy pollution. What mattered was whether policies created Bovenberg and deMooij showed that for scarcity rents and whether the government An influential paper by Bovenberg and reasonable consumer preferences the opti- received the rents and used them to lower deMooij touched off a large research agenda mal tax would, in fact, be lower than social other distortionary taxes. on the optimal design of environmental marginal damages. Their insight was that The result — that the second-best tax on taxes in a second-best world with pre-exist- while an environmental tax would enhance pollution was below social marginal damag- ing taxes.1 It had long been understood that efficiency by discouraging pollution, it was es — was troubling to many environmental- taxes on pollution could help to internal- still a distortionary tax and could interact ists who were concerned that it implied that ize pollution externalities. Beginning in the with other distortionary taxes with first- in a world with distortionary taxation more 1980s, analysts began to argue that the rev- best efficiency losses. Building on this initial pollution should be allowed. Such a con-

16 NBER Reporter Winter 2006/7 clusion confuses price and quantity effects. tax or from selling tradable permits) were icy. It may be that consumers do not believe That a first-best price rule (“set pollution recycled by reducing existing taxes, certain that future energy tax increases are credible. taxes equal to social marginal damages”) is European countries might do worse by low- Or, it may be that the publicity effects from modified in the presence of tax distortions ering particular taxes than by giving the investment credits influence consumers’ pur- (“set pollution taxes below social marginal money back in a lump sum. While this is chasing behavior. damages”) does not imply anything about a standard theoretical result from the the- Energy conservation will be an impor- changes in the optimal level of pollution. ory of the second-best, the CGE modeling tant component of any policy to reduce Using a simple analytic general equilibrium results confirm that it is more than an intel- energy consumption and to enhance energy model, I provide a counter-example to show lectual curiosity. security in the United States. The United that having a tax below social marginal dam- In addition to considering the efficiency States already has made impressive gains in ages could be consistent with a higher level effects of environmental taxation, I have stud- how efficiently it uses energy. Energy inten- of environmental quality.3 ied the distributional issues that arise with sity (energy use per dollar of GDP) has The analytic general equilibrium frame- environmental taxation. Many environmen- steadily fallen from a 1917 peak of thirty- work constructed for the research just tal taxes are regressive. For example, a carbon five thousand BTUs per dollar of GDP (year described was easily extended to a consider- tax would raise the price of energy products, 2000 dollars) to a current level of 9.3 thou- ation of monopoly behavior among pollut- products that are necessities in household sand BTUs. In recent research, I document ing firms and instrument design when pol- budgets. I carried out an incidence analysis that roughly two-thirds of this decline can icymakers cannot target pollution directly of a mix of environmental taxes and showed be attributed to improvements in energy but rather must target some proxy for pol- that the taxes, while regressive on an annual efficiency and one-third to changes in the lution.4 The interest in second-best envi- income basis, are less regressive when ana- composition of economic activity in the ronmental policy design was widespread at lyzed on a lifetime income basis.8 This is a United States.12 I also investigate the mech- this point and the NBER co-sponsored a common finding for consumption taxes.9 In anism through which increases in energy conference on environmental policy with addition, I noted that while an environmen- prices affect energy intensity. I find that the FEEM in Italy that Carlo Carraro and I co- tal tax might be regressive, an environmental dominant effect is through energy efficiency organized.5 One of the hotly debated topics tax reform could have any desired degree of rather than through an inducement to shift during this period was whether tradable per- progressivity. In particular, I constructed a from energy-intensive to non-energy-inten- mits for pollution (like those for SO2 trad- tax reform where the revenue from a mix of sive activities. In other words, whatever ing under the Clean Air Act Amendments environmental taxes is used to lower other forces have contributed to a shift towards a of 1990) should be given away or sold. One taxes in a distributionally neutral fashion. service economy in the United States, higher paper from that conference made the impor- More recently, I’ve broadened the discussion energy prices are not among them. tant point that this was not an either-or situ- of how one might use the proceeds from a Energy policy was at the forefront ation; rather, some of the permits could be carbon tax to fund corporate tax integration. of Congressional attention in 2005 when traded and some sold.6 The paper showed In particular, I measure the industry impacts Congress passed the first major energy leg- that only a small portion of permits need be from such a reform.10 islation since 1992. This legislation con- given away in order to preserve the equity tained tax incentives worth $14 billion over value of the energy industries because most Energy Policy a ten-year period. Some of these incentives of the burden of the permit price is passed were extensions of existing initiatives while forward to consumers in the form of higher In addition to work on environmental others were new. I recently reviewed the prices. policy, I have long focused on energy policy new legislation and federal energy tax policy I also have applied insights from the lit- with a particular emphasis on energy conser- more generally.13 In considering tax policy erature on second-best environmental taxa- vation. Early work with Kevin Hassett iden- initiatives towards energy, it is worth not- tion in my research on climate modeling. In tified the impact of energy tax credits for ing the four major reasons for government particular, an empirical analysis of European home conservation improvements.11 That intervention in energy markets: externalities energy and climate policy suggested that research identified an interesting asymme- from energy production and consumption, the benefits from auctioning permits from a try between price policies and investment national security, market failures in energy European carbon cap and trade system vary policies. Consider a conservation invest- conservation, and Hotelling rent expropria- substantially across countries, suggesting the ment that will reduce energy by a known tion on imported oil. Federal energy pol- need for country-specific policy guidance.7 amount over some future period. A govern- icy is not well targeted towards those four That research also showed that when envi- ment policy to double energy prices hence- concerns. I also show in that research that ronmental revenues (either from a carbon forth should have the same impact on the current energy tax policies make clean coal propensity to make this conservation invest- increasingly competitive with pulverized * Gilbert E. Metcalf is a Research Associate in ment as an alternative policy to subsidize coal electricity generating plants. The initia- the NBER’s Program on Public Economics and half the cost of the investment. However, we tives also make wind and biomass competi- a Professor in the Department of Economics at found that the investment subsidy was sub- tive with natural gas electricity generation. Tufts University. stantially more effective than the price pol- Finally, despite the United States being the

NBER Reporter Winter 2006/7 7 third largest producer of petroleum prod- choose how often to visit the park. The gov- Journal of Public Economics 87, 2003: pp. ucts in the world, the federal tax initia- ernment also uses a non-linear income tax 313–22. tives towards energy supply have a negligible to effect redistribution from high-ability to 4 D. Fullerton, I. Hong, and G. E. impact on world supply or prices. low-ability types (where ability is unobserv- Metcalf, “A Tax on Output of the Polluting able). The tax structure relies on a self-selec- Industry is not a Tax on Pollution: The Taxation and Public Pricing tion mechanism to achieve a separating equi- Importance of Hitting the Target”, NBER librium. We show that if the public good is Working Paper No. 7259, July 1999, and A third strand of research focuses on a complement to leisure, then it is optimal in Behavioral and Distributional Effects taxation and public pricing issues more gen- to set a positive price on the public good. of Environmental Policy, C. Carraro and erally. One aspect of that research consid- The higher price on the public good induces G. E. Metcalf, eds., Chicago, University ers the interplay between market structure more labor supply, which discourages high- of Chicago Press, 2001: pp. 13–38; D. and the appropriate form of commodity ability workers from choosing the consump- Fullerton and G. E. Metcalf, “Cap and taxation when firms produce differentiated tion-labor bundle designed for low-ability Trade Policies in the Presence of Monopoly products and thus can exert some degree of workers. In effect, the public good price and Distortionary Taxation”, NBER market power. It has long been known that helps us to discriminate the high from low- Working Paper No. 8901, April 2002, tax policy can substitute for direct regula- ability workers. and Resource and Energy Economics 24, tion to achieve the socially optimal market 2002: pp. 327–47. structure.14 Research with George Norman Summary 5 Behavioral and Distributional Effects suggests that the role of tax policy is more of Environmental Policy, op.cit. nuanced once one allows for more gen- Much of my current research is directly 6 A. L. Bovenberg and L. Goulder, eral market structures and technologies.15 or indirectly focused on the economics of “Neutralizing the Adverse Industry Impacts Whereas the previous literature found that climate change. Climate change is a topic at of CO2 Abatement Policies: What Does It positive ad valorem taxes could help effect the intersection between environmental and Cost?” in Behavioral and Distributional optimal market structure, we find that taxes energy economics and is one of the most dif- Effects of Environmental Policy, pp. 45– may be required under some circumstances ficult issues facing policymakers today. Any 85. and subsidies in other circumstances. The effort to reduce greenhouse gas emissions 7 M. Babiker, G. E. Metcalf, and J. degree of spatial contestability plays a key will require a shift in the forms of energy we Reilly, “Tax Distortions and Global role in determining the sign of the optimal currently use as well as a reduction in overall Climate Policy”, NBER Working Paper tax rate. Once one allows for flexible man- energy consumption. My current research No. 9136, August 2002, and Journal ufacturing technologies, the story changes focuses on how governments can best evalu- of Environmental Economics and considerably. Flexible manufacturing allows ate and design policies to address this criti- Management 46, 2003: pp.269–87. firms to switch product specifications easily cally important problem. 8 G. E. Metcalf, “A Distributional with the result that firms can easily custom- Analysis of Green Tax Reforms”, NBER ize products for consumers. Flexible man- 1 A. L. Bovenberg and R. de Mooij, Working Paper No. 6546, May 1998, and ufacturing can occur in traditional indus- “Environmental Levies and Distortionary National Tax Journal 52(4), 1999: pp. tries (for example, textiles) as well as in the Taxation”, American Economic Review 655–81. Internet based economy. Internet shopping 94, 1994: pp. 1085–9 9 See, for example, J. M. Poterba, provides us the opportunity to get our own 2 D. Fullerton and G. E. Metcalf, “Lifetime Incidence and the Distributional personalized web pages at sites like Amazon. “Environmental Controls, Scarcity Rents, Burden of Excise Taxes”, American It may well be, with some sites, that we also and Pre-Existing Distortions”, NBER Economic Review 79(2), 1989: pp.325– get our own personalized prices. We show Working Paper No. 6091, July 1997, 30, and E. Caspersen and G. E. Metcalf, “Is that with flexible manufacturing, commod- and Journal of Public Economics 80, a Value Added Tax Regressive? Annual ver- ity taxes are now ineffective at helping to (2), 2001: pp. 249–67. We document sus Lifetime Incidence Measures”, National achieve optimal market structure. the history of the double-dividend litera- Tax Journal 47(4), 1994: pp. 731–46. Another example of this research ture in D. Fullerton and G. E. Metcalf, 10 G. E. Metcalf, “Corporate Tax Reform: agenda concerns the optimal pricing of an “Environmental Taxes and the Double Paying the Bills with a Carbon Tax”, excludable public good in the presence of Dividend Hypothesis: Did You Really NBER Working Paper No. 11665, October distortionary taxation. With Jongsang Park, Expect Something For Nothing?”, NBER 2005, forthcoming in Public Finance I posited a model of excludable public goods Working Paper No. 6199, September 1997, Review. where consumers obtain utility based on the and Chicago-Kent Law Review 73(1), 11 K. Hassett and G. E. Metcalf, “Energy amount of the public good provided and 1998: pp. 221–56. Tax Credits and Residential Conservation the number of times the public good is con- 3 G. E. Metcalf, “Environmental Levies Investment: Evidence from Panel Data”, sumed. One example would be an uncon- and Distortionary Taxation: Pigou, NBER Working Paper No. 4020, August gested national park.16 The government Taxation, and Pollution”, NBER Working 1995, and Journal of Public Economics chooses the size of the park and consumers Paper No. 7917, September 2000, and 57, 1995: pp.201–17.

18 NBER Reporter Winter 2006/7 12 G. E. Metcalf, “Energy Conservation in 14 J. A. Kay and M. J. Keen, “How Policy 2(1), 2003. the United States: Understanding its Role Should Commodities Be Taxed?” European 16 G. E. Metcalf and J. S. Park, “A in Climate Policy”, NBER Working Paper Economic Review 23(3), 1983: pp. 339– Comment on the Role of Prices for No. 12272, June 2006. 58. Excludable Public Goods”, NBER Working 13 G. E. Metcalf, “Federal Tax Policy 15 G. E. Metcalf and G. Norman, Paper No. 12535, September 2006, forth- Towards Energy”, NBER Working Paper “Oligopoly Deregulation and the Taxation coming in International Tax and Public No. 12568, October 2006. This paper was of Commodities”, NBER Working Finance. written for the NBER’s 2006 Conference on Paper No. 9415, January 2003, and in Tax Policy and the Economy. Contributions to Economic Analysis and

NBER Profile: Charles H. Dallara

Charles H. Dallara was elected ics from the University of South Carolina to the NBER’s Board of Directors at in 1970 and an honorary doctorate from its Fall 2006 meeting. Dallara, who is the university in 1991. He did gradu- a Managing Director of the Institute of ate work at the Fletcher School of Law International Finance (IIF), will serve as a and Diplomacy, Tufts University, receiv- Director-at-Large. ing his M.A. in 1975, an M.A. in Law and Prior to joining the IIF, Dallara was a Diplomacy in 1976, and a Ph.D. in 1986. Managing Director at J.P. Morgan & Co. Dallara is a member of the Council from 1991 to 1993. During the 1980s, on Foreign Relations and serves on the he served in a variety of senior financial International Advisory Board of the positions in the U.S. government. He was Instituto de Empresa. He is also a member appointed by former President Bush to the of the Board of Overseers of the Fletcher position of Assistant Secretary of the U.S. School of Law and Diplomacy of Tufts Treasury for International Affairs in May University. He resides in Oak Hill, Virginia 1989 and served in that position until June with his wife, Peixin, and has three chil- 1991. dren, Stephen, Emily, and Bryan. A native of Spartanburg, South Carolina, Dallara earned a B.S. in econom-

NBER Reporter Winter 2006/7 9 NBER Profile: Roger W. Ferguson, Jr.

Roger W. Ferguson, Jr. was elected From 1984 to 1997, before becoming a Director-at-Large by NBER’s Board a member of the Federal Reserve Board, of Directors at its Fall 2006 meeting. Ferguson was a Partner at McKinsey & Ferguson is the Chairman of Swiss Re Company, an international management America Holding Corporation and Head consulting firm. From 1981 to 1984, he of Financial Services and member of the was an attorney at the New York City Executive Committee for Swiss Re, a global office of Davis Polk & Wardwell. reinsurance company. He joined Swiss Re Ferguson received his B.A. in econom- in June 2006, after serving as vice chairman ics in 1973, a J.D. in 1979, and a Ph.D. of the Board of Governors of the Federal in economics in 1981, all from Harvard Reserve System. Ferguson initially was University. He is a member of Harvard’s appointed to the Fed Board in November Board of Overseers. 1997 to fill an unexpired term ending Ferguson is married to Annette on January 31, 2000. In July 2001 he was Nazareth and they have two children. sworn in for a full term, but he resigned in 2006 to return to the private sector.

NBER Profile: Kristin J. Forbes

Kristin Forbes is an NBER Research hold this position). She also worked in the Associate in the Program on International U.S. Treasury Department as the Deputy Finance and an Associate Professor of Assistant Secretary of Quantitative Policy Economics at MIT’s Sloan School of Analysis, Latin American and Caribbean Management. She received her B.A., summa Nations. Forbes was recently honored as a cum laude with highest honors, from Williams “Young Global Leader” as part of the World College and her Ph.D. in Economics from Economic Forum at Davos and has won sev- MIT. Her research focuses on policy-related eral MIT teaching awards. questions in international finance and devel- Forbes grew up hiking and skiing in New opment, with recent work on international Hampshire and continues to enjoy trekking capital flows, capital controls, financial mar- and scaling peaks around the world. She also ket contagion, and currency crises. plays tennis and squash. Currently she lives in Over the last few years, Forbes has Boston with her husband, Steven Calhoun, rotated between academia and economic pol- and their new son, Leighton. Leighton icy positions in the U.S. government. She already has taken to hiking, although it may served a term as a Member of the White simply reflect an appreciation of his excellent House’s Council of Economic Advisers vantage point from his parents’ pack. (where she was the youngest person to ever

20 NBER Reporter Winter 2006/7 Conferences

Conference on Retirement Research

The NBER’s Program on Aging held Discussant: Erzo Luttmer, Harvard Alexander Ludwig, University of a Conference on Retirement Research on University and NBER Mannheim; Axel Boersch-Supan, October 20 and 21. The conference orga- University of Mannheim and NBER; nizers were NBER Research Associates Jeffrey R. Brown and Scott J. and Dirk Krueger, Goethe University, Jeffrey R. Brown of University of , Weisbenner, University of Illinois, Frankfurt, “Demographic Change, Urbana-Champaign, and Jeffrey Liebman Urbana-Champaign and NBER, “Who Relative Factor Prices, International of , and David A.Wise Chooses Defined Contribution Plans?” Capital Flows, and Welfare” of Harvard University, who directs the Discussant: Brigitte C. Madrian, Harvard Discussant: James M. Poterba NBER’s Aging Program. University and NBER The NBER has an ongoing grant Andrew Samwick, Dartmouth College from the Social Security Administration, Alan J. Auerbach and Ronald Lee, and NBER, “Changing Progressivity as a as part of the Retirement Research University of California, Berkeley and Means of Risk Protection in Investment- Consortium. The grant has funded anal- NBER, “Notional Defined Contribution Based Social Security” ysis of a wide range of issues related Pension Systems in a Stochastic Context: Discussant: Michael Hurd, RAND to Social Security. Selected papers writ- Design and Stability” Corporation and NBER ten under the grant were presented at Discussant: Jeffrey Liebman the conference, which was also funded Martin S. Feldstein, Harvard University through the grant. The papers were: Andrew Biggs and Clark A. Burdick, and NBER, “Reducing the Risk of Social Security Administration, and Kent Investment Based Social Security”(NBER Jeffrey Liebman, and Emmanuel Saez, Smetters, University of Pennsylvania Working Paper No. 11084) University of California, Berkeley and and NBER, “Pricing Personal Account Discussant: David Wilcox, Federal NBER, “Earnings Responses to Increases Benefit Guarantees: A Simplified Reserve Board in Payroll Taxes” Approach” Discussant: Bruce Meyer, University of Discussant: George Pennacchi, University John Beshears, Harvard University; Chicago and NBER of Illinois at Urbana-Champaign James J. Choi, Yale University and NBER; David Laibson, Harvard James M. Poterba, MIT and NBER; John Geanakoplos, Yale University, and University and NBER; and Brigitte Joshua Rauh, University of Chicago Stephen P. Zeldes, C. Madrian, “The Importance of and NBER; Steven Venti, Dartmouth and NBER, “Facilitating Comparisons Default Options for Retirement Saving College and NBER; and David A. Wise, between DB and DC Systems: Can a Outcomes: Evidence from the United “Reducing Social Security PRA Risk at PRA System Have the Same Features as States” the Individual Level— Lifecycle Funds the Current Social Security System” Discussant: Jeffrey R. Brown and No-loss Strategies” Discussant: Jason Furman, New York Discussant: Douglas Elmendorf, Federal University David M. Cutler and Edward L.Glaeser, Reserve Board Harvard University and NBER, and James M. Poterba, Steven Venti, and Allison B. Rosen, University of Gopi Shah Goda, ; David A.Wise, “The Decline of Defined Michigan, “Trends in Risk Factors in John Shoven, Stanford University and Benefit Retirement Plans and Asset the United States, 1971–5 versus 1999– NBER; and Sita Slavov, Occidental Flows” 2002” College, “Social Security and Medicare: Discussant: Jonathan Skinner, Discussant: Jim Smith, RAND Removing the Disincentives for Long Dartmouth College and NBER Corporation Careers”

Liebman and Saez use SIPP data lyze earnings responses to increases in tax bution of workers around the current matched to longitudinal uncapped earn- rates and to inform discussions about the taxable maximum is inconsistent with an ings records from the Social Security likely effects of raising the Social Security annual model in which people are highly Administration for 1981 to 1999 to ana- taxable maximum. The earnings distri- responsive to the payroll tax rate, even

NBER Reporter Winter 2006/7 21 in the subset of self-employed individu- Higher risk aversion or lower expected income. The authors find that a major- als. Panel data on married men with high returns on stocks raises the expected util- ity of participants fail to make an active earnings display a tremendous increase in ity of portfolios that include less risky decision and are thus defaulted into the earnings over the 1980s and 1990s rela- assets. There often exists a fixed-propor- traditional DB plan after 6 months. They tive to other groups, with no clear breaks tions portfolio of stocks and inflation- also find that those individuals who are around the key tax reforms. This suggests indexed government bonds that yields most likely to be financially sophisti- that other income groups cannot serve as expected utility at retirement that is at cated are most likely to choose the self- a control group for the high earners. This least as high as that from typical lifecycle managed DC plan, despite the fact that, analysis does not support the finding of investment strategies. When asset alloca- given current plan parameters, the DC a large behavioral response to taxation by tion is near the allocation that generates plan is inferior to the portable DB plan wives of high earners. The authors actu- the highest expected utility, variation in under reasonable assumptions about ally find a decrease in the labor supply of expense ratios is more important than future financial market returns. They dis- wives of high earners around both the variation in asset allocation for affecting cuss both rational and behavioral reasons 1986 and the 1993 tax reforms, which retirement utility. that might explain this finding. they attribute to an income effect caused Implicit taxes in Social Security — Around the world, Pay-As-You-Go by the surge in primary earnings at the which measure Social Security contribu- (PAYGO) public pension programs face top. Policy simulations suggest that with tions, net of benefits accrued, as a per- serious long-term fiscal problems primar- an earnings elasticity of 0.5, lost income centage of earnings — tend to increase ily because of actual and projected popu- tax revenue and increased deadweight over the life cycle. Goda, Shoven, and lation aging, and most appear unsustain- loss would swamp any benefits from the Slavov examine the effects of three poten- able as currently structured. Some have increase in payroll tax revenue. In con- tial policy changes on implicit Social proposed the replacement of such plans trast, with an elasticity of 0.2, the ratio Security tax rates: extending the number with systems of fully funded private or of the gain in OASDI revenue to lost of years used in the Social Security for- personal Defined Contribution (DC) income tax revenue and deadweight loss mula from 35 to 40; allowing individuals accounts, but the difficulties of transi- would be much greater. who have worked more than 40 years to tion to funded systems have limited their Poterba, Rauh, Venti, and Wise be exempt from payroll taxes; and distin- implementation. Recently, a new vari- examine how different personal retire- guishing between lifetime low-income ety of public pension program known ment account (PRA) asset allocation earners and high-income earners who as “Notional Defined Contribution” or strategies over the course of a worker’s work short careers. These three changes “Non-financial Defined Contribution” career would affect the distribution of can be achieved in a benefit- and reve- (NDC) has been created, with the objec- retirement wealth and the expected util- nue-neutral manner, and create a pattern tives of addressing the fiscal instabil- ity of wealth at retirement. They con- of implicit tax rates that are much less ity of traditional plans and mimicking sider rules that allocate a constant port- distortionary over the life cycle, elimi- the characteristics of funded DC plans folio fraction to various assets at all ages, nating the high implicit tax rates faced while retaining PAYGO finance. Using as well as “lifecycle” rules that vary the by many elderly workers. The effects different versions of the system recently mix of portfolio assets as the worker of these policies on progressivity and adopted in Sweden, calibrated to U.S. ages. Their analysis simulates retirement women are also examined. demographic and economic parameters, wealth using asset returns that are drawn Brown and Weisbenner provide Auerbach and Lee evaluate the success from the historical return distribution. new evidence on what types of individ- of the NDC approach in achieving fis- The expected utility associated with dif- uals are most likely to choose a defined cal stability in a stochastic context. (In ferent PRA asset allocation strategies, and contribution (DC) plan over a defined a companion paper, these authors will the ranking of these strategies, is sensitive benefit (DB) plan. Making use of admin- consider other aspects of the perfor- to four parameters: the expected return istrative data from the State Universities mance of NDC plans in comparison to on corporate stock; the worker’s relative Retirement System (SURS) of Illinois, traditional PAYGO pensions.) They find risk aversion; the amount of non-PRA they study the decisions of nearly 50,000 that, despite its built-in self-correction wealth that the worker will have avail- new employees who make a one-time, mechanisms, the basic NDC scheme is able at retirement; and the expense ratios irrevocable choice between a traditional still subject to fiscal instability: there charged for the investment. At modest DB plan, a portable DB plan, and an is a high probability that the system’s levels of risk aversion, or in the presence entirely self-managed DC plan. Because debt-payroll ratio will explode over time. of substantial non-PRA wealth at retire- the SURS-covered earnings of these With adjustments, however, the NDC ment, the historical pattern of stock and employees are not covered under the approach can be made considerably more bond returns implies that the expected Social Security system, this choice pro- stable. utility of investing completely in diver- vides insight into the DB versus DC A number of proposals to introduce sified stocks is greater than that from preferences of individuals with regard personal accounts to the Social Security any of the more conservative strategies. to a primary source of their retirement program contain provisions that would

22 NBER Reporter Winter 2006/7 “guarantee” account holders against rel- requires the creation of a new kind of tal, international capital flows, and the atively poor investment performance derivative security (which they call a distribution of welfare in the OECD. that would make their total benefits fall PAAW for Personal Annuitized Average To quantify these effects, they develop below the level scheduled under current Wage security) — it pays its owner one a multi-country large scale Overlapping law. Presently, most of the focus is on inflation-corrected dollar during every Generations model with uninsurable the expected cost of such guarantees, as year of life after the fixed retirement labor productivity and mortality risk. few estimates evaluate the potential mar- date, multiplied by the economy wide Because of the predicted relative abun- ket cost of insuring against the associ- average wage at the retirement date. dance of the factor capital, the rate of ated risk. Biggs, Burdick, and Smetters Redistribution occurs via a variable gov- return falls between 2005 and 2080 by demonstrate how a simple modification ernment match of private contributions. roughly 90 basis points. The simulations of parameter inputs used to calculate the The authors show that by choosing a indicate that capital flows from rapidly expected cost of guarantees would allow particular variable match and restricting aging regions to the rest of the world analysts to estimate the market cost of accounts to hold only PAAWs, it is possi- will initially be substantial, but that the underlying risk. ble to create a system of progressive per- trends are reversed when households de- During most of 2005, the United sonal accounts that exactly mimics the cumulate savings. Capital flows from the States was engaged in a heated debate promised taxes and payouts of the cur- important European countries — France, about whether to replace part of the rent system. They describe how to cre- Germany and Italy are initially posi- current, defined-benefit Social Security ate pools of PAAWs that could be traded tive at 2 percent and then decrease to system with a system of defined contri- in financial markets (yielding market about about -2 percent by 2040. In con- bution personal accounts. A political prices), and then consider allowing indi- trast, the U.S.current account deficit is stalemate has emerged. Democrats who viduals some (limited) flexibility to sell predicted to increase by 2 percentage advocate retaining the current system PAAWs in exchange for other market- points until 2030. In terms of welfare, will not budge from three core goals, able securities. Finally, they discuss ways the model suggests that young agents related to regarding Social Security to modify the system to make it self-bal- with few assets and currently low labor as social insurance: 1) Social Security ancing — reducing or avoiding the need productivity gain up to 1 percent in should redistribute wealth from those for politicians to alter the tax and benefit consumption from higher wages associ- who have earned more over their work- rules affecting participants. They argue ated with population aging. Older, asset- ing lives to those who have earned less; that this would lead to enhanced prop- rich households tend to lose because of 2) different generations should share erty rights and reduced political risk rel- the predicted decline in real returns to the risks of aggregate shocks; and 3) ative to the current system. capital. security should be achieved via infla- Demographic change has an impor- Samwick analyzes the progressivity tion-indexed payments received for life. tant effect on the stock of assets held in of the Social Security benefit formula as Republicans, on the other hand, will defined benefit pension plans. Poterba, a means of lessening the risk inherent in not give up on what they regard as their Venti, and Wise project the impact of investment-based Social Security reform. core goals, namely 1) individual owner- changes in the age structure of the U.S. Focusing on a single cohort of workers ship within Social Security accounts of population between 2005 and 2040 on who will reach their normal retirement tangible assets that cannot be revoked the stock of assets held by these plans. age as the Social Security trust fund is by a future government, 2) market valu- They project the net contributions to, projected to be exhausted, he simulates ation of account assets, 3) transparency and withdrawals from, these plans. These the distribution of benefits subject to regarding accrual of assets, 4) equity-like projections are combined with estimates both earnings and financial risk. The returns, and 5) individual choice of asset of the future evolution of the stock simulations show that the maximally allocation. Geanakoplos and Zeldes of assets in defined contribution plans progressive traditional benefit allows seek a common ground between these to describe the prospective impact of over two thirds of the equity risk to be two approaches that preserves the core demographic change on the stock of eliminated. Similarly, using progressive goals of each. They show that it is possi- retirement saving assets. Information benefit reductions in which replacement ble to convert Social Security into a sys- on demography-linked changes in asset rates for lower earnings are reduced by tem of personal accounts, with irrevoca- demand is a critical input to evaluating less than those for higher earnings, about ble ownership of assets that have market the potential impact of population aging half of the equity risk can be eliminated. prices and market rates of return, while on asset returns. Sensitivity tests show that these patterns at the same time redistributing bene- Boersch-Supan, Krueger, and hold over a wide range of assumptions fits based on lifetime income, sharing Ludwig present the main findings of about risk aversion, the equity premium, risks across generations, and providing recent research on the impact of the and the size of the personal retirement retirees an inflation-indexed life annuity. demographic transition towards an older accounts established by the reform. They call this system progressive personal population in industrialized countries Feldstein describes the risks implied accounts. Implementing these accounts on world-wide rates of return to capi- by a mixed system of Social Security

NBER Reporter Winter 2006/7 23 pension benefits with different combi- niques can raise expected utility relative Cutler, Glaeser, and Rosen com- nations of pay-as-you-go taxes and per- to the plans with no guarantees. The pare the risk factor profile of the popu- sonal retirement account (PRA) saving. ability to do so depends on the individ- lation in the early 1970s with that of the The analysis shows how these risks can ual’s risk aversion level. This underlines population in the early 2000s. They find be reduced by using alternative private the idea that different individuals would that for the population aged 25–74, the market guarantee strategies. The first rationally prefer different investment 10-year probability of death fell from such strategy uses a blend of equities and strategies and risk reduction options. 9.8 percent in 1971–5 to 8.4 percent in TIPS to guarantee at least a positive real Beshears, Choi, Laibson, and 1999–2002. Among those aged 55–74, rate of return on each year’s PRA saving. Madrian summarize the empirical evi- the same measure fell from 25.7 percent The second is an explicit zero-cost collar dence on how defaults affect retirement to 21.7 percent. The largest contribu- that guarantees an annual rate of return savings outcomes. After outlining the tors to these changes were the reduction by giving up all returns above a certain salient features of the various sources of in smoking and better control of blood level. One variant of these guarantees retirement income in the United States, pressure. Despite substantial increases uses a two-stage procedure: a guaran- they present the empirical evidence on in obesity in the past three decades, the teed return to age 66 and then a separate how defaults affect retirement savings overall population risk profile is health- guarantee on the implicit return in the outcomes at all stages of the savings life- ier now than it was. annuity phase. An alternative strategy cycle, including savings plan participa- These papers will be published by provides a combined guarantee on the tion, savings rates, asset allocation, and the University of Chicago Press in an return during both the accumulation and post-retirement savings distributions. NBER Conference volume. Its availabil- the annuity phase. Simulations are pre- They then discusses why defaults have ity will be announced in a future issue of sented of the probability distributions such a tremendous impact on savings the NBER Reporter. of retirement incomes relative to the outcomes. The paper concludes with a “benchmark” benefits specified in cur- discussion of the role of public policy rent law. Calculations of expected util- towards retirement saving when defaults ity show that these risk reduction tech- matter.

Macroeconomics and Individual Decision Making

The NBER’s Working Group on Princeton University and NBER, and Discussant: Eric Zitzewitz, Stanford Macroeconomics and Individual Decision John Morgan, University of California University Making held a conference in Boston on at Berkeley November 4. Working Group Directors Discussant: Petra Geraats, University of “The Rising-Tide Tax System: Indexing George Akerlof, University of California Cambridge (at Least Partially) for Changes in at Berkeley, and Robert Shiller, NBER Inequality” — Leonard Burman and and Yale University, organized the meet- “Why Has CEO Pay Increased So Jeffrey Rohaly, Tax Policy Center, and ing. These topics were discussed: Much?” — Xavier Gabaix, MIT and Robert Shiller NBER, and Augustin Landier, New Discussant: Christopher Foote, Federal “Utility and Happiness” — Miles York University Reserve Bank of Boston and NBER Kimball, University of Michigan and Discussant: George Baker, Harvard NBER, and Robert Willis, University University and NBER “A Cognitive Theory of Identity, of Michigan Dignity, and Taboos” — Roland Discussant: Christopher Hsee, “Coarse Thinking and Benabou, Princeton University and University of Chicago Persuasion” — NBER, and Jean Tirole, IDEI and and Andrei Shleifer, Harvard GREMAQ, Toulouse “Leadership in Groups: A Monetary University and NBER, and Joshua Discussant: Robert Oxoby, University of Policy Experiment” — Alan S. Blinder, Schwartzstein, Harvard University Calgary

Psychologists have developed effec- relationship between how happy a per- ics either ignores happiness data or tive survey methods of measuring how son feels and utility is an unresolved assumes that felt happiness is more or happy people feel at a given time. The question. Existing work in econom- less the same thing as flow utility. The

24 NBER Reporter Winter 2006/7 approach Kimball and Willis propose smaller groups by a very small margin. indexation would insure, at least par- steers a middle course between the two Finally, they successfully replicate their tially, against future increases in after- polar views that “happiness is irrele- Princeton results, at least qualitatively: tax inequality. Tax rates would endog- vant to economics” and the view that Groups perform better than individu- enously adjust to changes in inequality. “happiness is a sufficient statistic for als and they do not require more “time” Burman, Rohaly, and Shiller develop utility.” They argue that felt happi- to do so. a method of implementing the system ness is not the same thing as flow util- Gabaix and Landier develop a sim- using U.S. tax returns data and the Tax ity, but that it does have a systematic ple equilibrium model of CEO pay. Policy Microsimulation Model. They relationship to utility. In particular, CEOs have different talents and are study the outcomes if inequality index- they propose that happiness is the sum matched to firms in a competitive ation had begun in 1979, or 1994, and of two components: 1) elation — or assignment model. In market equilib- describe the distributive and incentive short-run happiness — which depends rium, a CEO’s pay changes one-for- effects. on recent news about lifetime utility; one with aggregate firm size, while Benabou and Tirole analyze social and 2) baseline mood — or long-run changing much less with the size of his and economic phenomena involving happiness — which is the output of own firm. The model determines the beliefs that people value and invest a household production function like level of CEO pay across firms and over in. Uncertain about their “deep val- the household production functions time, offering a benchmark for calibrat- ues”, agents infer them from their own for health, entertainment, and nutri- able corporate finance. The six-fold choices, which then come to define tion. Because happiness is itself one of increase in CEO pay between 1980 and “who they are”. Identity investments the arguments of the utility function, 2003 can be fully attributed to the six- increase in unfamiliar settings or when the determinants of happiness affect fold increase in market capitalization a greater endowment (wealth, career, behavior. Moreover, because happiness of large U.S. companies during that family, culture) raises the stakes on depends on recent news about lifetime period. The authors find a very small viewing an asset as valuable (escalating utility, short-run movements in happi- dispersion in CEO talent, which none- commitments). Taboos against certain ness data provide important informa- theless justifies large pay differences. transactions or their mere contempla- tion about preferences. This theory of The data broadly support the model. tion arise to protect beliefs about the the relationship between utility and The size of large firms explains many of “pricelessness” of certain assets (life, happiness provides a new explanation the patterns in CEO pay, across firms, freedom, love, faith) or things one for the Easterlin paradox of secularly over time, and between countries. “would never do”. Whether such behav- non-increasing happiness as the conse- Mullainathan, Schwartzstein, and iors are welfare-enhancing or reducing quence of the Baumol cost disease for Shleifer present a model of coarse depends on whether beliefs are sought happiness and an explanation for loss- thinking, in which individuals group for a functional value (self-disciplines, aversion based on the dependence of situations into categories, and transfer sense of direction) or as affective con- happiness on recent news. information from situations in a cat- sumptions (self-esteem, anticipatory In an earlier paper, Blinder and egory where it is useful to situations feelings). Thus, a “hedonic treadmill” Morgan created an experimental appa- where it is not. The model explains may arise, or competing identities may ratus in which Princeton University how uninformative messages can be cause dysfunctional failures to assimi- students acted as ersatz central bank- persuasive, particularly in low involve- late, acquire new skills, or adapt to glo- ers, making monetary policy decisions ment situations, and how objectively balization. In social interactions, norm both as individuals and in groups. In informative messages can be dropped violations trigger forceful reaffirma- this study, they manipulate the size and by the persuader without the audience tion, exclusion, or harassment when leadership structure of monetary policy assuming the worst. The model sheds threatening a strongly held identity, but decisionmaking. They find no evidence light on several aspects of mutual fund further erode an initially weak morale. of superior performance by groups that advertising. Dignity, pride, or wishful thinking lead have designated leaders. Groups with- Based on experience over the to inefficient breakdowns of bargain- out such leaders do as well as or bet- past three decades, growing inequal- ing even under symmetric information, ter than groups with well-defined lead- ity appears to be a serious risk. A as partners seek to self-enhance and ers. Furthermore, they find rather little change in the tax system to index shift blame by rejecting “insultingly difference between the performance of against changes in inequality is moti- low” offers. four-person and eight-person groups; vated both by financial theory and by the larger groups outperform the classical welfare economics. Inequality

NBER Reporter Winter 2006/7 25 Inter-American Seminar on Economics: Informality, Corruption, and Institutions The NBER and FEDESARROLLO Fasecolda, Colombia Andes, Colombia, and Mauricio co-sponsored this year’s Inter-American Cardenas, “Informality in Colombia: Seminar on Economics (IASE) on William F. Maloney and Edwin Goni, The Case of Child Labor” “Informality, Corruption, and Insti­ , and Mariano Bosch, Comments: Stefania Scandizzo, CAF, tutions.” The conference, which was held London School of Economics, “The Venezuela in Bogota, Colombia, on December 1 Determinants of Rising Informality in and 2, was organized by NBER Research Brazil: Evidence from Gross Worker Ernesto Schargrodsky, Universidad Associate Sebastian Edwards of Uni­ Flows” Torcuato Di Tella, and Sebastian versity of California, Los Angeles, and Comments: Maria Laura Alzua, IERAL Galiani, Washington University, St. Mauricio Cardenas, FEDESARROLLO, de Fundacion Mediterranea, Argentina Louis, “Property Rights of the Poor: Colombia. The following papers were Effects of Land Titling” discussed: Hugo Maul, Universidad Francisco Comments: Daniel Mejia, Banco de la Marroquin, Guatemala, “From Penny Republica, Colombia Alberto Alesina, Harvard University Capitalism to Global Markets: The Case and NBER, and Guido Tabellini, of the Guatemalan Informal Sector” Jennifer Hunt, McGill University and IGIER, Bocconi, “Why is Fiscal Policy Comments: Pablo Acosta, CAF, NBER, “ Bribery in Health Care in Peru often Procyclical? Venezuela and Uganda” Comments: Leonardo Villar, Banco de Comments: Mauricio Olivera, la Republica, Colombia Pablo Fajnzylber and William F. Fedesarrollo, Colombia Maloney, World Bank, and Gabriel Norman V. Loayza and Jamele V. Montes, University of Illinois, Eric V. Edmonds, Dartmouth College Rigolini, World Bank, “Informality Champaign-Urbana, “ Does Formality and NBER, and Salil Sharma, Trends and Cycles” Improve Micro-Firm Performance? Dartmouth College, “Institutional Comments: Marcela Melendez, Quasi-Experimental Evidence from the Influences on Human Capital InterAmerican Development Bank Brazilian SIMPLES Program” Accumulation: Micro Evidence from Comments: Maurice Kugler, Harvard Children Vulnerable to Bondage” Sebastian Edwards, “Capital Controls, University Comments: Alejandro Gaviria, Contagion, and Capital Flows” Universidad de los Andes, Colombia Comments: Roberto Junguito, Raquel Bernal, Universidad de los

Many countries, especially devel- ment. They first present a theoretical to formal firms, in the form of more oping ones, follow procyclical fis- model in which the size of informal rigid business regulations, less valuable cal polices; that is, spending goes up employment is determined by the rela- police and judicial services, and weaker (taxes go down) in booms and spend- tive costs and benefits of informality monitoring of informality. In the short ing goes down (taxes go up) in reces- and the distribution of workers’ skills. run, informal employment is found sions. Alesina and Tabellini provide In the long run, informal employment to be countercyclical for the majority an explanation for this suboptimal fis- varies with the trends in these vari- of countries, with the degree of coun- cal policy based upon political distor- ables, and in the short run it reacts to ter cyclicality being lower in coun- tions and incentives for a less-than- accommodate transient shocks and to tries with larger informal employment benevolent government to appropriate close the gap that separates it from its and better police and judicial services. rents. Voters have incentives similar trend level. They then use an error-cor- Moreover, informal employment fol- to the “starving the Leviathan” clas- rection framework to examine empiri- lows a stable, trend reverting process. sic argument, and demand more pub- cally informality’s long-and short-run These results are robust to changes in lic goods or fewer taxes to prevent relationships. For this purpose, they use the sample and to the influence of out- governments from appropriating rents country-level data at annual frequency liers, even when only developing coun- when the economy is doing well. They for a sample of developed and devel- tries are considered in the analysis. test this argument against more tra- oping countries, with the share of self- Edwards investigates whether ditional explanations based purely on employment in the labor force as the restrictions to capital mobility reduce borrowing constraints, with a reason- proxy for informal employment. They countries’ vulnerability to major exter- able amount of success. find that, in the long run, informality nal shocks. More specifically, he asks if Loayza and Rigolini study the is larger in countries that have lower countries that restrict the free flow of trends and cycles of informal employ- GDP per capita and impose more costs international capital have a lower prob-

26 NBER Reporter Winter 2006/7 ability of experiencing a sudden stop remainder seems driven by the rising on firm performance. They find that and being subject to contagion than labor costs and reduced flexibility aris- newly created firms that opt for oper- countries with a freer degree of capi- ing from Constitutional reform. ating formally use more paid workers, tal mobility. Edwards uses three new Maul deals with the processes of are more capital intensive, and exhibit indexes on the degree of international integrating into the global markets an higher levels of total factor productiv- financial integration and a large multi- economy with a large informal sec- ity. Increased access to credit and gov- country dataset for 1970–2004 to esti- tor. He argues that Guatemala’s XXI ernment provided technical assistance mate a series of random-effect probit century informal economy can be is not responsible for more than a small equations. The results from these pro- described as a “Penny Capitalism” sys- fraction of those formality effects. bit equations are used to compute mar- tem, a term coined by Sol Tax, and Rather, the observed greater willing- ginal effects of different indicators on later used by T. Schultz, to describe ness of formal firms to operate out of the likelihood of a country facing a Guatemalan’s indigenous communities a fixed locale appears to be responsible major external crisis. Edwards also ana- in the 1930s. This capitalist system for a large share of the formality-firm lyzes the role played by other variables works on a “microscopic scale,” and performance link. Further, the impact in determining the probability of expe- is characterized by limited access to seems largest on poorly performing riencing a sudden stop, including large larger markets, technology, and credit, firms. current account deficits; the exchange but highly efficient on very small scale. Bernal and Cardenas study the rate regime — fixed or flexible; hold- Most of the literature reduces informal- magnitude and nature of child labor ings of international reserves; fiscal ity to a single dimension (labor market, in Colombia. In particular, they eval- imbalances; world interest rates; and property rights, or Micro and Small uate the factors that determine joint the degree of dollarization, among oth- Enterprises phenomenon). In contrast, child labor and school attendance deci- ers. The most important finding is that in this paper Maul defines economic sions within the household. In addi- countries with greater capital mobil- informality as a process of individual tion, they investigate the characteristics ity are somewhat more subject to con- adaptation to a spectrum of regulatory of children, parents, and households tagion than countries that restrict the contracts that the political authorities that are associated with certain types free mobility of capital. impose on workers and business firms. of child labor, for example, work with Maloney, Goni, and Bosch study In this context, and given the ubiqui- relative versus work with non-relatives, gross worker flows to explain the ris- tous nature of the informal economy, and work in certain economic sectors. ing informality in Brazilian metropol- if globalization means the integration The authors use various sources of itan labor markets from 1983–2002. of formal firms to international trade, data including the Child Labor Survey This period covers two economic then for poor underdeveloped econ- (2001), Child Labor Module Follow- cycles, several stabilization plans, a omies, such as Guatemala, there is a up included in the National Household far-reaching trade liberalization, and big challenge coming. Reducing infor- Survey (2003), the Living Standards changes in labor legislation through mality in such an institutional envi- Survey (2003) and the Familias en the Constitutional reform of 1988. ronment implies much more than just Accion database (2002-3). Total child Focusing first on cyclical patterns, they reducing the costs and increasing the labor participation, which includes all confirm Bosch and Maloney’s (2006) benefits of becoming formal. In such a working children from 5 to 17 years findings for Mexico that the patterns context, limiting the access to national, of age, was approximately 14.5 percent of worker transitions between formal- regional, and global markets to “Doing in 2001. Approximately two-thirds of ity and informality correspond primar- Business” type of firms will seriously this group also attended school. Child ily to the job-to-job dynamics observed curtail the options for the country, as labor participation rates in rural areas in the United States and not to the tra- whole, and for the great majority of the were twice as large as those in urban ditional idea of informality constitut- labor force that operates in the infor- areas. Around two-thirds of working ing the inferior sector of a segmented mal economy, especially women, young children worked for their parents or market. However, they also confirm people and indigenous communities. other relatives. The majority of chil- distinct cyclical patterns of job find- Fajnzylber, Maloney, and Montes dren work in the agriculture sector ing and separation rates that lead to the use Regression Discontinuity methods (approximately 38 percent) followed informal sector absorbing more labor to identify the impact of the reduction by retail, manufacturing, and services. during downturns. Second, focusing on of registration costs and taxes on newly Bernal and Cardenas estimate a model secular movements in gross flows and born Brazilian micro firms. The intro- in which child labor and school atten- the volatility of flows, they find the rise duction of the SIMPLES program in dance are simultaneous decisions. in formality to be driven primarily by 1996 provides a quasi-natural experi- Higher educational attainment of the a reduction in job finding rates in the ment that permits them to eliminate head of the household, older head of formal sector. A small fraction of this many of the endogeneity issues sur- household, and higher adult employ- is driven by trade liberalization, and the rounding the impact of registration ment rates within the household are

NBER Reporter Winter 2006/7 27 all positively and significantly associ- pensation, while others are still dis- Hunt does not find evidence that the ated with higher probabilities of child puting the compensation payment in public health care sector in either Peru labor. In addition, the probability that the slow Argentine courts. These dif- or Uganda is able to price-discriminate a child works increases for those in ferent decisions by the former own- less effectively than public institutions the lowest income quintile, for chil- ers generated an allocation of property with less competition from the private dren living in larger households and rights that is exogenous in equations sector. living with extended family. The most describing the behavior of the squat- How do weak private property insti- vulnerable children (for example, eth- ters. The authors find that entitled tutions influence human capital invest- nic minorities, with low-educated par- families increased housing investment, ment decisions? Empirical challenges ents, living in very poor households, reduced household size, and improved have limited research on this poten- and the like) are more likely to work in the education of their children relative tially important facet of how institu- agriculture, whereas less disadvantaged to the control group. However, effects tions influence prosperity. Edmonds children (that is, those with highly on credit access are modest and there and Sharma argue that a debt-bond- educated parents, higher adult employ- are no effects on labor income. age institution prevalent in the western ment rates within the household, male Hunt examines the role of house- plains of Nepal is an unusually good head of household, and so on) are more hold income in determining who bribes setting in which it is possible to explore likely to work in retail and in particu- and how much they bribe in health how institutions affect investments in lar, in family-owned businesses. Finally, care in Peru and Uganda. She finds education. They observe substantially the authors find robust evidence that that rich patients are more likely than more child labor, lower schooling atten- a conditional cash transfer program, other patients to bribe in public health dance and attainment, and significantly Familias en Accion, has had significant care: doubling household consump- elevated fertility in families vulnerable effects on child labor, especially in the tion increases the bribery probability to debt-bondage. The data are most case of boys. by 0.2–0.4 percentage points in Peru, consistent with diminished returns to Secure property rights are consid- compared to a bribery rate of 0.8 per- education in the vulnerable population ered a key determinant of economic cent; doubling household expendi- as an explanation for the findings. They development. However, the evaluation ture in Uganda increases the bribery argue that this diminished return to of the causal effects of property rights probability by 1.2 percentage points education owes to a substantive expro- is a difficult task, as their allocation compared to a bribery rate of 17 per- priation threat. That is, the absence of is typically endogenous. Schargrodsky cent. The income elasticity of the bribe secure property rights over human cap- and Galiani exploit a natural experi- amount cannot be precisely estimated ital’s output is a significant deterrent to ment in the allocation of land titles to in Peru, but is about 0.37 in Uganda. educational investments. overcome this identification problem. Bribes in the Ugandan public sector It is anticipated that these pro- More than twenty years ago, a group appear to be extorted from the richer ceedings will be published in an NBER of squatters occupied a piece of land in patients amongst those exempted by conference volume. Its availability will a poor suburban area of Buenos Aires. government policy from paying the be announced in a future issue of the When the Congress passed a law expro- official fees, and reflect the same com- NBER Reporter. priating the land from the former own- bination of fee for service and price ers with the purpose of entitling it to discrimination as official fees. Bribes in the occupants, some of the original the private sector are flat-rate fees paid owners accepted the government com- by patients who do not pay official fees.

28 NBER Reporter Winter 2006/7 Financial Reporting and Taxation

The NBER held a conference on John R. Graham, Duke University and MIT, “New Evidence on the Importance “Financial Reporting and Taxation” in NBER, and Lillian F. Mills, University of Deferred Tax Assets and Liabilities Cambridge on December 7. Douglas of Texas, “Using Tax Return Data to and on Managerial Manipulation of Tax Shackelford, NBER and University of Simulate Corporate Marginal Tax Rates Expense” North Carolina, Chapel Hill, organized Discussants: Oliver Li, University Discussants: James R. Hines, University this program: of Notre Dame, and Clemens Sialm, of Michigan and NBER, and Andrew University of Michigan and NBER Schmidt, Columbia University Michelle Hanlon, University of Michigan, and Edward Maydew, George A. Plesko, University of Leslie A. Robinson and Richard University of North Carolina, Connecticut, “Estimates of the Sansing, Dartmouth College, “Tax “Book‑Tax Conformity: Implications for Magnitude of Financial and Tax Incentives versus Financial Reporting Multinational Firms” Reporting Conflicts” Costs: The Case of Internally Developed Discussants: Mihir A. Desai, Harvard Discussants: Raj Chetty, University of Intangible Assets” University and NBER, and Peter California, Berkeley and NBER, and Discussants: Christian Leuz, University Merrill, PricewaterhouseCoopers David Weisbach, University of Chicago of Chicago, and Emmanuel Saez, University of California, Berkeley and Douglas Shackelford; Joel B. Slemrod, Michelle Hanlon, University of NBER University of Michigan and NBER; and Michigan, and Joel B. Slemrod, “What James Sallee, University of Michigan, Does Tax Aggressiveness Signal? Jennifer L. Blouin and Irem Tuna, “A Unifying Model of How Taxes Affect Evidence from Stock Price Reactions to University of Pennsylvania, “Tax the Real and Accounting Decisions of News About Tax Aggressiveness” Contingencies: Cushioning the Blow to Corporations” Discussants: Joseph Bankman, Stanford Earnings?” Discussants: Alan J. Auerbach, University, and Dhammika Dharmapala, Discussants: Mary Margaret Frank, University of California, Berkeley and University of Connecticut University of Virginia, and Thomas NBER, and Terry Shevlin, University of Neubig, Ernst & Young Washington James M. Poterba, MIT and NBER, and Nirupama Rao and Jeri Seidman,

Hanlon and Maydew examine the als. The authors describe several possi- reporting. Accountants usually focus on implications for multinational firms of ble approaches to implementing book- the coordination of tax and account- recent proposals to conform tax and tax conformity for firms that have both ing choices, with little attention to real financial reporting (that is, book-tax con- domestic and foreign operations. They economic effects. Both approaches yield formity). Proponents of book-tax confor- discuss issues likely to arise with each important insights, but neither is com- mity argue that the current dual system in approach and conjecture at the behavioral plete because real and accounting deci- the United States allows firms to simul- responses to each. Using firm-level finan- sions interact in important ways. As a taneously manage their taxable income cial data from Compustat, they simulate result, studies of real (accounting) choices down while managing their book income the effects of book-tax conformity on pub- that ignore the incentives and constraints upward. By requiring book-tax confor- licly traded U.S. firms. Specifically, they that affect accounting (real) decisions mity, they contend that firms will be simulate the effects of book-tax confor- may misinterpret behavioral responses to forced to trade-off reporting high earn- mity on the level and variability of book taxation. Therefore, a unifying framework ings numbers to shareholders and report- earnings and tax payments/collections. that incorporates both real and account- ing low earnings to the taxing authority, Shackelford, Slemrod, and Sallee ing choices is needed for comprehen- resulting in improved financial report- model the impact of taxes on both the sive treatment of the behavioral responses ing and less tax avoidance. Reduced com- real decisions and the accounting choices to taxation. The authors propose that pliance costs and easier auditing also of firms. Their goal is to enhance under- corporations make real and accounting have been cited as potential benefits of standing of the behavioral effects of taxa- decisions to maximize a function whose book-tax conformity. Aspects of book- tion by merging the approach to the study arguments include the present value of tax conformity that have not been exam- of corporate taxation taken by econo- the after-tax cash flows and also after- ined, however, include its international mists and accountants. typi- tax book income in each of two periods. implications, particularly regarding the cally study the effects of taxes on real Corporations have limited discretion to foreign operations of U.S. multination- decisions, ignoring the role of financial shift across periods pre-tax book income,

NBER Reporter Winter 2006/7 29 taxable income, and the book tax provi- takes advantage of the contemporaneous suggesting that the negative reaction to sion. Some real decisions are more attrac- nature of reporting to mitigate the econo- tax shelter news is not predominantly a tive than others because they provide metric problems identified in earnings reputation effect. All in all, their analysis managers with discretion over the tim- management studies. He finds that the suggests that tax shelter news is viewed as ing of taxable income and book income; extent that tax reporting reflects discre- a negative event by the market, although this is especially true for companies for tionary financial reporting varies dramati- the stock price reaction is much smaller which this discretion has relatively high cally by industry, profitability, and the than the reaction to major accounting value. The authors show how the -pres sign of discretionary accruals measured mishaps. ence of discretion modifies the optimal under the tax system. Further, focusing on Poterba, Rao, and Seidman collect decisions of firms, in theory, and provide tax reporting, he finds that managers are data from the tax footnotes of a sample examples that illustrate this behavior in able to undertake tax reducing activities of large U.S. corporations between 1994 the real world. The source of this discre- with less of an effect on financial report- and 2004 and use it to investigate two tion may be either the accounting rules or ing than tax increasing accruals, consis- issues concerning financial accounting for the tax law. It is critical to know whether tent with recent evidence on the differen- taxes. They document the importance of the book and tax accounting either must tial growth of book and tax income, and deferred tax assets and liabilities for their by law be conformed, or will be con- with tax avoidance activities. sample firms and demonstrate the sub- formed by choice because of the private Hanlon and Slemrod study the stock stantial heterogeneity in firm tax posi- costs of maintaining separate accounting price reaction to news about tax aggres- tions. In 2004, 47 firms in the sample of systems. When they are not conformed, siveness. They find that, on average, a com- 71 reported net deferred tax assets and divergence between the two may be costly pany’s stock price declines when there is 24 reported net deferred tax liabilities. In to the extent that it alerts the IRS to pos- news about its involvement in tax shelters, this sample, total deferred tax assets for sibly aggressive tax planning or the capital but the reaction is small relative to reac- firms with such assets in 2004 were $57.6 markets to poor earnings quality. tions to other corporate misdeeds. They billion, while total deferred tax liabilities Graham and Mills simulate marginal find some limited evidence for cross-sec- for firms with such liabilities were $212.8 tax rates (MTRs) from 1998 to 2000 tional variation in the reaction. For exam- billion. This implies that a 5 percentage using U.S. tax return data for public cor- ple, the stock price decline is smaller for point decline in the federal statutory cor- porations. They compare the tax-return companies with relatively high effective porate tax rate would reduce net income tax rates to tax rates calculated from pub- tax rates, possibly because news about tax by roughly $8.2 billion at sample firms lic financial statement data (Compustat) aggressiveness is more favorably viewed with net deferred tax assets, because these and find that Graham’s (1996a) simu- for those firms where public informa- assets would decline in value and this in lated tax rate is the book variable most tion would suggest otherwise. The stock turn would reduce net income. Firms highly correlated with the tax-return vari- price decline is also smaller for firms that differ substantially in the composition able. They also find that the correlation have good governance, which is consis- of deferred tax assets and liabilities. The between book and tax MTRs improves tent with the idea that, for these firms, authors demonstrate this by disaggregat- substantially for firms with similar con- the news is less likely to trigger concern ing deferred tax accounts for their sample solidated groups. They identify ways about insiders’ aggressiveness toward the firms. They also explore the role of mana- to improve upon Compustat MTRs in investors themselves. Indeed, these results gerial discretion in reporting tax expense. terms of more closely approximating tax- suggest that for well-governed firms with They find evidence of tax management return based tax rates. Finally, they find especially high effective tax rates, news when firms will otherwise miss earnings that tax return MTRs are significantly that they have been involved in a tax shel- targets, and extend prior research by ana- correlated with financial statement cor- ter is received favorably by the market. lyzing which components of tax expense porate debt ratios, although less so than The reaction is more negative for firms are most likely to be used for earnings the correlation between book MTRs and in the retail sector, suggesting that part management. debt ratios. of the reaction may be a consumer/tax- Robinson and Sansing develop and Plesko examines the tax report- payer backlash. Hanlon and Slemrod also analyze a model in which tax consider- ing consequences of financial report- explore the stock price reaction to reports ations and financial reporting consider- ing discretion. Using a matched sample of effective tax rate calculations released ations have countervailing effects on a of financial statements with tax returns, by Citizens for Tax Justice. They hypoth- firm’s investments in internally developed he estimates the accuracy of tax return esize that these reports signal tax aggres- intangible assets. They also propose and information inferred from financial state- siveness without the implications for tax estimate a new measure of tax preferences, ments. To examine the trade-offs between penalties or illegal behavior that tax shel- the economic effective tax rate. This mea- financial and tax reporting, he models ter news carries, and therefore any mar- sure reflects both investments in intan- the relation that discretionary financial ket reaction represents a pure reputation gible assets and the use of debt financing, accounting accruals have to discretion- effect. They find no statistically - signifi neither of which generates book-tax dif- ary federal tax accruals. The methodology cant stock price reaction to the reports, ferences. Their measure indicates that the

30 NBER Reporter Winter 2006/7 economic effective tax rate was about 16 contingencies (aka tax cushion). A recent option incentive pay as a proportion of percent between 1988 and 2005, when call for corporate tax reform has high- total compensation and larger implicit the statutory tax rate was either 34 or 35 lighted the disparity between financial claims. Finally, these findings are consis- percent. On average, about two-thirds of and income tax reporting. In this paper, tent with firms asymmetrically reporting the difference between their measure and the authors create a broad-based mea- good news, providing additional evidence the statutory tax rate is attributable to sure of a cushion that appears to capture that firms strategically report non-recur- intangible assets and about one-third is cross-sectional variation in tax aggressive- ring income components (Schrand and attributable to the use of debt financing. ness. After controlling for tax aggressive- Walther 2000). Overall, the findings sup- Both the effect of intangible assets and ness, they find some evidence that firms port the need for FIN48, which attempts the use of debt financing on their measure appear to be using cushions to smooth to improve conformity in the reporting of vary across industries. earnings. Specifically, a tax cushion is used tax contingencies. Blouin and Tuna study firms’ tax to smooth earnings by firms with larger

Behavioral Responses to Taxation and Social Insurance Programs

An NBER-Universities Research Jennifer Huang, University of Anil Kumar, Federal Reserve Bank Conference on “Behavioral Responses Texas at Austin; Gene Amromin, of Dallas, and Gary V. Engelhardt, to Taxation and Social Insurance Federal Reserve Bank of Chicago; Syracuse University, “Employer Programs” took place in Cambridge on and Clemens Sialm, University of Matching and 401(k) Saving: Evidence December 8–9. Organizers Raj Chetty Michigan and NBER, “The Tradeoff from the Health and Retirement Study” and Emmanuel Saez, both of NBER between Mortgage Prepayments and Discussant: John Karl Scholz, University and University of California, Berkeley, Tax‑Deferred Retirement Savings” of Wisconsin, Madison and NBER chose these papers for discussion: Discussant: Brigitte Madrian, Harvard University and NBER Adam Looney, Federal Reserve Board, Anton Korinek, Columbia University, “Trading Tax Benefits for Child and Joseph E. Stiglitz, Columbia Nicole Maestas, RAND Corporation, Support” University and NBER, “Dividend and Dana Goldman, NBER and Discussant: Melissa Kearney, The Taxation and Intertemporal Tax RAND Corporation, “Medical Brookings Institution and NBER Arbitrage” Expenditure Risk and Household Discussant: James M. Poterba, MIT and Portfolio Choice” , Yale University, and NBER Discussant: , MIT and John A. List, NBER and University NBER of Chicago, “Does Price Matter in John D. Wilson, Michigan State Charitable Giving? Evidence from a University, and Joel B. Slemrod, James P. Ziliak, University of Kentucky, Large‑Scale Natural Field Experiment” University of Michigan and NBER, “Taxes, Transfers, and the Labor Supply Discussant: Lise Vesterlund, University “Tax Competition with Parasitic Tax of Single Mothers” of Pittsburgh Havens” Discussant: Bradley Heim, Department Discussant: James R. Hines, University of the Treasury of Michigan and NBER

Korinek and Stiglitz develop a life- internally growing and mature firms, as allow firms to engage in inter-temporal cycle model of the firm to analyze the postulated by the new view of dividend tax arbitrage so as to reduce investors’ effects of dividend tax policy on aggre- taxation. Since aggregate investment is tax burden. This can significantly distort gate investment. They find that new firms dominated by these latter two categories, aggregate investment. Anticipated tax raise less equity and invest less the higher the level of dividend taxation, as well as cuts (increases) delay (accelerate) firms’ the level of dividend taxes, in accordance unanticipated changes in dividend tax dividend payments, which leads them to with the traditional view of dividend tax- rates, have only a minor impact on aggre- hold higher (lower) cash balances and, ation. However, the dividend tax rate is gate investment and output. Anticipated for capital constrained firms, can signifi- irrelevant for the investment decisions of dividend tax changes, on the other hand, cantly increase (decrease) aggregate invest-

NBER Reporter Winter 2006/7 31 ment for periods after the tax change. be driven by liquidity considerations and ment is 1.3), but a small effect on hours of Furthermore, the authors show that the that self-reported debt aversion and risk work (average compensated wage elastic- analysis of dividend taxation in a contest- aversion variables explain, to some extent, ity of 0.16) unless the wage change also able democracy has to take into account the preference for paying off debt obliga- alters the mother’s decision to partici- expectations about future regime changes tions early and hence the propensity to pate in AFDC/TANF, food stamps, or and the ensuing dividend tax changes. forgo the proposed tax arbitrage. SSI. These estimates are consistent with a This can significantly change the evalua- As health care costs continue to rise, recent theoretical model by Saez (2002) tion of a given dividend tax policy. medical expenses have become an increas- that suggests that the optimal transfer pol- Slemrod and Wilson develop a tax ingly important contributor to finan- icy is one that involves a modest income competition framework in which some cial risk. Economic theory suggests that guarantee for non-workers coupled with jurisdictions, called tax havens, are para- when background risk rises, individuals subsidies for low-income workers much sitic on the revenues of other countries. will reduce their exposure to other risks. like the current EITC program. The havens use real resources to help com- Maestas and Goldman present a test of Employer matching of employee panies camouflage their home-country this theory by examining the effect of 401(k) contributions can provide a pow- tax avoidance, and countries use resources medical expenditure risk on the willing- erful incentive to save for retirement in an attempt to limit the transfer of ness of elderly Medicare beneficiaries to and is a key component in pension-plan tax revenues to the havens. The equilib- hold risky assets. They measure exposure design in the United States. Using detailed rium price for this service depends on the to medical expenditure risk by whether administrative contribution, earnings, demand and supply for such protection. an individual is covered by supplemental and pension-plan data from the Health Recognizing that taxes on wage income insurance through Medigap, an employer, and Retirement Study, Englehardt and are also evaded, the authors solve for or a Medicare HMO. They account for Kumar formulate a life-cycle-consistent the equilibrium tax rates on mobile capi- the endogeneity of insurance choice by econometric specification of 401(k) sav- tal and immobile labor, and demonstrate using county variation in Medigap prices ing and estimate the determinants of sav- that the full or partial elimination of tax and non-Medicare HMO market pen- ing, accounting for non-linearities in the havens would improve welfare in non- etration. They find that having Medigap household budget set induced by match- haven countries, in part because countries or an employer policy increases risky asset ing. The participation estimates indicate would be induced to increase their tax holding by 6 percentage points relative to that an increase in the match rate by 25 rates, which they have set at inefficiently those enrolled in only Medicare Parts A cents per dollar of employee contribu- low levels in an attempt to attract mobile and B. HMO participation increases risky tion raises 401(k) participation by 3.75 capital. They also demonstrate that the asset holding by 12 percentage points. to 6 percentage points, and the estimated smaller countries choose to become tax Given that just half of their sample holds elasticity of participation with respect havens, and show that the abolishment risky assets, these are economically sizable to matching ranges from 0.02–0.07. The of a sufficiently small number of the rela- effects. It also suggests an important link parametric and semi-parametric estimates tively large havens leaves all countries bet- between the availability and pricing of for saving indicate that an increase in ter off, including the remaining havens. health insurance and the financial behav- the match rate by 25 cents per dollar of Amromin, Huang, and Sialm show ior of the elderly. employee contribution raises 401(k) sav- that a signicant number of households can How wages and non-labor income ing by $400–$700 (in 1991 dollars). The perform a tax arbitrage by cutting back affect both the decision to work and estimated elasticity of 401(k) saving to on their additional mortgage payments hours of work among single mothers is matching is also small and ranges from and increasing their contributions to tax- critical to understanding the work dis- 0.09–0.12 overall, with just under half deferred accounts (TDA). Using data incentive effects of tax and welfare poli- of this effect on the intensive margin. from the Survey of Consumer Finances, cies, and the attendant design of optimal Overall, the analysis reveals that matching the authors show that at least 38 percent income tax and transfer schemes. Ziliak is a rather poor policy instrument with of U.S. households that are accelerating uses data from the Current Population which to raise retirement saving. their mortgage payments instead of sav- Survey and variation induced by funda- Looney examines the economic inci- ing in tax-deferred accounts are mak- mental reforms to the U.S. tax and wel- dence of non-refundable child-related tax ing the wrong choice. For these house- fare systems over the 1979–2001 period benefits in low-income single parent fam- holds, reallocating their savings can yield a to estimate the labor-supply response of ilies. Because non-refundable tax ben- mean benefit of 11 to 17 cents per dollar, single mothers to changes in their after- efits only reduce taxes due, many argue depending on the choice of investment tax and transfer wage rate and nonlabor that they cannot help low-income families assets in the TDA. In the aggregate, these income, conditional on whether or not without tax liability. Single parents may mis-allocated savings are costing U.S. they also participate in AFDC/TANF, be an exception because tax law permits households about 1.5 billion dollars per food stamps, or SSI. He finds that wage separated or divorced parents to exchange year. Finally, the authors show empirically changes have a large effect on the decision the tax benefits tied to their children. If that this inefficient behavior is unlikely to to work (the average elasticity of employ- parents exchange child-related tax ben-

32 NBER Reporter Winter 2006/7 efits and custodial parents are compen- opportunities and that custodial moth- They find that the match offer increases sated for the trade, then the incidence of ers receive increased child support as a both the revenue per solicitation and the such benefits is more progressive than a result. response rate. However, larger match naive estimate would suggest. Using data Karlan and List conducted a nat- ratios (that is, $3:$1 and $2:$1) relative from the Survey of Income and Program ural field experiment to further their to a smaller match ratio ($1:$1) had no Participation (SIPP) and exploiting varia- understanding of the economics of char- additional impact. The results provide tion in the value of child-related tax ben- ity. Using direct mail solicitations to over avenues for future empirical and theoret- efits due to differences in state tax systems 50,000 prior donors of a non-profit orga- ical work on charitable giving, cost-ben- and changes in federal tax law, Looney nization, they tested the effectiveness of efit analysis, and the private provision of finds that parents exploit these arbitrage a matching grant on charitable giving. public goods.

20th Annual TRIO Conference The Twentieth Annual TRIO Incentives and Ownership Rights” Industry from 1984 to 2002” Conference, so-named because it is jointly Discussants: Makoto Hanazono, Nagoya Discussants: George P. Baker, and sponsored by the NBER, the Centre for University, and Paul Oyer, Stanford Tatsuya Kikutani, Kyoto University Economic Policy Research (CEPR), and University and NBER the Tokyo Center for Economic Research Hirofumi Uchida, Wakayama (TCER), took place on December 15 Hideshi Itoh; Osamu Hayashida, University; Gregory F. Udell, Indiana and 16 in Tokyo. This year’s conference Osaka Keizai University; and University, and Wako Watanabe, focused on “Organizational Innovation Tatsuya Kikutani, Kyoto University, Tohoku University, “Bank Size and and Firm Performance.” It was organized “Complementarities among Authority, Lending Relationships in Japan” by George P. Baker, NBER and Harvard Responsibility, and Monitoring: Discussants: Takeo Hoshi, and Ayako University; Takeo Hoshi, NBER and Evidence from Japanese Business Yasuda, University of Pennsylvania University of California, San Diego; Groups” and Hideshi Itoh and Sadao Nagaoka, Discussants: Steven Tadelis, UC, Paul Oyer, “Ability and Employer Hitotsubashi University and TCER. The Berkeley, and Wako Watanabe, Tohoku Learning: Evidence from the Economist program was: University Labor Market” Discussants: Daiji Kawaguchi, George P. Baker; Robert Gibbons, Steven Tadelis, and Jonathan D. Hitotsubashi University and TCER, NBER and MIT; and Kevin J. Murphy, Levin, Stanford University, “A and Hideo Owan, Aoyama Gakuin University of Southern California, Costly Contracting Approach to the University “Strategic Alliances: Bridges Between Organization of Production” Islands of Conscious Power” Discussants: Ola Kvaloy and Sadao Arghya Ghosh and Hodaka Morita, Discussants: Hideshi Itoh, and Hodaka Nagaoka University of New South Wales, “An Morita, University of New South Wales Economic Analysis of Platform Sharing” Akira Takeishi and Sadao Nagaoka, Discussants: Reiko Aoki, Hitotsubashi Ola Kvaloy, University of Stavanger, Hitotsubashi University; and Yoshihisa University and TCER, and Hans and Trond E. Olsen, Norwegian Noro, Mitsubishi Research Institute, Gottinger, Hitotsubashi University School of Economics and Business “Determinants of Firm Boundaries: Administration, “Peer‑Dependent Empirical Analysis of the Japanese Auto

Strategic alliances range from short- structures. Their model focuses on two arrangements (coopetition), strategic dives- term cooperative projects, through long- issues emphasized by practitioners: spill- titures, total divestitures, licensing agree- term partnerships and joint ventures, to over effects (as opposed to hold-ups moti- ments, and royalty agreements. They show transactions that permanently restruc- vated by specific investments), and con- that any of these possible strategic alliances ture firm boundaries and asset ownership. tracting problems ex post (as opposed to could be optimal. Baker, Gibbons, and Murphy draw on only ex ante). They use this model to gen- In an earlier paper, Kvaloy and Olsen detailed discussions with practitioners to erate a large number of strategic alliance analyzed the conditions for implement- present a rich model of feasible governance possibilities, including simple cooperative ing peer-dependent incentive regimes when

NBER Reporter Winter 2006/7 33 agents have ownership rights. They showed action cost economics, agency theory and banks do not necessarily have a compara- that compensation tied to peer-perfor- property rights theory, Levin and Tadelis tive advantage in extending transactions- mance can induce employee-hold-up and develop a general procurement model that based lending. They also find, like BMPRS, obstruct the implementation of relational highlights the trade-off between produc- that smaller banks have strong relation- incentive contracts. In this paper, they pres- tive efficiency and the costs of adminis- ships with their borrowers. However, the ent some extensions: they argue that the trating performance contracts. They recast banking relationships in the United States costs of transferring ownership rights to the question of firm boundaries as one of and Japan are strong in somewhat different agents may depend on whether there exist contracting over inputs or outputs, result- dimensions. This paper clarifies these and conditions that call for peer-dependent ing in empirically testable predictions that other interesting similarities and differences incentives. In particular, they show that if are consistent with several previous studies. between the United States and Japan. there exists common noise that makes rel- Their results demonstrate why control and Oyer studies the human capital devel- ative performance evaluation optimal, or cost incentives will shift in complementary opment and firm-worker matching pro- peer pressure that makes joint performance ways, laying some foundations to the defi- cesses for Ph.D. economists. This group is evaluation optimal, then the firm will be nition of hierarchy and market transactions useful for this purpose because the types more reluctant to give up ownership rights. within transaction cost economics. of jobs they hold can be easily categorized Itoh, Hayashida, and Kikutani offer Since Coase’s (1937) seminal work, the and they have an observable productivity an empirical test of complementarities boundaries of the firm have long been one measure (that is, publications.) He derives among delegated authority, responsibil- of the most important issues for researchers, a two-period model to motivate an empir- ity, and monitoring, using unique survey and the auto industry has been one of the ical analysis of economist job matching data collected from group-affiliated compa- most investigated industries. One example upon graduation, matching ten years later, nies in Japan. The survey provides informa- is Monteberde and Teece (1982), which and productivity in the first ten years. He tion about how various decisions are made demonstrated that transaction cost — mea- shows that matching to a higher ranked within business groups, each of which con- sured by engineering efforts and firm-speci- institution affects productivity. He pres- sists of a large core parent firm and its net- ficity to design component — indeed did ents evidence that employers improve their work of affiliated firms, such as subsidiaries matter for the vertical integration deci- estimates of economists’ ability early in and related companies. The authors find sion by OEMs (GM and Ford). Takeishi, their career in a way that determines lon- some evidence that delegated authority and Nagaoka, and Noro extend their analysis ger-term job placement. He also finds that responsibility are complementary, imply- in three directions. First, for the dependent the initial placement of economists to insti- ing that increasing assigned responsibility variable, in addition to the two choices tutions does not show much evidence of raises the marginal return from increas- (make internally or buy from the mar- systematic misallocation along observable ing delegated authority. They also obtain ket), they put forward the third choice, characteristics. a stronger result, that performance is likely “buy from affiliated (‘keiretsu’) suppliers.” Ghosh and Morita explore the man- to be higher under the combination of low Second, for independent variables, they agerial implications and economic conse- authority and low responsibility, or that of examine a set of new variables to mea- quences of platform sharing under models high authority and high responsibility, than sure multiple dimensions of contractibil- of horizontal and vertical product differen- under the “mix and match” combinations ity. Third, they use a set of panel data of tiation. By using a common platform across where one of them is low and the other the Japanese auto industry, which they have different products, firms can save on fixed high. They then study the effects of moni- built up to cover the make-or-buy decisions costs for platform development. At the toring intensity on the authority-responsi- of 7 OEMs on 54 types of components for same time, platform sharing imposes restric- bility pair and find that performance of the almost two decades from 1984 to 2002. tions on firms’ ability to differentiate their firm with the combination of high author- Current theoretical and empirical products, and this reduces their profitabil- ity and high responsibility is increasing in research suggests that small banks have a ity. It might appear that platform sharing monitoring intensity, while the combina- comparative advantage in processing soft across firms makes consumers worse off tion of low authority and low responsibil- information and delivering relationship because firms cooperate in their product ity is not. The result is consistent with the lending. The most comprehensive anal- development processes to maximize their theoretical hypothesis that increasing mon- ysis of this view found using U.S. data joint profit. The authors find, however, itoring intensity raises the marginal return is by Berger, Miller, Petersen, Rajan, and that platform sharing benefits consumers in from increasing delegated authority and Stein, 2005 (BMPRS). Uchida, Udell, and their framework because it intensifies com- responsibility. Watanabe use essentially the same meth- petition in their horizontal differentiation What determines the boundaries of odology as BMPRS on a unique Japanese model, and because it increases the qual- organizations and the cost-based incen- dataset. Like BMPRS, they find that larger ity of the lower-end product in our vertical tives within and between contracting enti- firms tend to borrow from larger banks. differentiation model. They also show new ties? These questions have received much However, unlike BMPRS, they do not find channels through which a merger makes attention, and many approaches have been that this is because larger firms are more consumers worse off in the presence of plat- developed. Building on ideas from trans- transparent. Their results imply that large form sharing.

34 NBER Reporter Winter 2006/7 NBER News

2006 Awards and Honors A number of NBER researchers of the Canadian Institute for Advanced Erik Brynjolfsson won the Best received honors, prizes, awards, and pro- Research. Paper Award, International Conference fessional kudos during 2006. In alphabet- Jonathan Berk shared the TIAA- on Information Systems, for “Information ical order, those so honored (excluding CREF Paul Samuelson Award that was Worker Productivity: Task-level Evidence”, honors by the individual’s own univer- presented in January 2006. He was also co-authored with Sinan Aral and Marshall sity) are: cited for publishing one of the two best van Alstyne. He also won the award Daron Acemoglu received the 2005 papers ever published in The Review of for Best Paper in Valuing IT Track at John Bates Clark Medal in economics Financial Studies, 2006: “A Critique of Size the 2006 International Conference on and the Turkish Academy of Sciences Related Anomalies.” Finally, he received Information Systems for “Which Came Distinguished Science Award; he also was the Bernstein-Fabozzi/Jacobs Levy Award First, IT or Productivity? The Virtuous named a Fellow of the Society of Labor (Outstanding Article) from the Journal of Cycle of Investment and Use in Enterprise Economics and a Fellow of the American Portfolio Management, 2006. Systems” with Sinan Aral and D.J. Wu. Academy of Arts and Sciences. Marianne Bertrand won the 2006 Laurent Calvet received the “Best Alberto Alesina, winner of the 2006 John T. Dunlop Award of the Labor finance researcher under the age of 40” Distinguished CES (Center for Economic and Employment Relations Association prize for 2006 from Le Monde (news- Studies) Fellow Prize in Economics, for contributions to research that address paper) and the Europlace Institute of granted at the University of Munich. He industrial relations/employment prob- Finance (EIF). was also made a Fellow of the American lems of national significance. Jeff Campbell received the Journal Academy of Arts & Sciences. Rebecca Blank became a Fellow of of Industrial Economics 2nd Annual “Best Heitor Almeida and Daniel the Society of Labor Economists and a Article of the Year’’ Prize for his paper Wolfenzon won the (second Vice President of the American Economic with Hugo Hopenhayn, “Market Size prize) for the “Best Paper Published in Association. Matters.” 2005 in the Journal of Francine D. Blau is President, Society David Card and Alan Krueger won in the Areas of Corporate Finance and of Labor Economists. the IZA Prize in Labor Economics. Organizations” for their paper, “The Effect Alan Blinder delivered the Presi­ Alessandra Casella was awarded a of External Finance on the Equilibrium dential Address to the Eastern Economic Guggenheim Fellowship for 2006-7 for Allocation of Capital.” Association in Philadelphia in February. her work on Storable Votes. Lee J. Alston was elected President Richard Blundell was elected a Raj Chetty was recipient of a Smith of the International Society for New Fellow of the Society of Labor Economists. Richardson Public Policy Fellowship, one Institutional Economics. Jacob Boudoukh, Matthew of three given annually to young econo- Joseph Altonji was elected a Fellow Richardson, and Robert Whitelaw won mists working in public economics and of the Society of Labor Economists. “The Goldman Sachs Asset Management public policy in the United States. He Torben G. Andersen was awarded Award for the Best Paper in Investments” also received an NSF CAREER grant, the a Certificate of Appreciation from the at the 2006 WFA Meetings in Colorado. NSF’s major award in support of early- American Statistical Association for editor- John F P Bridges was awarded the career scientists. ship of the Journal of Business & Economic International Society of Pharmaco­ was elected a Statistics 2004–6. economics and Outcomes Research Fellow of the American Academy of Arts Joshua Angrist was elected a Fellow (ISPOR) Bernie O’Brien New Investi­ and Sciences. of the Society of Labor Economists and a gator Award for 2006. John H. Cochrane won the Fama/ Fellow of the American Academy of Arts Charles Brown was named a Fellow DFA second prize from the Journal of & Sciences. of the Society of Labor Economists. Financial Economics for “The Risk and Abhijit Banerjee received the first Jeffrey R. Brown was appointed to Return of Venture Capital.” Michael Wallerstein Award from the the Social Security Advisory Board. Janet Currie was elected a Fellow American Political Science Association in Markus K. Brunnermeier received of the Society of Labor Economists in September 2006. the BGI-Michael Brennan best paper 2006. Lucian Bebchuk is President-Elect award (runner up) for his paper David M. Cutler and Jonathan of the American Association of Law and “Information Leakage and Market Gruber were joint recipients of the first Economics. Efficiency” published in the Review of ASHE medal (American Society of Health Roland Benabou became a Fellow Financial Studies. Economists), given to the outstanding

NBER Reporter Winter 2006/7 35 health economist aged 40 and younger. the Vaccine Industry.” of Political and Social Science in 2006. Peter M. DeMarzo won Barclays Roderick Floud was awarded the Finally, she delivered the Ely Lecture to Global Investors/Michael Brennan Prize honorary degree of Doctor of Letters the American Economic Association in for the Best Paper published in Volume by the University of Westminster in January 2006. 18 of The Review of Financial Studies: November 2006. Austan Goolsbee was named a “The Pooling and Tranching of Securities: Robert W. Fogel was named Fulbright Scholar for 2007 at the London A Model of Informed Intermediation.” Indispensable Person of the Year for 2006 School of Economics and the Institute of Barry Eichengreen received a doc- by the Alliance for Aging Research for Fiscal Studies for his research on taxes and tor honoris causa from the American contributions to the study of health and the Internet in the European Union and University in Paris. aging. the United States. Stanley L. Engerman (and Kenneth Richard Freeman and Edward Gary Gorton and Frank A. Schmid L. Sokoloff) won the Arthur H. Cole Lazear shared the 2006 Society of Labor received the 2006 Hicks-Tinbergen Prize for the outstanding article published Economists’ prize honor- Medal for their paper “Capital, Labor in the Journal of Economic History from ing lifetime achievements in the field of and the Firm: A Study of German September 2005-6 for “The Evolution labor. Codetermination” at the 21st Congress in of Suffrage Institutions in the New Barbara Fraumeni received a gold Vienna on August 27, 2006. The medal is World.” Engerman was also named a medal from the Secretary of Commerce awarded every two years for the best paper Distinguished Fellow of the American for her work on R and D while Chief in the Journal of the European Economic Economic Association. Economist at the Bureau of Economic Association. Isaac Ehrlich was named Founding Analysis of the U.S. Department of John R. Graham and Campbell Editor-in-Chief of the Journal of Human Commerce. She also received the 2006 Harvey won a number of best paper Capital. He was also named NYSTAR Carolyn Shaw Bell Award from the awards: for their paper with Shiva Rajgopal, Distinguished Professor of 2005 by Committee on the Status of Women in “The Economic Implications of Corporate the New York state office of Science, the Economics Profession of the American Financial Reporting,” the 2006 Notable Technology, and Academic Research Economics Association, given annually to Contribution to Accounting Literature (NYSTAR) Faculty Development the individual who has furthered the sta- Award and the 2006 FARS (Financial Program for his work on human capital tus of women in the economics profession and Reporting Section) best paper and economic growth and development. through example, achievements, increas- award from the American Accounting Robert Feenstra received the ing our understanding of how women can Association. Also, for “Payout Policy in Bernhard Harms Prize for 2006, from the advance in the profession, and the men- the 21st Century” with Alon Brav and Kiel Institute for World Economics. toring of others. Roni Michaely, the 2006 Jensen Prize for Martin Feldstein was appointed to Victor Fuchs was elected an honor- best corporate finance paper published in the President’s Foreign Intelligence Advis­ ary member of Alpha Omega Alpha (the the Journal of Financial Economics. ory Board. Phi Beta Kappa of Medical Schools) in Robert E. Hall became a Fellow of Raquel Fernandez presented the fall 2006. the Society of Labor Economists in 2006. Marshall Lecture at the European Economic Xavier Gabaix received the Lars Peter Hansen won the Erwin Association meetings in Vienna in August “Young Scientist Award for Socio- and Plein Nemmers Prize in Economics, 2006. Econophysics” from the German Physical awarded biennially by Northwestern Jesus Fernandez-Villaverde and Juan Society for his work on the origins of scal- University and designed to recognize Rubio-Ramirez won the Richard Stone ing laws in economics (QJE 1999, “Zipf ’s “work of lasting significance” in the dis- prize in applied . law for cities”) and finance (QJE 2006, cipline. He was also voted a Fellow of the Amy Finkelstein received CES-Ifo’s “Institutional Investors and Stock Market American Finance Association. Distinguished Research Affiliate Award, Volatility” and Nature 2003, “Theory of Eric Hanushek became a member for “best paper in Public Economics by Power Law Distributions in Financial of the National Academy of Education a young scholar” for “The Interaction of Market Fluctuations”). and a Fellow of the Society of Labor Public and Private Insurance: Medicaid Sherry Glied was elected to the Economists. and the Long-Term Care Insurance Institute of Medicine of the National Oliver Hart is President of the Market”(co-authored by Jeffrey R. Academy of Sciences. American Law and Economics Association Brown). She also was given Research received the 2005 and Vice President of the American America’s Eugene Garfield Economic Carolyn Shaw Bell Award from the Economic Association. Impact of Medical and Health Research Committee on the Status of Women in Geoffrey Heal was named a Fellow Award for “outstanding research on the Economics Profession of the American of the Association of Environmental and how medical or health research impacts Economics Association. She also became Resource Economists in recognition of the economy” for “Static and Dynamic a Fellow of the National Academy of a lifetime of significant contributions to Effects of Health Policy: Evidence from Sciences and of the American Academy the area of environmental and resource

36 NBER Reporter Winter 2006/7 economics. All Perils Insurance” was selected the Enrico Moretti received the 2006 won the Ulysses 2005 article of the year by the Journal of IZA Young Labor Economist Award. Medal from University College Dublin Insurance Regulation. Tobias Moskowitz won (with co- for his contribution to research in eco- Edward Lazear and Richard authors Annette Vissing-Jorgensen and nomics and human behavior. Freeman shared the 2006 Society of Marianne Bitler) the Journal of Finance Vernon Henderson was elected a Labor Economists’ Jacob Mincer prize Distinguished Paper for Fellow of the Regional Science Association honoring lifetime achievements in the 2005. He also won the AFA Fisher Black International. field of labor. Prize for 2007, which honors the top Takeo Hoshi will be an inaugural David K. Levine is incoming finance scholar under age 40. recipient of the Enjoji Jiro Memorial President of the Society for Economic Joseph P. Newhouse was presi- Prize. Dynamics. dent of the American Society of Health Erik Hurst and Mark Aguiar won Gary Libecap was President of Economists in 2006 and became a member the TIAA-CREF Paul A. Samuelson the Western Economics Association of the Committee on National Statistics Award for their paper “Consumption vs. and President of the Economic History of the National Research Council. Expenditure” (published in the October Association in 2006. Muriel Niederle received a Sloan 2005 JPE). The Award is designed to rec- Jens Ludwig was the winner of the Fellowship beginning in summer 2007. ognize the best scholarly writing on issues 2006 David N. Kershaw Prize, presented Lubos Pastor and Pietro Veronesi related to lifelong financial security. by the Association for Public Policy won the 2006 Barclays Global Investors Takatoshi Ito was appointed a Analysis and Management, established to Prize for the best paper presented at the member of the Council of Economic honor persons who, under the age of 40, European Finance Association Meetings and Fiscal Policy (Japanese equivalent have made a distinguished contribution in Zurich: “Technological Revolutions of the President’s Council of Economic to the field of public policy analysis and and Stock Prices.” Advisers). management. Mark Pauly’s paper “Terrorism Michael C. Jensen was named Lisa M. Lynch is the next Chairman Losses and All Perils Insurance” (co- Honoris Causa Professor by HEC Business of the Board of the Federal Reserve Bank authored with Howard Kunreuther) was School, Paris in November 2006. of Boston. selected the 2005 article of the year by the In March 2006 Jensen received the Thomas MaCurdy became a Fellow Journal of Insurance Regulation. Dean’s Leadership Award in Corporate of the Society of Labor Economists. Tomas Philipson was awarded the Governance, Drexel University, LeBow Ulrike Malmendier won the Kenneth Arrow prize of the International College of Business, and in January 2006 Citation of Excellence Award given annu- Health Economics Association for best he was awarded the LECG prize for his ally by Emerald Management Reviews paper in a given year. lifetime contributions to finance. for top articles among the 15,000 peer- received the Elaine Dirk Jenter won “The CRA reviewed articles in the Reviews’ database Bennett Award from the American International Award for the Best Paper for her article “CEO Overconfidence and Economics Association’s Committee on in Corporate Finance” at the 2006 WFA Corporate Investment”. the Status of Women in the Economics Meetings in Colorado. Robert A. Margo was the 2006 Profession. She also won the Bernacier Zorina Khan received the Economic recipient of the Clio Award from the Prize as best European economist under 40 History Association’s Alice Hanson Jones Conference, awarded annu- working on finance or macroeconomics. Biennial Prize for an outstanding book ally for “exceptional service” to the field James M. Poterba is 2nd Vice on North American economic history for of cliometrics. President, National Tax Association. The Democratization of Invention: Patents Grant Miller received the inaugu- Vincenzo Quadrini and José-Víctor and Copyrights in American Economic ral best student paper award from the Ríos-Rull shared the 2005 Arrow Prize Development, 1790–1920. American Society of Health Economists for Senior Economists with Paul Klein for Leonid Kogan won the 2006 Smith and the Biennial Award for Distinguished their paper, “Optimal Time-Consistent Breeden Prize (first prize) for his Journal Contribution to Scholarship in Population Taxation with International Mobility of of Finance paper, “The Price Impact and from the American Sociological Capital”. Survival of International Traders”, co- Association’s Section on Population (joint Joshua Rauh received the Brattle authored by Stephen A. Ross, Jiang Wang, with David Cutler.) Prize, First Prize Paper, for the best cor- and Mark M. Westerfield. Olivia S. Mitchell received an hon- porate finance paper published in the Alan Krueger, in addition to the orary Doctorate in Economics from Journal of Finance, 2006. IZA Prize in Labor Economics shared the University of St. Gallen. She also Ricardo A. Reis was the W. Glenn with David Card, became a Fellow of the was appointed to Singapore’s Central Campbell Rita Ricardo-Campbell Society of Labor Economists. Provident Fund Advisory Board. National Fellow and the Arch W. Shaw Howard Kunreuther and Mark Robert Moffitt was named a Fellow Fellow at the Hoover Institution in 2006. Pauly’s paper, “Terrorism Losses and of the Society of Labor Economists. Matthew Richardson, Jacob

NBER Reporter Winter 2006/7 37 Boudoukh, and Robert Whitelaw won Fellow of the American Finance Association prize) for an NYU Stern best paper of the “The Goldman Sachs Asset Management in 2006. He is also 2006–7 President of year: “Information Immobility and the Award for the Best Paper in Investments” the Eastern Economic Association. His Home Bias Puzzle”. at the 2006 WFA Meetings in Colorado. term as vice president of the American W. Kip Viscusi received the 2006 Richard Rogerson was elected a Economic Association expired in January Distinguished Economist award from Fellow of the Econometric Society. 2006. the Kentucky Economics Association. Kenneth Rogoff was chosen a Vice Jody L. Sindelar is president of the Also, he and Richard J. Zeckhauser won President of the American Economic recently formed American Society of the 2006 Ronald H. Coase Prize for the Association (for 2007). Health Economics (ASHE). best paper published by a University of Christina Romer served as Vice Kenneth L. Sokoloff (and Stanley Chicago Law School Journal. President of the American Economic L. Engerman) won the Arthur H. Cole Annette Vissing-Jorgensen won Association. Prize for the outstanding article published the Journal of Finance Brattle Prize David Romer was named a Fellow in the Journal of Economic History from Distinguished Paper for 2005 (with co- of the American Academy of Arts & September 2005–6 for “The Evolution of authors Tobias Moskowitz and Marianne Sciences. Suffrage Institutions in the New World.” Bitler.) Stephen A. Ross won the 2006 Smith Pablo T. Spiller is 2nd Vice President Jiang Wang won the 2006 (first prize) for his Journal of the International Society of New Breeden Prize (first prize) for his Journal of Finance paper, “The Price Impact and Institutional Economics. of Finance paper, “The Price Impact and Survival of International Traders”, co- James H. Stock became a Fellow Survival of International Traders”, co- authored by Leonid Kogan, Jiang Wang, of the American Academy of Arts & authored by Leonid Kogan, Stephen A. and Mark M. Westerfield. Sciences. Ross, and Mark M. Westerfield. Michael Rothschild was named a Richard Sutch was elected a Fellow Robert Whitelaw, Matthew Distinguished Fellow of the American of the American Association for the Richardson, and Jacob Boudoukh won Economic Association in 2005. The award Advancement of Science, inducted in “The Goldman Sachs Asset Management was presented in 2006. February 2006, and elected Honorary Award for the Best Paper in Investments” Emmanuel Saez won the Doug President of the International Economic at the 2006 WFA Meetings. Purvis Memorial Prize for an article co- History Association in August 2006. Barbara L. Wolfe was appointed to the authored with Michael Veall of McMaster Alan M. Taylor was on leave at Advisory Committee to the Director of the University, “The Evolution of High London Business School for part of 2006 National Institute of Health. Incomes in North America: Lessons from as the recipient of a fellowship from the Daniel Wolfenzon won the Jensen Canadian Evidence.” The Purvis Prize is John Simon Guggenheim Memorial Prize (second place) for “Best Paper awarded annually to the authors of a Foundation. Published in 2005 in the JFE.” highly significant written contribution to Werner Troesken received the Lu Zhang won the Smith Breeden Canadian economic policy. Economic History Association’s Alice Award (First Prize) for 2005 from Thomas Sargent was elected Hanson Jones Biennial Prize for an out- American Finance Association and President of the American Economic standing book on North American eco- Journal of Finance. Association. nomic history: Water, Race, and Disease Richard J. Zeckhauser and W. Kip Douglas A. Shackelford received (MIT Press, 2004). Viscusi won the 2006 Ronald H. Coase the Ray M. Sommerfeld Outstanding Aleh Tsyvinski received the NSF Prize for the best paper published by Tax Educator Award from the American Career Award for 2007-12 in 2006. a University of Chicago Law School Taxation Association. Stijn Van Nieuwerburgh won the Journal. Robert J. Shiller was inducted as Glucksman Institute Research Prize (first

38 NBER Reporter Winter 2006/7 Program and Working Group Meetings

China Working Group Meeting

The NBER’s Working Group on Oberholzer‑Gee, Harvard University, North Carolina, “Financial Openness China met in Cambridge on October “The Political Economy of Firm and the Chinese Growth Experience” 13. NBER Research Associate Shang‑Jin Size Distributions: Evidence from Discussant: Zhiwu Chen, Yale Wei of the IMF organized the meeting. Post‑Reform China” University The following papers were discussed: Discussant: Bruce Reynolds, University of Virginia Chang‑Tai Hsieh, University of Douglas Almond, Columbia University California, Berkeley, and Peter Klenow, and NBER; Lena Edlund, Columbia David Dollar, World Bank, and Stanford University and NBER, University; and Hongbin Li and Junsen Shang‑jin Wei, “Das (Wasted) Kapital: “Misallocation and Manufacturing TFP Zhang, Chinese University of Hong Firm Ownership and Investment in China and ” Kong, “Long‑Term Effects of China’s Efficiency in China” Discussant: Lee Branstetter, Carnegie Great Famine in Hong Kong and Discussant: Galina Hale, Federal Mellon University and NBER Mainland China” Reserve Bank of San Francisco Discussant: Xiaobo Zhang, Yi Qian, Northwestern University, International Food Policy Research Geert Bekaert, Columbia University “Pricing and Marketing Impacts of Entry Institute and NBER; Campbell R. Harvey, by Counterfeiters and Imitators” Duke University and NBER; and Discussant: Nancy Qian, Brown Tarun Khanna and Felix Christian Lundblad, University of University

Almond and his co-authors eval- political economy. They exploit a new will) and that directed at potential pro- uate whether the Great China Famine database of up to five million obser- vincial outsiders (potential entrants, if had negative effects on its survivors. vations in two years in China, 1999 you will) and show that the effects on According to the fetal-origins hypothesis, and 2003. They are particularly inter- the size distribution are opposite. The cohorts in utero during the famine should ested in studying the effects of China’s results are consistent with local govern- have suffered the greatest long-term dam- uneven march to the market on firms ments — in an attempt to protect the age. Consistent with this hypothesis, the of different ownership, namely state- autonomy granted them by the center authors finds a broad spectrum of - com owned enterprises, collectively owned during the reform process — “hitting promised outcomes for cohorts born in enterprises, foreign invested enterprises, back” at central government efforts 1960 who appear in the 2000 Chinese and private firms. Their results show to contain them, perhaps in order to Census. These effects are greatest for those that massive liberalization in China encourage their own local (provincial) in rural areas, but extend to those who has encouraged the growth of foreign- firms. The authors conclude that the were born in urban areas. The authors also invested enterprises and, to a lesser simplest measures appear more impor- find that Hong Kong residents who were extent, collective enterprises (includ- tant than conventional measures of born in China exhibit inferior health out- ing Township and Village Enterprises), financial constraints in determining comes, including reduced birth weight of but they have never encouraged genu- firm size distributions, even after con- children born to parents who themselves inely private firms. The best thing that trolling for industry (technological) were in utero during the famine. Health can be said for private enterprise in effects. effects exist among emigrants from main- China is that foreign direct investment Based on a survey that Dollar and land China despite the selective effects appears to spur the entry of small firms. Wei designed, which covers a strati- of emigration, which are generally posi- Surprisingly, price flexibility, an impor- fied random sample of 12,400 firms tive. Moreover, no corresponding dam- tant form of liberalization, does not in 120 cities in China with firm-level age among cohorts born in Hong Kong, help private firms, although it does help accounting information for 2002–4, and thereby shielded from the famine, is large foreign firms and large collectives they examine the presence of system- observed. to become even larger. The research- atic distortions in capital allocation Khanna and Oberholzer‑Gee ers are also able to distinguish between that result in uneven marginal returns study the relationship between firm government interference directed at to capital across firm ownership, size distributions and some aspects of provincial insiders (incumbents, if you regions, and sectors. The survey pro-

NBER Reporter Winter 2006/7 39 vides a systematic comparison of invest- tutions. Standard growth regressions Counterfeit and imitative prod- ment efficiency among wholly and par- cannot explain China’s extraordinary ucts appear similar to authentic prod- tially state-owned, wholly and partially growth experience, and the authors fail ucts but usually have lower quality. foreign-owned, and domestic privately to find an important role for foreign However, unlike imitation, counterfeit- owned firms, conditioning on their trade and foreign direct investment. In ing infringes upon intellectual property sector, location, and size. The research- contrast, the sheer volume of invest- rights by claiming a brand name that it ers find that even with a quarter-cen- ment has played a significant role in does not own. Qian models the pric- tury of reforms, state-owned firms still China’s growth. As China’s per cap- ing, quality, and marketing strategies of have significantly lower returns to capi- ita GDP continues to grow, it must producers of authentic and counterfeit tal, on average, than domestic private find sustainable sources of growth. The goods in a setting of oligopolistic com- or foreign-owned firms. Similarly, cer- authors identify a more efficient finan- petition under both complete and asym- tain regions and sectors have consis- cial sector, less state ownership, higher metric information. His model explains tently lower returns to capital than quality of government institutions, and the effects of both counterfeit and imi- other regions and sectors. A back-of- full financial openness as important tative entry with different parameter the-envelope calculation suggests that factors. Interaction analysis suggests specifications. He collects data from if China could allocate capital more that the beneficial effects of financial Chinese shoe companies from 1993– efficiently, it would reduce its invest- openness first require further financial 2004 to test the theoretical predictions. ment intensity from the current 40 and institutional development. China is Exploiting the discontinuity of govern- percent of GDP to 35 percent without less of an outlier in its growth variabil- ment enforcement efforts for the foot- sacrificing its economic growth (and ity experience but has achieved high wear sector in 1995, and the differences hence deliver a greater improvement to growth with surprisingly low growth in authentic companies’ relationships its citizens’ living standard). volatility. with their local governments, Qian uses Bekaert and his co-authors reflect Resource misallocation can lower two different techniques to measure the on China’s economic performance from aggregate total factor productivity effects of counterfeit entry on authen- the perspective of the experiences of a (TFP). Hsieh and Klenow use micro tic manufacturers’ prices, qualities, and broad panel of countries. The authors data on manufacturing establishments profits. The empirical results are con- formulate an econometric framework, to quantify the extent of this misalloca- sistent with the theoretical predictions. building on standard growth regres- tion in China and India in recent years. First, low-quality counterfeit entrants sions that allows them to measure the For each country, they measure sizable induce authentic producers to both impact of various factors on economic gaps in the marginal products of labor produce higher quality products and growth and growth variability. Because and capital across plants within nar- raise prices. Second, there is empirical China has become more and more inte- rowly-defined industries. When capital evidence for the presence of asymmet- grated into the world’s economic and and labor are hypothetically reallocated ric information, under which authen- financial landscape, the authors devote to equalize the marginal products, they tic prices rise further to signal quality special attention to measures of (de calculate manufacturing TFP gains on (or authenticity). However, this price- jure) financial openness. They also doc- the order of a factor of 2. Output gains signaling effect diminishes over time. ument how the real effects of openness are nearly a factor of 4 if physical capi- Third, other costly non-price devices are affected by financial development, tal accumulates to restore the original are used for signaling and reducing political risk, and the quality of insti- average marginal product of capital. counterfeit sales.

40 NBER Reporter Winter 2006/7 Health Care Program Meeting The NBER’s Program on Health Dhaval Dave, Bentley College and David Meltzer, and Domenico Care met in Cambridge on October 20. NBER, and Robert Kaestner, University Salvatore, Universita Bocconi, “Sex and David Meltzer, NBER and University of of Illinois at Chicago and NBER, Physician Practice Variation” Chicago, organized the program. These “Medicare and Health Behaviors” papers were discussed: David M. Cutler and Robert S. Joshua Graff Zivin, Columbia University Huckman, Harvard University and Amy Finkelstein and Daron and NBER; Harsha Thirumurthy, Yale NBER, and Jonathan T. Kolstad, Acemoglu, MIT and NBER, “Input University; and Markus Goldstein, Harvard University, “Is Entry Efficient and Technology Choices in Regulated The World Bank, “AIDS Treatment and When Inputs are Constrained? Lessons Industries: Evidence from the Health Intrahousehold Resource Allocations: from Cardiac Surgery” Care Sector”(NBER Working Paper No. Children’s Nutrition and Schooling in 12254) Kenya” Joshua Lerner, Harvard University and NBER, and Ulrike Malmendier, Darius Lakdawala and Neeraj Sood, Sean Nicholson, Cornell University University of California, Berkeley and The RAND Corporation and NBER, and NBER; Michael Waldman, Cornell NBER, “Contractibility and the Design “Health Insurance as a Two‑Part Pricing University; and Nodir Adilov, Purdue of Research Agreements” Contract”(NBER Working Paper No. University, “Does Television Cause 12681) Autism?”

Finkelstein and Acemoglu ask Monopolies appear throughout ciency in the innovation market. how regulatory change might affect medical care markets, as a result of pat- Basic economic theory suggests the input mix and technology choices ents, limits to the extent of the mar- that health insurance coverage may of regulated industries. They present ket, or the presence of unique inputs cause a reduction in prevention activi- a simple neoclassical framework that and skills. Economists typically think ties, but empirical studies have yet to emphasizes the change in relative fac- of such monopolies as necessary evils provide evidence to support this pre- tor prices associated with a shift from or even pure inefficiencies. However, diction. Dave and Kaestner extend the full-cost to partial-cost reimbursement, in the health care industry, the dead- analysis of the effect of health insur- and investigate how this affects firms’ weight costs of monopoly may be much ance on health behaviors by allowing technology choices through substitu- smaller or even absent. Health insur- for the possibility that health insur- tion of (capital embodied) technol- ance, frequently implemented as an ex ance has a direct (ex ante moral hazard) ogies for tasks previously performed ante premium coupled with an ex post and indirect effect on health behav- by labor. Empirically, they study the co-payment per unit consumed, oper- iors. The indirect effect works through change from full-cost to partial-cost ates as a two-part pricing contract. This changes in health promotion infor- reimbursement under the Medicare allows monopolists to extract consumer mation and the probability of illness Prospective Payment System (PPS) surplus without inefficiently constrain- that may be a byproduct of insurance- reform, which increased the relative ing quantity. Lakdawala and Sood induced greater contact with medi- price of labor faced by U.S. hospitals. note that this view of health insurance cal professionals. The authors iden- Using the interaction of hospitals’ pre- contracts has several novel implica- tify these two effects and in doing so PPS Medicare share of patient days tions: 1) medical care monopolies may identify the pure ex ante moral hazard with the introduction of these regula- have smaller or no deadweight costs effect. They find limited evidence that tory changes, they document a substan- in the goods market, because insured obtaining health insurance reduces pre- tial increase in capital-labor ratios and consumers face low co-payments; 2) vention and increases unhealthy behav- a large decline in labor inputs associ- since monopolists have incentives to iors among elderly persons. ated with PPS. Most interestingly, they seek low co-payments, price regu- The provision of life-saving antiret- find that the PPS reform seems to have lation of health care monopolies is roviral (ARV) treatment has emerged encouraged the adoption of a range of inferior to laissez-faire or simple tax- as a key component of the global new medical technologies. They also and-transfer schemes that redistribute response to HI V/AIDS, but very lit- show that the reform was associated monopoly profits; and 3) competitive tle is known about the impact of this with an increase in the skill composi- health insurance markets or optimally intervention on the welfare of children tion of these hospitals, which is con- designed public health insurance can in the households of treated persons. sistent with technology-skill or capital- eliminate static losses in the goods mar- Zivin, Thirumurthy, and Goldstein skill complementarities. ket while still improving dynamic effi- estimate the impact of ARV treatment

NBER Reporter Winter 2006/7 41 on children’s schooling and nutrition Washington — characterized by high test their theoretical predictions con- outcomes using longitudinal household precipitation variability. Employing a cerning the welfare effects of free entry survey data collected in collaboration variety of tests, they show that in each in the presence of scarce inputs and het- with a treatment program in western of the three states (and across all three erogeneous quality by considering how Kenya. They find that children’s weekly states when pooled) there is substan- the 1996 repeal of certificate-of-need hours of school attendance increase tial evidence that county autism rates (CON) legislation in Pennsylvania by over 20 percent within about six are indeed positively related to county- affected the market for coronary artery months after treatment is initiated for wide levels of precipitation. In the final bypass graft (CABG) surgery in that the adult household member. For boys set of tests, they use California and state from 1993 to 2003. Variation in treatment households, these increases Pennsylvania data on children born in entry across markets following this are closely related to decreases in their between 1972 and 1989 to show, again exogenous regulatory change allows market labor supply. Similarly, young consistent with the television-as-trigger them to estimate the effect of entry on children’s short-term nutritional sta- hypothesis, that county autism rates are market quantity, output quality, and tus — as measured by their weight-for- also positively related to the percentage welfare. The 1996 repeal of CON led height Z-score — also improves dramat- of households that subscribe to cable to a 56 percent increase in the number ically. The researchers argue that these television. The precipitation tests indi- of hospitals offering CABG by 2003. treatment effects will be considerably cate that just under 40 percent of autism This dramatic entry was not associ- larger when compared to the counter- diagnoses in the three states studied are ated with an increase in the number factual scenario of no ARV treatment. the result of television watching because of surgeries performed in the state. Their results show how intrahouse- of precipitation, while the cable tests Rather, entry led to a redistribution of hold resource allocation is altered in indicate that approximately 17 percent surgeries from lower- to higher-qual- response to significant health improve- of the growth in autism in California ity surgeons. Further, the researchers ments. Because the improvements in and Pennsylvania during the 1970s and argue, the inelastic supply of high-qual- children’s schooling and nutrition will 1980s is attributable to the growth ity cardiac surgeons limited the degree affect their socioeconomic outcomes in of cable television. These findings are of excess entry following the repeal adulthood, the provision of ARV treat- consistent with early childhood televi- of CON. While they cannot observe ment is likely to generate significant sion viewing being an important trigger cost, the fact that quality improved long-run macroeconomic benefits. for autism. and that excessive entry was limited Autism is currently estimated to In 2003, 281 attending physicians by the availability of surgeons together affect approximately one in every 166 on the general medicine services of suggest that free entry increased wel- children, yet the cause or causes of six academic hospitals were surveyed fare in Pennsylvnia’s market for cardiac the condition are not well understood. concerning demographic and profes- surgery. One of the current theories concern- sional attributes. They also were asked Lerner and Malmendier analyze ing the condition is that among a set to name three other general medicine how variations in contractibility affect of children vulnerable to developing attendings to whom they were most the design of contracts in the context the condition because of their under- likely to turn for advice during a typi- of biotechnology research agreements. lying genetics, the condition manifests cal month on the inpatient general A major concern of firms financing itself when such a child is exposed medical services. Based on the sur- biotechnology research is that the R to a (currently unknown) environ- vey results, Meltzer and Salvatore find and D firms might use the funding mental trigger. Nicholson, Waldman, that both professional and personal fac- to subsidize other projects or to sub- and Adilov empirically investigate the tors affect who such physicians ask for stitute one project for another. The hypothesis that early childhood tele- advice. Attendings are more likely to researchers develop a model based on vision viewing serves as such a trigger. name colleagues who are more experi- the property-rights theory of the firm Using the Bureau of Labor Statistics’ enced (older), graduated from a higher- that allows for researchers in the R and American Time Use Survey, they first ranked medical school, and read more D firms to pursue multiple projects. establish that the amount of televi- journals. However, personal factors are When research activities are non-veri- sion a young child watches is positively also highly influential, with attend- fiable, it is optimal for the financing related to the amount of precipitation ings less likely to seek advice from col- company to obtain the option right to in the child’s community. This sug- leagues who differed from them in terminate the research agreement while gests that, if television is a trigger for gender, experience (age), sub-specialty, maintaining broad property rights to autism, then autism should be more medical school rank, and journal read- the terminated project. The option prevalent in communities that receive ership. Women are more likely to name right induces the biotechnology firm substantial precipitation. Next they men as advisors than men are to name researchers not to deviate from the pro- look at county-level autism data for women . posed research activities. The contract three states — California, Oregon, and Cutler, Huckman, and Kolstad prevents opportunistic exercise of the

42 NBER Reporter Winter 2006/7 termination right by conditioning pay- occurs in contractually dificult envi- The additional empirical results allow ments on the termination of the agree- ronments in which there is no specifi- them to distinguish the property-rights ment. Using a new dataset on 584 bio- able lead product candidate. They also explanation from alternative stories, technology research agreements, the analyze how the contractual design based on uncertainty and asymmetric researchers find that the assignment varies with the R and D firm’s finan- information about the project quality of termination and broad intellectual cial constraints and research capacities or research abilities. property rights to the financing firm and with the type of financing firm.

Entrepreneurship Working Group Meeting

The Entrepreneurship Working World Bank; and Chris Woodruff, Discussant: Dirk Bergemann, Yale Group met in Cambridge on October University of California, San Diego, University 20. Group Director Josh Lerner, NBER “Returns to Capital in Microenterprises: and Harvard Business School, organized Evidence from a Field Experiment” Panel Discussion: “Where is the Venture this program: Discussant: Shawn Cole, Harvard Capital Going? And Does it Matter?” University Paul Gompers, Harvard University Steven J. Davis, NBER and University and NBER; Bill Helman, Greylock of Chicago; John Haltiwanger, NBER Thomas Hellmann, University of Partners; and Philip Rotner, MIT and University of Maryland; and Javier British Columbia; Laura Bottazzi, Miranda and Ron Jarmin, Bureau of Bologna University; and Marco Da Rin, Yael Hochberg, Northwestern the Census, “Volatility and Dispersion Tilburg University, “The Importance University; and Alexander Ljungvist in Business Growth Rates: Publicly of Trust for Investment: Evidence from and Yang Lu, New York University, Traded versus Privately Held Firms” Venture Capital” “Networking as a Barrier to Entry and (NBER Working Paper No. 12354) Discussant: Rebecca Zarutskie, Duke the Competitive Supply of Venture Discussant: Richard Caves, Harvard University Capital” University Discussant: Toby Stuart, Harvard Morten Sorensen, University of University Suresh De Mel, University of Chicago, “Learning by Investing: Peradeniya; David McKenzie, The Evidence from Venture Capital”

Davis and his co-authors study the using the LBD, but show that its impact ing the return to capital is complicated variability of business growth rates in the is overwhelmed by declining volatility by unobserved factors such as entre- U.S. private sector from 1976 onwards. among privately held firms. This pattern preneurial ability and demand shocks, They exploit the recently developed holds in every major industry group. which are likely to be correlated with Longitudinal Business Database (LBD), Employment shifts toward older busi- capital stock. De Mel, McKenzie, and which contains annual observations on nesses account for 27 percent or more Woodruff use a randomized experiment employment and payroll for all U.S. of the volatility decline among privately to overcome this problem, and to mea- businesses. Their central finding is a large held firms. Simple cohort effects that sure the return to capital for the average secular decline in the cross-sectional dis- capture higher volatility among more microenterprise in their sample, regard- persion of firm growth rates and in the recently listed firms account for most of less of whether or not it applies for average magnitude of firm-level volatil- the volatility rise among publicly traded credit. The researchers provide cash and ity. Measured as in other recent research, firms. equipment grants to small firms in Sri the employment-weighted mean vola- Small and informal firms account Lanka, and measure the increase in prof- tility of firm growth rates has declined for a large share of employment in devel- its arising from this exogenous (posi- by more than 40 percent since 1982. oping countries. The rapid expansion tive) shock to capital stock. They find This stands in sharp contrast to pre- of microfinance services is based on the average real return to capital to be vious findings of rising volatility for the belief that these firms have produc- around 4 percent per month, substan- publicly traded firms in COMPUSTAT tive investment opportunities and can tially higher than the market interest data. The researchers confirm the rise in enjoy high returns to capital if given rate. They then use the heterogeneity volatility among publicly traded firms the opportunity. However, measur- of treatment effects to explore whether

NBER Reporter Winter 2006/7 43 missing credit markets or missing insur- tion, a variety of transaction costs, and acterized by strong relationships and ance markets are the most likely cause even investor and company fixed effects. networks, rather than arm’s length, of the high returns. Returns vary with They also consider the relationship spot-market transactions. Hochberg, entrepreneurial ability and with mea- between trust and contracts and find Ljungvist and Lu examine the potential sures of other sources of cash within the no evidence that sophisticated contracts entry-deterring effects of this organiza- household, but not with risk aversion or can be used to overcome lack of trust. tional choice in the context of relation- uncertainty. The researchers therefore They conclude that trust is a fundamen- ships established when VCs syndicate conclude that credit constraints are the tal force driving investment choices. portfolio company investments using main reason for the high returns. Venture capital investors (VCs) can U.S. data for the period 1980 to 2003. The social capital literature finds a create value by actively exploring new The results show that networking does positive relationship between trust and investment opportunities to learn about help reduce entry: VC markets with economic growth or trade. Yet the use their returns. In traditional financial more extensive networking among the of macro-level data makes it difficult markets, a free-rider problem reduces incumbent players experience less entry, to identify the direction of causality. exploration and learning, but VCs’ orga- and the economic effect is sizeable. Hellmann, Bottazzi, and Da Rin exam- nizational structure may limit informa- However, potential entrants can use their ine hand-collected micro data on the tion spillovers and reduce this prob- prior relationships with the incumbents patterns of venture capital investments, lem. Sorensen presents a basic model of as well as previous investment experi- where the trust between investors and learning, based on the statistical Multi- ence in the industry or state to overcome companies’ countries is clearly exoge- Armed Bandit model. The value of an this barrier to entry. The researchers nous. The researchers find that trust investment consists of both its immedi- also document that companies seeking among nations has a significant effect ate return and an option value of learn- venture capital raise money on worse on the likelihood that a venture capital- ing. When he estimates the model, it terms in more densely networked mar- ist invests in a company. This holds even turns out that VCs who explore more kets, and that increased entry into a mar- after accounting for alternative factors, have higher returns. ket is associated with companies receiv- such as geographic distance, informa- Many financial markets are char- ing increased valuations.

International Finance and Macroeconomics The NBER’s Program on International University; Rene M. Stulz, Ohio State Exchange Rate Regimes” Finance and Macroeconomics met in University and NBER; and Frank Discussant: Christian Broda, University Cambridge on October 27. Research E. Warnock, University of Virginia of Chicago and NBER Associates Charles Engel and Linda and NBER, “Financial Globalization, Tesar, organized the program. The fol- Governance, and the Evolution of the Craig Burnside, Duke University lowing papers were discussed: Home Bias”(NBER Working Paper No. and NBER; Martin Eichenbaum 12389) and Sergio Rebelo, Northwestern Enrique G. Mendoza, International Discussant: Joshua Coval, Harvard University and NBER; and Issac Monetary Fund and Nber; Vincenzo University and NBER Kleshchelski, Northwestern University, Quadrini, University of Southern “The Returns to Currency Speculation” California and NBER; and Victor Ricardo J. Caballero and Guido Discussant: Eric van Wincoop, Rios‑Rull, University of Pennsylvania Lorenzoni, MIT and NBER, “Persistent University of Virginia and NBER and NBER, “Financial Integration, Appreciations, Overshooting, and Financial Deepness and Global Optimal Exchange Rate Interventions” Marianne Baxter, Boston University Imbalances” Discussant: Enrique G. Mendoza and NBER, “International Risk Sharing Discussant: Manuel Amador, Stanford in the Short Run and the Long Run” University and NBER Michael W. Klein, Tufts University Discussant: Fabrizio Perri, New York and NBER; and Jay C. Shambaugh, University and NBER Bong‑Chan Kho, Seoul National Dartmouth College, “The Nature of

Large global financial imbal- believe. Instead, Mendoza and his co- of financial integration when coun- ances need not be the harbinger of a authors show that large and persistent tries have different financial markets world financial crash as many authors global imbalances can be the outcome characteristics. In particular, countries

44 NBER Reporter Winter 2006/7 with more advanced financial markets tors in Korean firms is inversely related study the exchange rate effect of fixed accumulate foreign liabilities vis-a-vis to insider ownership and that the firms exchange rates. Fixed exchange rates countries with less developed financial that attract the most foreign portfolio exhibit greater bilateral exchange rate systems in a gradual, long-lasting pro- equity investment are large firms with stability today and in the future. Pegs cess. Moreover, differences in financial dispersed ownership. also display lower multilateral volatil- development affect the composition Most economies experience epi- ity, which may explain why exchange of foreign portfolios, so that a country sodes of large real exchange rate appreci- rate regimes have an effect on a number with negative net foreign asset posi- ations, when the question arises whether of different macroeconomic variables. tions can receive positive factor pay- there is a need for intervention to pro- Currencies that are at a forward ments. Three empirical observations tect the export sector. Caballero and premium tend to depreciate. This “for- support these arguments: 1) financial Lorenzoni present a model of irrevers- ward-premium puzzle”’ represents an deepness varies widely even amongst ible export destruction where exchange egregious deviation from uncovered industrial countries, with the United rate stabilization may be justified if the interest parity. Burnside and his co- States ranking at the top; 2) the sec- export sector is financially constrained. authors document the properties of ular decline in the U.S. net foreign However, the criterion for interven- returns to currency speculation strat- assets position started with a gradual tion is not whether there are inefficient egies that exploit this anomaly. The process of financial markets liberaliza- bankruptcies or not, but whether these first strategy, known as the carry trade, tion; and 3) net exports and current can cause a large exchange rate over- is widely used by practitioners. This account balances are negatively corre- shooting (and real wage decline) once strategy involves selling currencies for- lated with indicators of financial mar- the factors behind the appreciation sub- ward that are at a forward premium kets development. side. The optimal policy often involves and buying currencies forward that Despite the disappearance of for- a mild initial intervention followed by are at a forward discount. The second mal barriers to international invest- an increasingly aggressive stabilization strategy relies on a particular regres- ment across countries, Kho and his as the appreciation persists and the sion to forecast the payoff to selling co-authors find that the average home financial resources of the export sector currencies forward. The researchers bias of U.S. investors towards the 46 dwindle. In some instances, the policy show that these strategies yield high countries with the largest equity mar- also involves an exacerbation of the ini- Sharpe ratios which are not a compen- kets did not fall from 1994 to 2004 tial overshooting during the deprecia- sation for risk. However, these Sharpe if countries are equally weighted but tion phase. ratios do not represent unexploited fell if countries are weighted by mar- The impermanence of fixed exchange profit opportunities. In the presence ket capitalization. This is inconsistent rates has become a stylized fact in inter- of microstructure frictions, spot and with portfolio theory explanations of national finance. The combination of forward exchange rates move against the home bias, but is consistent with a view that pegs do not really peg with traders as they increase their positions. what we call the optimal insider owner- the “fear of floating” view that floats The resulting “price pressure” drives ship theory of the home bias. Since for- do not really float generates the con- a wedge between average and mar- eign investors can only own shares not clusion that exchange rate regimes are, ginal Sharpe ratios. The authors argue held by insiders, there will be a large in practice, unimportant for the behav- that marginal Sharpe ratios are zero home bias towards countries in which ior of the exchange rate. This is consis- even though average Sharpe ratios are insiders own large stakes in corpora- tent with evidence on the irrelevance positive. tions. Consequently, for the home bias of a country’s choice of exchange rate Baxter extends and refines the to fall substantially, insider ownership regime for general macroeconomic study of international risk-sharing in has to fall in countries where it is high. performance. Recently, though, more two dimensions. First, she investigates Poor governance leads to concentrated studies have shown that the exchange risk-sharing at different horizons. In insider ownership, so that governance rate regime does matter in some con- other words, countries might pool risks improvements make it possible for cor- texts. Klein and Shambaugh attempt associated with high-frequency shocks porate ownership to become more dis- to reconcile the perception that fixed (for example, seasonal fluctuations persed and for the home bias to fall. exchange rates are only a “mirage” with in crop yields) but might not share The researchers find that the home the recent research showing the effects risks associated with low frequency bias of U.S. investors decreased most of fixed exchange rates on trade, mon- shocks (for example, different long- towards countries in which the owner- etary autonomy, and growth. First they run national growth rates). Second, ship by corporate insiders was low and demonstrate that, while pegs frequently she studies bilateral risk-sharing, which countries in which ownership by cor- break, many do last and those that break is different from the approach taken porate insiders fell. Using firm-level tend to reform, so a fixed exchange in most previous studies of risk-shar- data for Korea, they find that portfo- rate today is a good predictor that one ing. Her focus on bilateral risk-shar- lio equity investment by foreign inves- will exist in the future. Second, they ing stems from the observation that,

NBER Reporter Winter 2006/7 45 because of such factors as financial with others. Baxter compares direct tive in evaluating the extent of interna- linkages, common cultural linkages, tests of risk-sharing to indirect tests tional risk-sharing. or simply proximity, countries might and finds that the indirect tests com- share risks with some countries but not mon in the literature are less informa-

Public Economics The NBER’s Program on Public Jonathan Gruber, MIT and NBER, Stephen Coate, Cornell University Economics met in Cambridge on and Daniel M. Hungerman, University and NBER, and Brian Knight, Brown November 2–3. Joshua Rauh and Austan of Notre Dame and NBER, “The University and NBER, “Socially Goolsbee, NBER and University of Church vs. The Mall: What Happens Optimal Districting: A Theoretical and Chicago, organized the meeting. These When Religion Faces Increased Secular Empirical Exploration” papers were discussed: Competition?”(NBER Working Paper Discussant: Richard Holden, MIT No. 12410) Mark Duggan, University of Maryland Discussant: Edward Glaeser, Harvard Woodrow T. Johnson, University of and NBER; Perry Singleton, University University and NBER Oregon, and James M. Poterba, MIT of Maryland; and Jae Song, Social and NBER, “Taxes and the Trading Security Administration, “Aching to Joseph E. Stiglitz, Columbia University Behavior of Mutual Fund Investors Retire? The Rise in the Full Retirement and NBER, and Anton Korinek, around Fund Distribution Dates” Age and its Impact on the Disability “Dividend Taxation and Intertemporal Discussant: William Gentry, Williams Rolls”(NBER Working Paper No. Tax Arbitrage” College 11811) Discussant: Raj Chetty, University of Discussant: Jonathan Gruber, MIT and California, Berkeley and NBER Leora Friedberg, University of Virginia NBER and NBER, and Anthony Webb, Stefania Albanesi, Columbia University Center for Retirement Research, “Life Jeffrey Liebman, Harvard University and NBER, “Optimal Taxation of is Cheap: Using Mortality Bonds to and NBER, and Emmanuel Saez, Entrepreneurial Capital with Private Hedge Aggregate Mortality Risk” University of California, Berkeley Information” Discussant: Jeffrey Brown, University of and NBER, “Earnings Responses to Discussant: Aleh Tsyvinski, Harvard Illinois and NBER Increases in Payroll Taxes” University and NBER Discussant: Austan Goolsbee

The Social Security Amendments SSDI enrollment by exploiting varia- ings records from the Social Security of 1983 reduced the generosity of ben- tion across birth cohorts in the pol- Administration for 1981 to 1999 to efits for retired workers in the United icy-induced reduction in the present analyze earnings responses to increases States by increasing the program’s value of retired worker benefits. They in tax rates and to inform discussions full retirement age from 65 to 67 and find that the Amendments significantly about the likely effects of raising the increasing the penalty for claiming increased SSDI enrollment since 1983, Social Security taxable maximum. The benefits at the early retirement age with an additional 0.6 percent of men earnings distribution of workers around of 62. These changes were phased in and 0.9 percent of women between the the current taxable maximum is incon- gradually, so that individuals born in, ages of 45 and 64 receiving SSDI ben- sistent with an annual model in which or before, 1937 were unaffected and efits in 2005 as a result of the changes. people are highly responsive to the pay- those born in 1960 or later were fully Their results further indicate that these roll tax rate, even in the subset of self- affected. No corresponding changes effects will continue to increase dur- employed individuals. Panel data on were made to the program’s disabled ing the next two decades, as those fully married men with high earnings dis- worker benefits, and thus the relative exposed to the reduction in retirement play a tremendous increase in earnings generosity of Social Security Disability benefit generosity reach their fifties relative to other groups over the 1980s Insurance (SSDI) benefits increased. and early sixties. and 1990s, with no clear breaks around Duggan and his co-authors investi- Liebman and Saez use SIPP data the key tax reforms. This suggests that gate the effect of the Amendments on matched to longitudinal uncapped earn- other income groups cannot serve as

46 NBER Reporter Winter 2006/7 a control group for the high earn- Stiglitz and Korinek develop a of their capital to outside investors, ers. This analysis does not support the life-cycle model of the firm to ana- returns to externally owned capital are finding of a large behavioral response lyze the effects of dividend tax policy subject to double taxation: at the level to taxation by wives of high earners. on aggregate investment. They find of the entrepreneur and at the level of The researchers actually find a decrease that new firms raise less equity and the outside investors. Even if entrepre- in the labor supply of wives of high invest less as the level of dividend taxes neurs can purchase private insurance earners around both the 1986 and the increases, in accordance with the tra- against their idiosyncratic risk, optimal 1993 tax reforms, which they attribute ditional view of dividend taxation. asset taxes are essential to implement to an income effect attributable to the However, the dividend tax rate is irrel- the constrained-efficient allocation if surge in primary earnings at the top. evant for the investment decisions of entrepreneurial portfolios are private Policy simulations suggest that with an internally growing and mature firms, as information. earnings elasticity of 0.5, lost income postulated by the new view of dividend Coate and Knight investigate the tax revenue and increased deadweight taxation. Since aggregate investment is problem of optimal districting in the loss would swamp any benefits from dominated by these latter two catego- context of a simple model of legisla- the increase in payroll tax revenue. In ries, the level of dividend taxation as tive elections. In the model, districting contrast, with an elasticity of 0.2, the well as unanticipated changes in divi- matters because it determines the seat- ratio of the gain in OASDI revenue dend tax rates have only a minor impact vote curve, which describes the rela- to lost income tax revenue and dead- on aggregate investment and output. tionship between seats and votes. The weight loss would be much greater. Anticipated dividend tax changes, on paper first characterizes the optimal Gruber and Hungerman identify the other hand, allow firms to engage seat-vote curve and shows that, under a a policy-driven change in the oppor- in inter-temporal tax arbitrage so as to weak condition, there exist districtings tunity cost of religious participation reduce investors’ tax burden. This can that generate this ideal relationship. based on state laws that prohibit retail significantly distort aggregate invest- The paper then develops an empirical activity on Sunday, known as “blue ment. Anticipated tax cuts (increases) methodology for computing seat-vote laws.” Many states have repealed these delay (accelerate) firms’ dividend pay- curves and measuring the welfare gains laws in recent years, raising the oppor- ments, which leads them to hold higher from implementing optimal districting. tunity cost of religious participation. (lower) cash balances and, for capital This is applied to the districting plans The researchers construct a model that constrained firms, can significantly in- used to elect U.S. state legislators dur- predicts, under fairly general condi- crease (decrease) aggregate investment ing the 1990s. tions, that allowing retail activity on for periods after the tax change. Capital gain distributions by open- Sundays will lower attendance levels Albanesi studies optimal taxation end mutual funds accelerate the cap- but may increase or decrease religious of entrepreneurial capital and finan- ital gains tax liability of their tax- donations. They then use a variety cial assets in economies with private able investors, thereby reducing their of datasets to show that when a state information. Returns to entrepreneur- after-tax return. Investors can reduce repeals its blue laws, religious atten- ial capital are risky and depend on this acceleration by delaying the pur- dance falls, and church donations and entrepreneurs’ hidden effort. The idio- chase of fund shares until after capi- spending fall as well. These results do syncratic risk in capital returns implies tal gain distributions tax place. Tax- not seem to be driven by declines in that the intertemporal wedge on entre- exempt investors, including individual religiosity prior to the law change, nor preneurial capital that characterizes investors who hold mutual funds in are comparable declines in membership constrained-efficient allocations can be their IRAs, Keogh plans, and 401(k) or giving to nonreligious organizations positive or negative. The properties accounts, have no such disincentives to after a state repeals its laws observed. of optimal marginal taxes on entre- invest prior to distributions. Johnson The authors then assess the effects of preneurial capital depend on the sign and Poterba examine data on fund changes in these laws on drinking and of this wedge. If the wedge is posi- inflows and outflows, and account drug use reported in the NLSY. They tive, the optimal marginal capital tax openings and closures, at a sample of find that repealing blue laws leads to is decreasing in capital returns, while open-end mutual funds offered by a an increase in drinking and drug use, the opposite is true when the wedge single fund complex. The data display and that this increase is found only is negative. Optimal marginal taxes on a significant increase in inflows to tax- among the initially religious individu- other assets depend on their correla- able accounts, and an increase in the als who were affected by the blue laws. tion with idiosyncratic capital returns. number of new accounts opened, in the The effect is economically significant; The optimal tax system equalizes after weeks following distribution dates rela- for example, the gap in heavy drink- tax returns on all assets, thus reduc- tive to the weeks prior to distributions. ing between religious and non-religious ing the variance of after tax returns These findings are consistent with the individuals falls by about half after the on capital relative to other assets. If hypothesis that at least some taxable laws are repealed. entrepreneurs are allowed to sell shares investors adapt their portfolio trades to

NBER Reporter Winter 2006/7 47 reduce their tax liabilities. be extended to private pension provid- providers could deal with aggregate Insurance companies, employers, ers and Social Security. Given the fore- mortality risk more efficiently by trans- and the U.S. government all provide casts of the Lee-Carter model, they ferring it to financial markets through annuities and therefore assume aggre- calculate that a markup of 3.9 percent mortality-contingent bonds. The gate mortality risk. Using the widely- on an annuity premium(or sharehold- researchers calculate the returns that cited Lee-Carter mortality model, ers’ capital equal to 3.9 percent of the one recently-proposed mortality bond Friedberg and Webb quantify aggre- expected present value of annuity pay- would have paid had it been available gate mortality risk as the risk that the ments) would be required to reduce over a long period. Using both the average annuitant lives longer than is the probability of insolvency resulting Capital and the Consumption Capital predicted by the model, and deter- from aggregate mortality shocks to 5 Asset Pricing Models, they determine mine that annuities expose providers percent, and a markup of 5.7 percent the risk premium that investors would to substantial risk. They also focus on would reduce the probability of insol- have required to hold the bond. At other recent actuarial forecasts, some vency to 1 percent. Based on the same plausible coefficients of risk aversion, of which lie at the edge or outside of model, they find that a projection scale annuity providers should be able to the 95 percent confidence interval of commonly referred to by the insur- hedge aggregate mortality risk via such Lee-Carter. They focus on the impli- ance industry underestimates aggregate bonds at very low cost. cations of aggregate mortality risk for mortality improvements and would insurance companies; this analysis can leave annuities underpriced. Annuity

Monetary Economics

The NBER’s Program on Monetary Olivier Blanchard, MIT and NBER, University of Pennsylvania and Economics met in Cambridge on and Jordi Gali, Universitat Pompeu NBER, “Aggregate Shocks or November 3. Christopher House Fabra and NBER, “A New Keynesian Aggregate Information? Costly and Matthew D. Shapiro, NBER and Model with Unemployment” Information and Business Cycle University of Michigan, organized the Discussant: Robert E. Hall, Stanford Comovement” meeting, at which these papers were University and NBER Discussant: Mirko Wiederholt, discussed: Northwestern University Monika Piazzesi, University of Michael Elsby and Gary Solon, Chicago and NBER, and Martin and Jon Steinsson, University of Michigan and NBER, Schneider, New York University, Harvard University, “Five Facts About and Ryan Michaels, University of “Inflation and the Price of Real Prices: A Reevaluation of Menu Cost Michigan, “Reassessing the Ins and Assets” Models” Outs of Unemployment Again: Discussant: Robert B. Barsky, Discussant: Mark Bils, University of Everyone’s a Winner” University of Michigan and NBER Rochester and NBER Discussant: Robert Shimer, University of Chicago and NBER Laura Veldkamp, New York University, and Justin Wolfers,

Responding to Shimer’s already- tion. But contrary to Shimer’s con- examination of particular components influential manuscript “Reassessing the clusions, they find that even his own of the inflow to unemployment. Ins and Outs of Unemployment,” Elsby, methods and data, when viewed in an Blanchard and Gali develop a util- Solon, and Michaels reconsider the appropriate metric, reveal an important ity based model of fluctuations, with extent to which the increased unem- role for increased inflows to unemploy- nominal rigidities and unemployment. ployment during a recession arises from ment as well. This finding is further In doing so, they combine two strands an increase in the number of unem- strengthened by their refinements of of research: the New Keynesian model, ployment spells versus an increase in Shimer’s methods of correcting for data with its focus on nominal rigidities, their duration. Like Shimer, they find problems and by an extension of his and the Diamond-Mortensen-Pissarides an important role for increased dura- approach that enables a more detailed model, with its focus on labor mar-

48 NBER Reporter Winter 2006/7 ket frictions and unemployment. Their inflation, both of which lowered the effects of common shocks, relative to analysis proceeds in two steps. First, overall propensity to save. They also sectoral shocks. they leave nominal rigidities aside and show that the Great Inflation led to a Nakamura and Steinsson establish show that, under a standard utility portfolio shift by making housing more five facts about prices in the U.S. econ- specification, productivity shocks have attractive than equity. Apart from tax omy: 1) The median duration of con- no effect on unemployment in the con- effects, a new channel is that disagree- sumer prices when sales are excluded strained efficient allocation. Then they ment about inflation across age groups at the product level is 11 months. The focus on the implications of alterna- drives up collateral prices when credit median duration of finished goods pro- tive real wage setting mechanisms for is nominal. ducer prices is 8.7 months. 2) One- fluctuations in unemployment. Next When similar patterns of expan- third of regular price changes are price they introduce nominal rigidities in the sion and contraction are observed decreases. 3) The frequency of price form of staggered price setting by firms. across sectors, we call this a business increases responds strongly to inflation They derive the relationship between cycle. Yet explaining the similarity and while the frequency of price decreases inflation and unemployment and dis- synchronization of these cycles across and the size of price increases and price cuss how it is influenced by the pres- industries remains a puzzle. Whereas decreases do not. 4) The frequency of ence of real wage rigidities. Finally, output growth across industries is price change is highly seasonal: it is they show the nature of the tradeoff highly correlated, identifiable shocks, highest in the first quarter and lowest between inflation and unemployment like shocks to productivity, are far less in the fourth quarter. 5) The hazard stabilization and draw the implications correlated. Previous work has exam- function of price changes for individual for optimal monetary policy. ined complementarities in production, consumer and producer goods is down- In the 1970s, U.S. asset markets but Veldkamp and Wolfers propose ward sloping for the first few months witnessed: 1) a 25 percent dip in the that sectors make similar input deci- and then flat (except for a large spike at ratio of aggregate household wealth rel- sions because of complementarities in 12 months in consumer services and all ative to GDP; and 2) negative co-move- information acquisition. Because infor- producer prices). These facts are based ment of house and stock prices that mation about driving forces has a high on CPI microdata and a new compre- drove a 20 percent portfolio shift out fixed cost of production and a low mar- hensive dataset of microdata on pro- of equity into real estate. Piazzesi and ginal cost of replication, it can be more ducer prices that they construct from Schneider use an overlapping genera- efficient for firms to share the cost raw production files underlying the tions model with uninsurable nominal of discovering common shocks than PPI. They show that the first, second, risk to quantify the role of structural to invest in uncovering detailed sec- and third facts are consistent with a change in these events. They attribute toral information. Firms basing their benchmark menu-cost model, while the the dip in wealth to the entry of baby decisions on this common informa- fourth and fifth facts are not. boomers into asset markets, and to the tion make highly correlated production erosion of bond portfolios by surprise choices. This mechanism amplifies the

NBER Reporter Winter 2006/7 49 Higher Education The NBER’s Working Group on Marko Tervio, University of California, Technological Knowledge to and from Higher Education met in Cambridge on Berkeley, “Network Analysis of Three American Universities” November 9. Working Group Director Academic Labor Markets” Discussant: William Kerr, Harvard Charles T. Clotfelter of Duke University Discussant: Richard Jensen, University University organized the meeting. These papers were of Notre Dame discussed: Zeynep Hansen, Washington Brian C. Cadena and Benjamin J. University and NBER; Hideo Owan, Susan M. Dynarski, Harvard University Keys, University of Michigan, “Self- Aoyama Gakuin University; and and NBER, and Judith E. Scott- Control Induced Debt Aversion: Jie Pan, Washington University, Clayton, Harvard University, “The Evidence from Interest-Free Student “The Impact of Group Diversity Cost of Complexity in Federal Student Loans” on Performance and Knowledge Aid: Lessons from Optimal Tax Theory Discussant: Ofer Malamud, University Spillover:An Experiment in a College and ” (NBER of Chicago Classroom”(NBER Working Paper No. Working Paper No. 12227) 12251) Discussant: Eric Bettinger, Case Western Megan MacGarvie, Boston University Discussant: Jacob Vigdor, Duke Reserve University and NBER, “Foreign Students University and NBER and the Diffusion of Scientific and

The federal system for distribut- sured by a similar method to that used Thus, debt-averse behavior arises even ing student financial aid rivals the tax by Google to rank web pages. In all among consumers who have no direct code in its complexity. Both have been disciplines, the distribution of influ- distaste for debt. Using the 2003-4 a source of frustration and a focus ence is significantly more skewed than wave of the National Post-Secondary of reform efforts for decades, yet the the distribution of academic place- Student Aid Study (NPSAS), the complexity of the student aid system ments because of a strong hierarchy of authors investigate students’ financial has received comparatively little atten- schools in which movements are sel- aid situations and subsidized loan take- tion from economists. Dynarski and dom upwards. This hierarchy is stron- up decisions. They exploit an institu- Scott-Clayton describe the complex- gest in economics. Tervio also finds tional detail of the financial aid pro- ity of the aid system, and apply lessons that, in economics, there are clusters of cess to identify a group of students who from optimal tax theory and behavioral departments that are significantly more should be especially vulnerable to self- economics to show that complexity is connected within than with each other. control problems. Their results suggest a serious obstacle to both efficiency These clusters are consistent with anec- that consumers choose to limit their and equity in the distribution of stu- dotal evidence about Freshwater and liquidity in economically meaningful dent aid. They show that complexity Saltwater schools of thought. There is situations, consistent with the predic- disproportionately burdens those with a similar but weaker division within tions of the behavioral model. the least ability to pay and undermines comparative literature, but not within MacGarvie combines counts of the redistributive goals. They use detailed mathematics. number of Science and Engineering data from federal student aid applica- Cadena and Keys use insights doctorates by country of origin at U.S. tions to show that a radically simplified from behavioral economics to offer universities with data on citations to aid process can reproduce the current an explanation for a particularly sur- and from U.S. universities’ patents to distribution of aid using a fraction of prising borrowing phenomenon: nearly study the relationship between labor the information now collected. 20 percent of undergraduate students mobility and international patterns of Tervio analyzes the academic labor who are offered interest-free loans turn diffusion of scientific and technologi- market as a citation network, where them down. The authors present a for- cal knowledge. Preliminary findings departments gain citations by plac- mal model of the financial aid pro- suggest that knowledge diffuses from ing their Ph.D. graduates into the fac- cess emphasizing that a rational agent U.S. universities to foreign countries ulty of other departments. The aim is would not reject interest-free student when doctoral recipients migrate inter- to measure the distribution of influ- loans because doing so requires forego- nationally, and there is some evidence of ence and the possible division into ing a significant government subsidy. A foreign knowledge acquisition by U.S. clusters between academic depart- student with time-inconsistent prefer- universities when doctoral recipients ments in three disciplines (economics, ences, however, may optimally choose move abroad. However, there appears mathematics, and comparative litera- to turn down subsidized loans to avoid to be little evidence that foreign coun- ture). Departmental influence is mea- excessive consumption during school. tries benefit from improved access to

50 NBER Reporter Winter 2006/7 U.S. science and technology contained the self-selection problem. Their results their own group contract form — mem- in patents when doctoral recipients indicate that male-dominant groups bers of “autonomous” groups receive remain in the U.S. after graduation. performed worse both in group work equal grade for their group work while Hansen, Owan, and Pan combine and in individually taken exams than those in “democratic” groups can adopt class performance data from an under- female-dominant and equally-mixed differentiated point allocation, thus graduate class with students’ personal gender groups. Individual members providing a proper mechanism to pun- records to explore diversity and knowl- from a group with more diversity in ish free riders. The estimation results edge spillover effects. A major advan- age and gender scored higher in exams. show a significant correlation between tage of their dataset is the exogenous Another novel aspect of this natural the choice of a democratic contract and assignment of groups, which rules out experiment is that each group chooses the group and individual performance.

Education Program Meeting

The NBER’s Program on Education, Impact of School Entry Laws on Moshe Justman, Ben Gurion University, directed by Caroline M. Hoxby of Educational Attainment and Labor and Yaakov Gilboa, Sapir Academic Harvard University, met in Cambridge Market Outcomes” College, “Equal Opportunity in on November 9 and 10. These papers Education: Lessons from the Kibbutz” were presented and discussed at the Philip Babcock, University of meeting: California, Santa Barbara, “From Ties Adalbert Mayer and Steven Puller, to Gains? Evidence on Connectedness, Texas A&M University, “The Old Boy Mark Hoekstra, University of Skill Acquisition, and Diversity” (and Girl) Network: Social Network Pittsburgh, “The Effect of Attending the Formation on University Campuses” Flagship State University on Earnings: A Andrea Ichino, European University Regression Discontinuity Approach” Institute; Pietro Garibaldi, University Sally Kwak, University of of Turin; Francesco Giavazzi, MIT and Hawaii-Manoa, “The Impact of Carlos Dobkin, University of NBER; and Enrico Rettore, University Intergovernmental Incentives on California, Santa Cruz, and Fernando of Padova, “College Cost and Time to Disability Rates and Special Education Ferreira, University of Pennsylvania, Obtain a Degree: Evidence from Tuition Spending” “Should We Care About the Age at Discontinuities” Which Children Enter School? The

By combining confidential admis- least squares estimates. However, he finds right after the cut-off date. This effect is sions records from a large state univer- no effect on earnings for white women significantly larger for minorities and the sity with earnings data collected through generally and only weak evidence of a pos- children of parents with less than a high the state’s Unemployment Insurance pro- itive effect for white women with strong school education. The researchers also gram, Hoekstra examines the effect of attachment to the labor force. find that almost one third of the initial attending the flagship state university on Dobkin and Ferreira examine how differences in enrollment rates disappear the earnings of 28–33 year-olds. To dis- state laws regulating the age at which stu- by ninth grade because the youngest chil- tinguish this effect from the effects of dents can enter school affect children’s dren in a cohort are considerably more confounding factors correlated with the progression through school, adult educa- likely to be held back a grade during ele- university’s admission decision, and/or tional attainment, and labor market out- mentary school. Despite the striking dif- the applicant’s enrollment decision, he comes. Using the exact day of birth from ferences in the timing of enrollment and uses a regression discontinuity approach the 2000 Long Form Decennial Census the rates at which students are held back, along with a conventional instrumental for the states of California and Texas, the the authors find only a modest impact on variable approach. The results indicate authors first show that students born right adult educational attainment, with indi- that attending the most selective state before the cut-off date for school enroll- viduals who enter school a year earlier university causes earnings to be at least 10 ment are 50–60 percentage points more having only about one percentage point percent higher for white men, an effect likely to enroll in kindergarten a year ear- increase in their probability of complet- that is considerably higher than ordinary lier than similar students that were born ing high school. They find slightly larger

NBER Reporter Winter 2006/7 51 effects for Hispanics and older cohorts. lar end. Using a Regression Discontinuity than would arise from the random selec- However, the additional education result- Design on data from Bocconi University tion of friends, even after controlling for ing from early school entry does not result in Italy, Ichino and his co-authors show a variety of measures of socioeconomic in differences in employment rates or that if tuition is raised by 1000 Euros in background, ability, and college activi- wages. This may be because other con- the last year of the program, the probabil- ties. Also, two students are more likely to founding mechanisms, such as the impact ity of late graduation decreases by 6.1 per- be friends if they share the same major, of school entry laws on the age relative to centage points with respect to a bench- participate in the same campus activi- peers, and the absolute age at which an mark average probability of 80 percent. ties, and, to a lesser extent, come from the individual is taught a particular material. The researchers conclude by showing that same socioeconomic background. Next, These results suggest that it is problematic an upward sloping tuition profile may be the authors develop a model of the for- to use school entry laws as an instrument efficient when effort is sub-optimally sup- mation of social networks that decom- for educational attainment when trying to plied in the presence of peer effects. poses the formation of social links into estimate the returns to education. Justman and Gilboa use the unique effects based upon the exogenous school Babcock uses micro-level data on circumstances of education in Israeli kib- environment and effects of endogenous social networks in middle and second- butzim — communal villages — to derive choice arising from preferences for certain ary schools to estimate the effects of con- two related findings. First, by regressing characteristics in one’s friends. They use nectedness on education attainment out- kibbutz members’ test scores on parental student-level data from an actual social comes and the association between racial education, the authors obtain an egalitar- network to calibrate the model, which diversity and connectedness. The analy- ian standard of equal opportunity in edu- generates many of the characteristics com- sis addresses concerns about unobserved cation, measured as origin-independence, mon to social networks. They then simu- neighborhood and school-level hetero- to which the degree of equal opportunity late network structures under alternative geneity by using within-school variation in other education systems can be com- university policies and find that changes between grade cohorts to identify effects pared. Second, by comparing this effect to in the school environment that affect the of connectedness. There are two main its counterpart for the general Israeli pop- likelihood that two students interact have findings: 1) Being part of a more con- ulation, they can quantitatively decom- only a limited potential to reduce the seg- nected cohort within a given second- pose the parental effect on test scores (in mentation of the social network. ary or middle school is associated with the general population) so as to distin- In California in 1998, the state con- significantly higher years of schooling guish between what money can and can- verted from a system that awarded funds attained and higher probability of hav- not buy. based on the number of disabled students ing attended college, seven years later. Mayer and Puller document the in a district to one based on total enroll- 2) Being part of a more racially diverse structure and composition of social net- ment. This change induced changes in the grade cohort, within a given school, is works on university campuses and inves- total funding awarded to different dis- associated with significantly lower levels tigate the processes that lead to their tricts, and reduced the marginal “price” of connectedness — rare micro-level evi- formation. They use a large dataset that of an additional disabled student to zero. dence to augment existing cross-region identifies students in one another’s social Kwak finds that the reform created both evidence on ethnic fractionalization and network on campus and link these data “income” and “substitution” effects on the disconnectedness. to university records on each student’s number of students classified as disabled. For many students throughout the demographic and school outcome char- In the short run, additional state special world, time to obtain an academic degree acteristics. The campus networks exhibit education grants translate into sizeable extends beyond the normal completion common features of social networks, such increases in special education spending, time, while college tuition is essentially as clusteredness. The authors show that but in the longer run special education constant during the years of enrollment race is strongly related to social ties. In funds appear fungible across other spend- and, in particular, does not increase when a particular, blacks and Asians have dis- ing needs. student remains in a program after its regu- proportionately more same race friends

52 NBER Reporter Winter 2006/7 Asset Pricing The NBER’s Program on Asset Robert Novy-Marx, University of Discussant: Jonathan Berk, University of Pricing met at The Wharton School, Chicago, “Investment Cash Flow California, Berkeley and NBER University of Pennsylvania, on November Sensitivity and the Value Premium” 10. NBER researchers Leonid Kogan, Discussant: Joao Gomes, University of Long Chen, Michigan State University, Sloan School of Management at MIT, Pennsylvania and Xinlei Zhao, Kent State University, and Amir Yaron, The Wharton School, “Return Decomposition” organized the meeting. These papers Arvind Krishnamurthy and Annette Discussant: John Heaton, University of were discussed: Vissing-Jorgensen, Northwestern Chicago and NBER University and NBER, “The Demand Stijn Van Nieuwerburgh, New for Treasury Debt” Lubos Pastor, University of Chicago York University and NBER, and Discussant: Monika Piazzesi, University and NBER, and Robert Stambaugh, Pierre‑Olivier Weill, University of of Chicago and NBER University of Pennsylvania and NBER, California, Los Angeles, “Why Has “Predictive Systems: Living with House Price Dispersion Gone Up? Martijn Cremers and Antti Petajisto, Imperfect Predictors” Discussant: Markus K. Brunnermeier, Yale University, “How Active is Your Discussant: Jonathan Lewellen, Princeton University and NBER Fund Manager? A New Measure That Dartmouth College and NBER Predicts Performance”

Nieuwerburgh and Weill investigate resulting from optimal investment behav- yield that investors attribute to Treasury the 30-year increase in the level and disper- ior, generates both investment-cash flow debt. Changes in the supply of Treasury sion of house prices across U.S. metropoli- sensitivity and a countercyclical value pre- debt trace out the demand for convenience tan areas, using a calibrated dynamic gen- mium. Novy-Marx explictly characterizes by investors. The authors further show that eral equilibrium island model. The model the investment strategies of heterogeneous the aggregate demand curve for the conve- is based on two main assumptions: house- firms as rules in industry average-Q that nience provided by Treasury debt is down- holds flow in and out of metropolitan depend on the industry’s concentration ward sloping; they provide estimates of the areas in response to local wage shocks, and and capital intensity. Firms are uncon- elasticity of demand. They also analyze dis- the housing supply cannot adjust instantly strained, investing when and because mar- aggregated data from the Flow of Funds because of regulatory constraints. Feeding ginal-q equals one, but investment is still Accounts of the Federal Reserve and show into the model the 30-year increase in associated strongly with positive cash-flow that individual groups of Treasury bond cross-sectional wage dispersion that is doc- shocks and only weakly with average-Q holders have downward sloping demand umented based on metropolitan-level data, shocks, because firm value is insensitive to curves. Even groups with the most elastic the authors generate the observed increase demand when demand is high. A value pre- demand curves have demand curves that in house price level and dispersion. In mium arises, both within and across indus- are far from flat. The authors discuss the equilibrium, workers flow towards excep- tries, because the market-to-book sorting implications for the behavior of corporate tionally productive metropolitan areas and procedure over-weights the value portfo- bond spreads, interest rate swap spreads, drive house prices up. The calibration also lio with high-cost producers, firms in slow and the value of aggregate liquidity and for reveals that, while a baseline level of regula- growing industries, and firms in industries the financing of the U.S. deficit, Ricardian tion is important, a tightening of regulation that employ irreversible capital, which are equivalence, and the effects of foreign cen- by itself cannot account for the increase in riskier, especially in “bad” times. The two tral bank demand on Treasury yields. house price level and dispersion: in equilib- puzzles are linked directly, with theory pre- To quantify active portfolio manage- rium, workers flow out of tightly regulated dicting value firms should exhibit stron- ment, Cremers and Petajisto introduce a towards less regulated metropolitan areas, ger investment-cash flow sensitivities than new measure they label “Active Share”. It undoing most of the price impact of addi- growth firms. describes the share of portfolio holdings tional local supply regulations. Finally, the Krishnamurthy and Vissing-Jorgensen that differ from the portfolio’s benchmark calibration with increasing wage dispersion show that the U.S. debt/GDP ratio is neg- index. They argue that to determine the suggests that the welfare effects of housing atively correlated with the spread between type of active management for a portfolio, supply regulation are large. corporate bond yields and Treasury bond they need to measure it in two dimensions, Firms’ equilibrium investment behav- yields. The result holds even when con- using both Active Share and tracking error. ior explains two seemingly unrelated eco- trolling for the default risk on corporate They apply this approach to the universe nomic puzzles. Endogenous variation in bonds. The authors argue that the corpo- of all-equity mutual funds to characterize firms’ exposures to fundamental risks, rate bond spread reflects a convenience how much and what type of active man-

NBER Reporter Winter 2006/7 53 agement each practices. The authors test ing forecasting variable can change the bal- with the finding in a related literature that how active management is related to char- ance of the two news components, and the value stocks have higher CF covariation acteristics such as fund size, expenses, and conclusions that compare the two parts. To with aggregate CFs. turnover in the cross-section, and look at illustrate this, the researchers apply their The standard regression approach to the evolution of active management over approach to Treasury bonds, which should investigating return predictability seems time. They find that active management have zero CF variance and zero CF betas. too restrictive in one way but too lax in predicts fund performance: the funds with Instead, they find that the variance of the another. A predictive regression assumes the highest Active Share significantly out- “CF news” is larger than that of the DR that expected returns are captured exactly perform their benchmark indexes, both news, and that bonds with longer maturi- by a set of given predictors but does not before and after expenses, while the non- ties have higher “CF betas.” Applying the exploit the likely economic property that index funds with the lowest Active Share approach to equity returns, they show that innovations in expected returns are nega- underperform. The most active stock pick- the relative importance of CF variance and tively correlated with unexpected returns. ers tend to create value for investors while DR variance of the market portfolio is sen- Pastor and Stambaugh develop an alter- factor bets and closet indexing tend to sitive to the choice of forecasting variables; native framework — a predictive sys- destroy value. and, for most forecasting-variable specifi- tem — that accommodates imperfect pre- Chen and Zhao study the robustness cations, value stocks usually do not have dictors and beliefs about that negative of the return decomposition approach, in higher CF betas. These results run coun- correlation. In this framework, the predic- which unexpected equity return is decom- ter to what is reported using the decom- tive ability of imperfect predictors is supple- posed into discount rate (DR) news and position approach in the current literature. mented by information in lagged returns as cash flow (CF) news. In this approach, DR Further decomposing the CF news in the well as lags of the predictors. Compared to news is modeled directly but CF news is current literature into directly-modeled CF predictive regressions, predictive systems usually backed out as a residual component. news and residual news, the authors show deliver different and substantially more The researchers argue that the approach that opposite conclusions can be drawn precise estimates of expected returns as has serious limitations because CF news, as depending on the nature of the residual well as different assessments of a given pre- a residual, depends critically on how suc- news. They also reconcile their finding that dictor’s usefulness. cessfully DR news is modeled. One miss- value stocks do not have higher CF betas

Corporate Finance

The NBER’s Program on Corporate Discussant: Michael Weisbach, Nicola Gennaioli, Stockholm Finance met in Cambridge on November University of Illinois and NBER University, and Stefano Rossi, 10. NBER Research Associate Andrei Stockholm School of Economics, Shleifer of Harvard University orga- Bart Lambrecht, University of “Optimal Resolutions of Financial nized the meeting. These papers were Lancaster, and Stewart C. Myers, Distress by Contract” discussed: NBER and MIT, “Debt and Managerial Discussant: Kenneth Ayotte, Columbia Rents in a Real‑Options Model of the University Camelia M. Kuhnen, Northwestern Firm” University, and Jeffrey Zwiebel, Discussant: Douglas W. Diamond, Simeon Djankov and Caralee McLiesh, Stanford University, “Executive Pay, University of Chicago and NBER World Bank, and Oliver D. Hart and Hidden Compensation, and Managerial Andrei Shleifer, Harvard University Entrenchment” Douglas Baird, University of Chicago; and NBER, “Debt Enforcement Around Discussant: Xavier Gabaix, MIT and Arturo Bris, Yale University; and Ning the World” NBER Zhu, University of California, Davis, Discussant: David S. Scharfstein, “The Dynamics of Large and Small Harvard University and NBER Jin Xu, University of Chicago, “What Chapter 11 Cases: An Empirical Study” Determines Capital Structure? Evidence Discussant: Michelle J. White, University from Import Competition” of California, San Diego and NBER

Kuhnen and Zwiebel consider a executive compensation, where top man- ject to limited entrenchment, instead of managerial optimal framework for top agement sets its own compensation sub- the conventional setting where such com-

54 NBER Reporter Winter 2006/7 pensation is set by a board that maximizes import penetration could be endogenous claims cast over small Chapter 11 reorga- firm value. Top management would like to capital structure because of strategic nizations, accounts of small Chapter 11 to pay themselves as much as possible, behaviors of firms, Xu uses the indus- must focus squarely on them. but are constrained by the need to ensure try tariff as an instrumental variable for Gennaioli and Rossi theoretically sufficient efficiency to avoid replacement. import penetration. The IV result is con- explore the possibility that parties might Shareholders can remove a manager, but sistent with the OLS result. Robustness efficiently resolve financial distress by only at a cost, and therefore will only do checks with alternative leverage measures contract as opposed to exclusively rely- so if the anticipated future value of the confirm the basic result. The results can ing on state intervention. They character- manager (given by anticipated future per- be explained best by the trade-off theory. ize which financial contracts are optimal formance net future compensation) falls Xu also finds some evidence consistent depending on legal protection of inves- short of that of a replacement by this with the disciplinary role of debt hypoth- tors against fraud, and how efficient is replacement cost. In this setting, observ- esis, but it only accounts for a small por- the resulting resolution of financial dis- able compensation (salary) and hidden tion of the competition-leverage effect. tress. They find that when legal protec- compensation (perks, pet projects, pen- Lambrecht and Myers present a the- tion against fraud is strong, issuing a con- sions) serve different roles for manage- ory of capital investment and debt and vertible debt security to a large, secured ment and have different costs, and both equity financing in a real-options model of creditor allows the parties to attain the are used in equilibrium. The authors a public corporation. The model assumes first best. Conversion of debt into equity examine the relationship between observ- that managers maximize the present value upon default allows the debtor to col- able and hidden compensation and other of their future compensation (managerial lateralize the whole firm to the creditor, variables in a dynamic model, and derive rents), subject to constraints imposed by not just certain physical assets, thereby a number of unique predictions regard- outside shareholders’ property rights to inducing the creditor to internalize the ing these two types of pay. They then test the firms’ assets. The authors show that upside from efficient reorganization. these implications and find results that managers adopt an optimal debt pol- When instead legal protection against generally support the predictions of their icy that generates efficient investment fraud is poor, straight debt with foreclo- model. and disinvestment decisions. Optimal sure is the only feasible contract, even if Xu uses the exogenous shock of debt equals the liquidation value of the it induces over-liquidation. The norma- greater import competition to study firms’ assets and is therefore default-risk tive implication of this analysis is that an the effect of product market competi- free. But managers’ personal wealth efficient resolution of financial distress is tion on corporate capital structure in the constraints can justify additional risky attained under freedom of contracting U.S. domestic textile and apparel sector. debt to fund positive-NPV investments. and strong protection against fraud. Theoretically, when import competition Changes in cash flow can cause changes Djankov, McLiesh, Hart, and increases, expected domestic profitabil- in investment by tightening or loosening Shleifer present insolvency practitioners ity drops, increasing the probability of the wealth constraints. from 88 countries with an identical case bankruptcy and reducing the tax benefit Baird and his co-authors show that of a hotel about to default on its debt, of debt. According to the trade-off the- the dynamics of Chapter 11 turn dramat- and ask them to describe in detail how ory, optimal financial leverage should go ically on the size of the business. The vast debt enforcement against this hotel will down. Xu finds that after a quota-elimi- majority of the assets administered in proceed in their countries. The research- nating trade law took effect, firms in the Chapter 11 are concentrated in a handful ers use the data on time, cost, and the sector significantly de-levered by reducing of large cases, but most of the businesses likely disposition of the assets (preserva- debt and increasing outside equity. The in Chapter 11 are small, and the smaller tion as a going concern versus piecemeal average textile and apparel firm reduced the business, the smaller the distribution sale) to construct a measure of the effi- its leverage by 0.10, a 30 percent reduc- to general unsecured creditors. For busi- ciency of debt enforcement in each coun- tion, while the rest of the manufactur- nesses with assets above $5 million, unse- try. They identify several characteristics ing sector barely de-levered. Xu extends cured creditors typically collect half of of the debt enforcement procedure, such the analysis by looking at the relation of what they are owed. Where the business’s as the structure of appeals and the avail- capital structure to industry-level import assets are worth less than $200,000, ordi- ability of floating charge finance, that penetration in a large sample of all U.S. nary general creditors usually recover influence efficiency. This measure of effi- manufacturing industries. Financial lever- nothing. In the typical small Chapter 11 ciency is strongly correlated with per cap- age is strongly negatively correlated with case, the tax collector is the central figure. ita income and legal origin and predicts industry-level import penetration, con- In small business bankruptcies, priority debt market development across coun- trolling for documented determinants tax liabilities are the largest unsecured tries. Interestingly, it is also highly corre- of leverage. Economically, one standard liabilities of the business. Tax obligations lated with measures of the quality of gov- deviation increase in import penetration are entitled to priority and are obliga- ernment obtained in other studies. corresponds to about a 20 percent stan- tions of both the corporation and those dard deviation decrease in leverage. Since who run it. Given the large shadow tax

NBER Reporter Winter 2006/7 55 Behavioral Economics The NBER’s Working Group on Szilagyi, Harvard University, “In Search Earnings News” Behavioral Economics met in Cambridge of Distress Risk” (NBER Working Paper Discussant: Stefano Della Vigna, on November 11. NBER Research No. 12362) University of California, Berkeley and Associates and Group Directors Robert Discussant: Tyler Shumway, University NBER J. Shiller of Yale University and Richard of Michigan H. Thaler, University of Chicago, orga- Robin Greenwood, Harvard University, nized this program: Andrea Frazzini, University of Chicago, and Stefan Nagel, Stanford University and Owen Lamont, Yale University and and NBER, “Inexperienced Investors Am Tal Fishman and Harrison Hong, NBER, “The Earnings Announcement and Speculative Bubbles” Princeton University, and Jeffrey Premium and Trading Volume” Discussant: Jeremy C. Stein, Harvard D. Kubik, Syracuse University, “Do Discussant: Steven Heston, University University and NBER Arbitrageurs Amplify Economic of Maryland Shocks?” Massimo Massa and Lei Zhang, Discussant: Michael Rashes, Bracebridge David Hirshleifer and Siew Hong INSEAD, “Cosmetic Mergers: The Capital Teoh, University of California, Irvine, Effect of Style Investing on the Market and Sonya Seongyeon Lim, DePaul for Corporate Control” John Y. Campbell, Harvard University University, “Driven to Distraction: Discussant: Malcolm Baker, Harvard and NBER, and Jens Hilscher and Jan Extraneous Events and Underreaction to University and NBER

Fishman, Hong, and Kubik con- failure and the pricing of financially this earnings announcement premium sider whether arbitrageurs amplify fun- distressed stocks using U.S. data for is large, robust, and strongly related damental shocks in the context of short 1963–2003. Firms with higher lever- to the fact that volume surges around arbitrage in equity markets. The ability age, lower profitability, lower market announcement dates. Stocks with high of arbitrageurs to hold on to short posi- capitalization, lower past stock returns, past announcement period volume earn tions depends on asset values: shorts more volatile past stock returns, lower the highest announcement premium, are often reduced following good news cash holdings, higher market-book suggesting some common underlying about a stock. As a result, the prices of ratios, and lower prices per share are cause for both volume and the pre- highly shorted stocks are excessively more likely to file for bankruptcy, mium. The authors show that high pre- sensitive to economic shocks. Using be de-listed, or receive a D rating. mium stocks experience the highest monthly short interest data and exploit- When predicting failure at longer hori- levels of imputed small investor buying, ing differences in short selling regula- zons, the most persistent firm char- suggesting that the premium is driven tions across stock exchanges to instru- acteristics — market capitalization, the by buying by small investors when the ment for the amount of shorting in a market-book ratio, and equity vola- announcement catches their attention. stock, the authors find: 1) The price of tility — become relatively more signif- Psychological evidence indicates a highly shorted stock is more sensitive icant. The model here captures much that it is hard to process multiple stim- to earnings news than a stock with little of the time variation in the aggregate uli and perform multiple tasks at the short interest. 2) Short interest changes failure rate. Since 1981, financially dis- same time. Hirshleifer and his co- around announcements (proxied by tressed stocks have delivered anoma- authors test the investor distraction share turnover) are more sensitive to lously low returns. They have lower hypothesis, which holds that the arrival earnings surprises for highly shorted returns but much higher standard devi- of extraneous news causes trading and stocks. 3) For highly shorted stocks, ations, market betas, and loadings on market prices to react sluggishly to rel- returns to shorting are higher following value and small-cap risk factors than evant news about a firm. They focus on better earnings news. 4) These differen- stocks with a low risk of failure. These the competition for investor attention tial sensitivities are driven by very good patterns hold in all size quintiles but between a firm’s earnings announce- earnings news as opposed to very bad are particularly strong in smaller stocks. ments and the earnings announce- earnings news. These findings point They are inconsistent with the conjec- ments of other firms. They find that to the importance of limited arbitrage ture that the value and size effects are the immediate stock price and volume in affecting asset price dynamics and compensation for the risk of financial reaction to a firm’s earnings surprise is the potentially destabilizing role of distress. weaker, and post-earnings announce- speculators. On average, stock prices rise around ment drift is stronger, when a greater Campbell, Hilscher, and Szilagyi scheduled earnings announcement number of earnings announcements by explore the determinants of corporate dates. Frazzini and Lamont show that other firms are made on the same day. A

56 NBER Reporter Winter 2006/7 trading strategy that exploits post-earn- agers, exhibit trend-chasing behavior effect” from the target to the bidder ings announcement drift is most profit- in their technology stock investments. that induces the market to evaluate the able for earnings announcements made As a result, young managers increase assets of the more neglected bidder at on days with a lot of competing news, their technology holdings during the the (inflated) market value of the less but it is not profitable for announce- run-up, and decrease them during the neglected target. Both bidder and tar- ments made on days with little compet- downturn. The economic significance get premiums are positively related to ing news. of young managers’ actions is ampli- the difference in neglectedness between Asset market experiments suggest fied by large inflows into their funds bidder and target. However, the tar- that inexperienced investors play a role prior to the peak in technology stock get’s ability to appropriate the gain is in the formation of asset price bub- prices. These results are unlikely to be reduced by the fact that its bargaining bles. Without first-hand experience of explained by standard career concerns position is weaker when the potential a downturn, these investors are more models or by differences in the abil- for asset appreciation of the bidder is optimistic and likely to exhibit trend ity to pick technology stocks between higher. The effect on the value of the chasing in their portfolio decisions. young and old managers. bidder is persistent in the medium run Greenwood and Nagel examine this Massa and Zhang study the impact (1–2 years). The authors document a hypothesis with mutual fund manager of style investing on the market for cor- better medium-term performance of data from the technology bubble. Using porate control. By using data on the more neglected firms taking over less age as a proxy for managers’ investment flows in mutual funds, they construct a neglected ones. The bidder managers experience, they find that around the measure of “neglectedness” that is not a engaging in these types of “cosmetic peak of the bubble, mutual funds run direct transformation of stock market mergers” take advantage of the tempo- by younger managers are more heavily data, but directly relies on the identifi- rary window of opportunity created by invested in technology stocks, relative cation of the sentiment-induced inves- the higher stock price induced by the to their style benchmarks, than their tor demand. They show that bidders M and A deal to reduce their stake in older colleagues. Consistent with the tend to pair with targets that are rela- the firm at convenient conditions. experimental evidence, the authors find tively less neglected. The merger with that young managers, but not old man- a less neglected target generates a “halo

Labor Studies

The NBER’s Program on Labor Daniel Parent, McGill University, and Aggregate Implications of Studies met in Cambridge on November “Performance Pay and Wage Inequality” Assortative Matching: A Nonparametric 17. NBER Research Associates Lawrence Analysis” F. Katz and Richard B. Freeman, both of Sandra E. Black, University of Harvard University, organized the pro- California , Los Angeles and NBER, Carmit Segal, Harvard University, gram. These papers were discussed: and Alexandra Spitz‑Oener, Humboldt “Motivation, Test Scores, and Economic University Berlin, “Explaining Women’s Success” Joseph G. Altonji, Yale University Success: Technological Change and the and NBER, and Anthony A. Smith Skill Content of Women’s Work” Anne Case and Christina Paxson, and Ivan Vidangos, Yale University, Princeton University and NBER, “Modeling Earnings Dynamics” Bryan S. Graham, University of “Stature and Status: Height, Ability, California, Berkeley and NBER; Guido and Labor Market Outcomes”(NBER Thomas Lemieux, University of British W. Imbens, Harvard University and Working Paper No. 12466) Columbia and NBER; W. Bentley NBER; and Geert Ridder, University of Macleod, Columbia University; and Southern California, “Complementarity

Altonji, Smith, and Vidangos use hours over a career. Their model incor- error components in both wages and generalized indirect inference to esti- porates state and duration dependence hours, and measurement error. They esti- mate a joint model of earnings, employ- in several variables, multiple sources of mate the dynamic response of wage rates, ment, job changes, wage rates, and work unobserved heterogeneity, job-specific hours, and earnings to various shocks,

NBER Reporter Winter 2006/7 57 and measure the relative contributions of more skill-intensive occupations. In National Longitudinal Survey of Youth the shocks to the variance of earnings in addition, the task changes are occurring 1979 (NLSY), may serve as a proxy for a given year and over a lifetime. Shocks most rapidly in occupations in which test-taking motivation. To gather more associated with job changes make a large computers have made major headway. definite evidence on the motivational contribution to the variance of career Overall — and in contrast to recent liter- component in the coding speed test, earnings and operate mostly through the ature that puts a strong emphasis on only Segal conducts a controlled experiment, job-specific error components in wages one dimension of activities on the job, inducing motivation by the provision and hours. Unemployment shocks also namely interactive tasks — the research- of incentives. In the experiment, the make a large contribution and operate ers show that changes in job content have average performance improved substan- mostly through long-term effects on the evolved differently for men and women tially and significantly once incentives wage rate. on several dimensions. were provided. More importantly, there An increasing fraction of jobs in the Graham, Imbens, and Ridder pres- were heterogeneous responses to incen- U.S. labor market explicitly pay workers ent methods for evaluating the effects of tives. Roughly a third of the participants for their performance using a bonus, a reallocating an indivisible input across improved their performance significantly commission, or a piece rate. In this paper, production units. When production in response to performance-based incen- Lemieux, Macleod, and Parent look at technology is nonseparable such reallo- tives, while the others did not. These two the effect of the growing incidence of cations, although leaving the marginal groups have the same test score distribu- performance pay on wage inequality. The distribution of the reallocated input tions when incentives are provided, sug- basic premise of the paper is that perfor- unchanged by construction, may none- gesting that some participants are less mance pay jobs have a more “competi- theless alter average output. Examples motivated and invest less effort when no tive” pay structure that rewards produc- include reallocations of teachers across performance-based incentives are pro- tivity differences more than other jobs. classrooms composed of students of vided. These participants, however, are Consistent with this view, the authors varying mean ability and altering assign- not less able. How do the coding-speed show that compensation in performance ment mechanisms for college roommates test scores relate to economic success? pay jobs is more closely tied to both in the presence of social interactions. Focusing on male NLSY participants, measured (by the econometrician) and The researchers focus on the effects of Segal shows that the coding speed scores unmeasured productive characteristics of reallocating one input while holding the are highly correlated with earnings 23 workers. The authors conclude that the assignment of another, potentially com- years after NLSY participants took the growing incidence of performance pay plementary input, fixed. They present a test, even after controlling for usual mea- accounts for 25 percent of the growth in class of such reallocations — correlated sures of cognitive skills, like the Armed male wage inequality between the late matching rules — that includes the status Forces Qualification Test (AFQT) scores. 1970s and the early 1990s, and for most quo allocation, a random allocation, and Moreover, for highly educated workers, of the growth in top-end wage inequal- both the perfect positive and negative the association between AFQT scores ity (above the 80th percentile) during assortative matching allocations as spe- and earnings is significantly larger than this period. cial cases. Their econometric approach the one between coding speed scores and Black and Spitz‑Oener adopt a involves first nonparametrically estimat- earnings, while for less educated workers task-based view of technological change ing the production function and then these associations are of similar size. and examine how the proliferation of averaging this function over the distribu- It has long been recognized that computers in the 1980s and 1990s has tion of inputs induced by the new assign- taller adults hold jobs of higher status affected women’s tasks relative to those ment rule. and, on average, earn more than other of men. Using data from West Germany, Segal investigates how low-stakes workers. A large number of hypotheses they find that women have witnessed test scores relate to economic success. have been put forward to explain the large relative increases in non-routine The inferences in the economic litera- association between height and earnings. analytic tasks and non-routine interactive ture on this subject are mostly based In developed countries, researchers have tasks between 1979 and 1999. However, on tests, without performance-based emphasized factors such as self esteem, the most notable difference between the incentives, administered to survey par- social dominance, and discrimination. genders is the pronounced decline in ticipants. Segal argues that the lack of In this paper, Case and Paxson offer a routine task inputs among women with performance-based incentives allows for simpler explanation. Prenatal and early almost no change in routine task input the possibility that higher test scores childhood health and nutrition have for men. Consistent with the skill-biased are caused by noncognitive skills asso- critical effects on both growth and cog- technological change hypothesis, task ciated with test-taking motivation, and nitive development. As a result, on aver- changes were most pronounced within not necessarily by cognitive skills alone. age, taller people earn more because they occupations, whereas only minor parts This suggests that the coding speed are smarter. As early as age 3 — before of the aggregate trends are attributable test, which is a short and very sim- schooling has had a chance to play a to women who were moving towards ple test available for participants in the role — and throughout childhood, taller

58 NBER Reporter Winter 2006/7 children perform significantly better on to select into higher paying occupa- mium in adult earnings can be explained cognitive tests. The correlation between tions that require more advanced verbal by childhood scores on cognitive tests. height in childhood and adulthood is and numerical skills and greater intel- Furthermore, they show that taller adults approximately 0.7 for both men and ligence, for which they earn handsome select into occupations that have higher women, so that tall children are much returns. Using four datasets from the cognitive skill requirements and lower more likely to become tall adults. As United States and the United Kingdom, physical skill demands. adults, taller individuals are more likely the researchers find that the height pre-

Productivity

The NBER’s Program on Productivity Nick Bloom; Raffaela Sadun, London Sabien Dobbelaere, Ghent University, met in Cambridge on December 1, School of Economics; and John Van and Jacques Mairesse, CREST 2006. Nick Bloom and Kathryn L. Shaw, Reenen, London School of Economics and NBER, “Product Market and Stanford University and NBER, orga- and NBER, “It Ain’t What You Do But Labor Market Imperfections and nized the program. These papers were the Way that You Do IT:Investigating Heterogeneity in Panel Data Estimates discussed: the U.S. Productivity Miracle Using of the Production Function” Multinationals” Discussant: Chad Syversson, University Jan De Loecker, New York University, Discussant: Susanto Basu, Boston of Chicago and NBER “Product Differentiation, Multi‑Product College and NBER Firms and Structural Estimation of Bronwyn H. Hall, University of Productivity” Anne P. Bartel and Casey Ichinowski, California, Berkeley and NBER; Discussant: Marc Muendler, University Columbia University and NBER; Grid Thoma, University of Bocconi; of California, San Diego Kathryn L. Shaw; and Ricard and Salvatore Torrisi, University of Correa, Federal Reserve Board of Bologna, “The Market Value of Patents Francine Lafontaine and Jagadeesh Governors, “International Differences and R&D: Evidence from European Sivadasan, University of Michigan, “The in the Adoption and Impact of New Firms” Microeconomic Implications of Input Information Technologies and New HR Discussant: Megan MacGarvie, Boston Market Regulations:Cross‑Country Practices: The Valve-Making Industry in University and NBER Evidence from Within the Firm” the U.S. and the U.K.” Discussant: Lee Branstetter, Discussant: Scott Stern, Northwestern Carnegie‑Mellon University and NBER University and NBER

De Loecker proposes a methodol- where he has matched firm-level with a unique outlet-level dataset obtained ogy for estimating (total factor) pro- product-level information. The result- from a multi-national food chain oper- ductivity in an environment of prod- ing production coefficients and pro- ating about 2,840 retail outlets in over uct differentiation and multi-product ductivity estimates change considerably 48 countries outside the United States. firms. In addition to correcting for the after taking into account the demand The dataset provides information on simultaneity bias in the estimation of variation and the product mix. Finally, output, input costs, and labor costs production functions, he controls for he analyzes the effects of trade liber- at a weekly frequency over a four- the omitted price bias, as documented alization in the Belgian textile indus- year period, allowing the authors to by Klette and Griliches (1996). By try. While he finds significant produc- examine the consequences of increased aggregating demand and production tivity gains from trade liberalization, rigidity at a much more detailed level from product space into firm space, he the estimated effects are approximately than has been possible with commonly can use plant-level data to estimate pro- half of those obtained with standard available annual frequency or aggregate ductivity. The productivity estimates techniques. data. They find that higher levels of are corrected for demand shocks and, Lafontaine and Sivadasan investi- the index of labor market rigidity are as by-products, he recovers the elastic- gate the microeconomic implications of associated with significantly lower out- ity of demand and implied mark-ups. labor regulations that protect employ- put elasticity of labor demand, as well He applies this methodology to the ment and are expected to increase as significantly higher levels of hyster- Belgian textile industry, using a dataset rigidity in labor markets. They exploit esis (measured as the elasticity of cur-

NBER Reporter Winter 2006/7 59 rent labor costs with respect to the pre- strong in the sectors that intensively price-cost mark-up and extent of rent vious week’s). Specifically, an increase use information technologies (such as sharing parameters. For manufactur- of one standard deviation in the labor retail and wholesale): the very same ing as a whole, their preferred estimates regulation rigidity index 1) reduces the industries that accounted for the U.S.- of these parameters are of an order of response of labor cost to a one standard European productivity growth differ- magnitude of 1.3 and 0.5 respectively. deviation increase in output (revenue) ential since the mid-1990s. Their sector-level results indicate that by about 4.7 percentage points (from There is now a well-developed sector differences in these parameters, 27.2 percent to 22.5 percent); and 2) body of macroeconomic evidence that and in the underlying estimated factor increases the response of labor cost information technology (IT) invest- elasticities and shares, are quite sizeable. to a one standard deviation increase ments are likely to have ‘‘paid-off ’’ with Since firm production function, behav- in lagged labor cost by about 9.6 per- higher levels of productivity growth ior, and market environment are very centage points (from 17.8 percent to in industries that invested more heav- likely to vary even within sectors, they 27.4 percent). These estimates imply an ily in IT in recent years. In their earlier also investigate firm-level heterogeneity increase in gross misallocation of labor work, Bartel, Ichniowski, and Shaw in estimated mark-up and rent-sharing of about 2 to 5 percent for a single stan- provided evidence confirming that parameters. To determine the degree dard deviation increase in the index of valve manufacturing plants that adopt of true heterogeneity in these parame- labor regulation. Finally, they find that new IT are in fact the same ones that ters, they adopt the P.A. Swamy (1970) the Company delayed entry, operates increase the customization and reli- methodology, allowing them to correct fewer outlets, and favors franchising in ability of their products and increase the observed variance in the firm-level countries with more rigid labor laws. the speed and efficiency of their oper- estimates from their sampling variance. Overall, the data imply a strong impact ations, thereby providing an explana- The median of the firm estimates of of rigid labor laws on labor input and tion of what lies behind the macro- the price-cost mark-up, ignoring labor related decisions at the micro level. level trends. An important question is market imperfections, is 1.10, while as Productivity growth in sectors that whether plants outside of the United expected it is higher — 1.20 — when intensively use information technolo- States gain as much from IT as U.S. taking them into account. The median gies (IT) appears to have accelerated plants. In this paper with Correa, they of the corresponding firm estimates of much faster in the United States than add data from the United Kingdom to the extent of rent sharing is 0.62. The in Europe since 1995, leading to the their U.S. dataset. Based on this com- Swamy corresponding robust estimates U.S. “productivity miracle”. If this was bined data, they find that the plants of true dispersion are about 0.18, 0.37, partly attributable to the superior man- in the United Kingdom have experi- and 0.35, yielding very sizeable within- agement or organization of U.S. firms enced the same changes that are evi- sector firm heterogeneity. The authors (rather than simply the advantages of dent in the United States: pronounced find that firm size, capital intensity, being located in the United States geo- increases in productivity and increased distance to the sector technology fron- graphically), we would expect to see skill demand associated with increases tier, and investing in R and D seem to a stronger association of productiv- in the purchase of new capital that has account for a significant part of this ity with IT for U.S. multinationals IT imbedded in the capital. heterogeneity. (compared to non-US multination- Embedding the efficient bargain- Hall and her co-authors provide als) located in Europe. Bloom and his ing model into the original R. E. Hall novel empirical evidence on the pri- co-authors examine a large panel of (1988) approach for estimating price- vate value of patents and R and D. U.K. establishments and provide evi- cost margins shows that imperfections They analyze an unbalanced sample of dence that U.S.-owned establishments in the product and labor markets gen- firms from five EU countries — France, do have a stronger relationship between erate a wedge between factor elastici- Germany, Switzerland, Sweden, and productivity and IT capital than either ties in the production function and the United Kingdom — in the period non-U.S. multinationals or domestic their corresponding shares in revenue. 1985–2005. They explore the relation- establishments. Indeed, the differen- Dobbelaere and Mairesse investigate ship between firm’s stock market value tial effect of IT appears to account for these two sources of discrepancies, both and patents, accounting for the “qual- almost all of the difference in total fac- at the sector level and the firm level, ity” of EPO patents. They find that tor productivity between U.S.-owned using an unbalanced panel of 10,646 Tobin’s q is positively and significantly and all other establishments. This find- French firms in 38 manufacturing sec- associated with R and D and patent ing holds in the cross section, when tors over the period 1978–2001. By stocks. In contrast to results for the including fixed effects, and even when estimating standard production func- United States, forward citations do not the authors examine a sample of estab- tions and comparing the estimated fac- add information beyond that in pat- lishments taken over by U.S. multina- tor elasticities for labor and materi- ents. However, the composite quality tionals. They find that the U.S. mul- als and their shares in revenue, they indicator based on backward citations, tinational effect on IT is particularly are able to derive estimates of average forward citations, and the number of

60 NBER Reporter Winter 2006/7 technical fields covered by the patent are of poor quality and that these pat- in contrast to results for the United is informative for value. Software pat- ents are applied for merely to build States. However, in Europe, such pat- ents account for a rising share of total portfolios rather than for protection of ents are highly concentrated, with 90 patents in the EPO. Moreover, some real inventions. Therefore, the authors percent of the software patents in this scholars of innovation and intellectual tested for the impact of software pat- sample held by just 15 of the firms. property rights argue that software and ents on the market value of the firm business-methods patents on average and did not find any significant effect,

International Trade and Investment The NBER’s International Trade and Dartmouth College and NBER; Tariff Reform” — Eric Edmonds and Investment Program met at the NBER’s Stephen J. Redding, London School of Nina Pavcnik, Dartmouth College office in California on December 1‑2, Economics; and Peter K. Schott, Yale and NBER, and Petia Topalova, 2006. Program Director Robert C. University and NBER International Monetary Fund Feenstra, University of California, Davis, organized the meeting. These papers were “Quality Pricing and Endogenous “Trade, Knowledge, and the Industrial discussed: Entry: A Model of Exchange Rate Revolution” — Kevin H. O’Rourke, Pass‑Through” — Raphael Auer, Swiss Trinity College, Dublin and NBER; “Trading Tasks: A Simple Theory of National Bank, and Thomas Chaney, Ahmed S. Rahman, United States Offshoring” — Gene M. Grossman and University of Chicago and NBER Naval Academy; and Alan M. Taylor, Esteban Rossi‑Hansberg, Princeton University of California, Davis and University and NBER “Explaining Import Variety and NBER Quality: The Role of the Income “Trade, Diffusion, and the Gains from Distribution” — Yo Chul Choi and “Buffalo Hunt: International Trade and Openness” — Andres Rodriguez‑Clare, Chong Xiang, Purdue University, and the Virtual Extinction of the North Pennsylvania State University and David Hummels, Purdue University American Bison” — M. Scott Taylor, NBER and NBER University of Calgary and NBER

“Multi‑Product Firms and Trade “Trade Adjustment and Human Capital Liberalization” — Andrew B. Bernard, Investments: Evidence from Indian

For centuries, most international shared gains for all domestic factors, in is that, compared to the model with- trade involved an exchange of com- contrast to the distributional conflict out diffusion, the gains from trade are plete goods. But, with recent improve- that typically results from reductions in smaller whereas the gains from open- ments in transportation and commu- the cost of trading goods. ness are much larger when diffusion is nications technology, it increasingly Rodriguez‑Clare presents and included in the model. entails different countries adding value then calibrates a model in which coun- Bernard, Redding, and Schott to global supply chains, or what might tries interact through trade and diffu- develop a general equilibrium model of be called “trade in tasks.” Grossman sion of ideas, and then quantifies the multi-product firms and analyze their and Rossi‑Hansberg propose a new overall gains from openness and the behavior during trade liberalization. conceptualization of the global pro- contribution of trade to these gains. Firm productivity in a given product duction process that focuses on trad- Having the model match the trade data is modeled as a combination of firm- able tasks and use it to study how (that is, the gravity equation) and the level “ability” and firm-product-level falling costs of offshoring affect fac- observed growth rate is critical for this “expertise”, both of which are stochastic tor prices in the source country. The quantification to be reasonable. It is and unknown prior to the firm’s pay- authors identify a productivity effect of shown that for this match it is necessary ment of a sunk cost of entry. Higher task trade that benefits the factor whose to introduce diffusion and/or knowl- managerial ability raises a firm’s pro- tasks are more easily moved offshore. edge spillovers to the basic model of ductivity across all products, which In the light of this effect, reductions in trade and growth in Eaton and Kortum induces a positive correlation between the cost of trading tasks can generate (2001). The main result of the paper a firm’s intensive (output per product)

NBER Reporter Winter 2006/7 61 and extensive (number of products) ine a generalized version of Flam and address a basic empirical problem fac- margins. Trade liberalization fosters Helpman’s (1987) model of vertical ing previous unified growth models, productivity growth within and across differentiation that maps cross-coun- exemplified by Galor and Weil (2000). firms and in the aggregate by induc- try differences in income distribu- In such models, the onset of industri- ing firms to shed marginally productive tions to variations in import variety alization leads to an increase in skill products and forcing the lowest-pro- and price distributions. The theoreti- premia, which is required in order to ductivity firms to exit. Though export- cal predictions are examined and con- induce families to limit fertility and ers produce a smaler range of products firmed using micro data on income increase the education of their children. after liberalization, they increase the from the Luxemburg Income Study for However, the onset of the Industrial share of products sold abroad as well as 30 countries over 20 years. The pairs of Revolution saw a marked decline in skill exports per product. All of these adjust- importers whose income distributions premia, and this cannot be explained by ments are shown to be relatively more look more similar have more export supply-side educational reforms, since pronounced in countries’ comparative partners in common and more similar these only came much later. Thus the advantage industries. import price distributions. Similarly, authors construct a model, in the tradi- Auer and Chaney develop a model the importers whose income distribu- tion of Galor and Weil and Galor and of international trade under perfect tions look more like the world buy Mountford (2004), which endogenizes competition and flexible prices that from more exporters and have import the direction of technical change. They accounts for the slow and incomplete price distributions that look more like show that technological change dur- pass through of exchange rate fluc- the world. ing the early phases of the Industrial tuations into consumer prices. They Can the short- and medium-term Revolution was inevitably unskilled- build an extension of the Mussa and adjustment costs associated with trade labor-biased. They also show that a Rosen (1978) model of quality pricing. liberalization have long-term con- growth in “Baconian knowledge” and Exporters sell goods of different qual- sequences through their impact on international trade can explain a shift in ity to consumers with heterogeneous schooling and child labor? Edmonds, the direction of technical change away preferences for quality. In equilibrium, Pavcnik, and Topalova examine this from unskilled-labor-intensive innova- higher quality goods are more expen- question in the context of India’s 1991 tions and towards skilled-labor-inten- sive. The authors derive three testable tariff reforms. Overall, in the 1990s, sive innovations. Simulations show that predictions. First, exchange rate fluctu- rural India experienced a dramatic the model does a good job in track- ations are only partially passed through increase in schooling and a decline in ing reality, at least until the late nine- to consumers. Second, there is more child labor. However, communities that teenth century with its mass education pass through in the long run than in relied heavily on employment in pro- reforms. the short run, and more pass through tected industries before liberalization In the sixteenth century, North for aggregate prices than for individ- do not experience as large an increase America contained 25–30 million buf- ual prices. Third, there is more pass in schooling or decline in child labor. falo; by the late nineteenth century through for low quality goods than for The data suggest that this failure to fewer than 100 remained. While remov- high quality goods. When the exchange follow the national trend of increas- ing the buffalo east of the Mississippi rate of an exporting country appre- ing schooling and diminishing work took settlers two centuries, the remain- ciates, existing exporters scale down is associated with a failure to follow ing 10 to 15 million buffalo on the their production, driving prices up. In the national trend in reduc- Great Plains were killed in a punctu- the long run, low quality exporters tion. Schooling costs appear to play a ated slaughter in a little over ten years. pull out, driving prices up even fur- large role in this relationship between Taylor uses theory, data from interna- ther. Since those goods are inexpen- poverty, schooling, and child labor. tional trade statistics, and first person sive, aggregate prices go up more than Extrapolating from these results, the accounts to argue that the slaughter individual prices. This exit of low qual- estimates here imply that roughly half on the plains was initiated by a for- ity exporters has a larger impact on the of India’s rise in schooling and a third eign-made innovation and fueled by a price of low quality goods than on the of the fall in child labor during the foreign demand for industrial leather. price of high quality goods. Low qual- 1990s can be explained by falling pov- Ironically, the ultimate cause of this ity goods prices adjust more than high erty and therefore improved capacity to sad chapter in American environmen- quality goods prices. afford schooling. tal history was of European, and not Choi, Xiang, and Hummels exam- O’Rourke, Rahman, and Taylor American, origin.

62 NBER Reporter Winter 2006/7 Bureau Books

The following volumes may be ordered directly from the University of Chicago Press, Order Department,11 030 South Langley Avenue, Chicago, IL 60628-2215; 1-800-621-2736. Academic discounts of 10 percent for individual volumes and 20 percent for standing orders for all NBER books published by the University of Chicago Press are available to university faculty; orders must be sent on university stationery.

Globalization and Poverty

Globalization and Poverty, edited the lives of the poor? Has increased finan- tive arguments. This volume provides the by Ann Harrison, is available from the cial integration led to more or less pov- more nuanced understanding necessary to University of Chicago Press for $99.00. erty? How have the poor fared during var- move that debate beyond the slogans. This NBER Conference Report brings ious currency crises? Does food aid hurt Harrison is a Research Associate in together experts on international trade or help the poor? the NBER’s Program on International and poverty to provide a detailed view of To date, poverty has been used as Trade and Investment and a Professor of the effects of globalization on the poor in a popular and convenient catch phrase Agricultural and Resource Economics at developing nations, answering such ques- by parties on both sides of the global- the University of California, Berkeley. tions as: Do lower import tariffs improve ization debate to further their respec-

The Risks of Financial Institutions The Risks of Financial Institutions, measure and forecast risk. The contribu- the Federal Reserve Board of Governors edited by Mark Carey and René M. Stulz, tors — academics, regulators, and bank- in Washington. Stulz is an NBER is available from the University of Chicago ers — bring a wide range of perspectives Research Associate in the Programs on Press for $99.00. This NBER Conference and experience to the issue. The result is Asset Pricing and Corporate Finance and Report examines the various risks affect- a volume that points a way toward greater the Everett D. Reese Professor of Banking ing financial institutions and explores a financial stability and better risk manage- and Monetary Economics at The Ohio variety of methods for helping institu- ment of financial institutions. State University. tions and regulators to more accurately Carey is a Finance Project Manager at

NBER Reporter Winter 2006/7 63 Nonprofit Org. U.S. Postage NBERReporter PAID National Bureau of NATIONAL BUREAU OF ECONOMIC RESEARCH Economic Research

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

Change Service Requested