NBER Reporter NATIONAL BUREAU OF ECONOMIC RESEARCH

A quarterly summary of NBER research No. 1, March 2019

Program Report ALSO IN THIS ISSUE

Misconduct and Employment of Financial Advisers Economics of Education

Advisers with history Advisers with history of no misconduct of misconduct 100% 100%

45% Caroline Hoxby * 54%

81% 81%

22% 37%

10% 9% 24% 18% 0% 9% 10% 0% Men Women Men Women The enterprise of education has a supply side, where institu-

Join dierent firm Leave the industry Remain with the firm tions produce education, and an investment side, where people

Source: Researchers’ calculations using data from the Financial Industry Regulatory Authority acquire education. I say an “investment side” and not a “demand side” because education is largely an investment and not a form of consumption. Because it is an investment, the economics of educa- The Labor Market for Financial Misconduct 9 tion is guided by a great deal of theory that would not apply if edu- The Return of Survey Expectations 14 cation were a consumption good like, say, bread. Education econo- mists study both the supply of and investment in education. What Inventions Are We Missing? 18 On the supply side, we think about institutions’ objectives and Do Government Grants to Charities constraints. We consider competition among institutions and how Crowd Private Donations Out or In? 22 institutions interact with governments and taxpayers. We use mod- NBER News 26 els from public economics, political economy, industrial organiza- tion, regulation, and finance. Conferences 28 On the investment side, we consider whether people are under- Program and Working Group Meetings 33 or over-investing in their own or others’ education — a determina- NBER Books 42 tion primarily based on whether they are earning a higher or lower rate of return than what they could earn in an alternative invest- ment, such as physical capital. We consider market failures because the financing of human capital investments is failure-prone, owing to issues like moral hazard that occur when human beings them- selves are the vehicles for investment. We often investigate the potential for market failures due to people being poorly informed or irrational about investing in themselves — mispredicting their returns, say, or discounting the future in a hyperbolic manner. Applying economic analysis to education is the defining fea- ture of the NBER’s Economics of Education Program. As an NBER program, there are also some distinctive features. First, the research

* Caroline Hoxby is the Scott and Donya Bommer Professor in Economics at Stanford University, an NBER research associate, and director of the Economics of Education Program. She is also a senior fellow of the Hoover Institution and of the Stanford Institute for Economic Policy Research.

Reporter Online at: www.nber.org/reporter carried out by education economists now relies, to an unusual extent, on extremely NBER Reporter high-quality administrative data recorded by schools, governments, and authoritative third parties. The data are so accurate and comprehensive that we often can use ambi- The National Bureau of Economic Research is a private, nonprofit research orga- nization founded in 1920 and devoted to objective quantitative analysis of the tious econometric tools that are impracti- American economy. Its officers and board of directors are: cal with data that are sample-based, sparse, President and Chief Executive Officer — James M. Poterba or prone to error. Second, the program fea- Controller — Kelly Horak tures young scholars to an unusual extent, Corporate Secretary — Alterra Milone because in most years the share of education BOARD OF DIRECTORS economists in the cohort of new PhDs is Chair — Karen N. Horn greater than the share in the previous year. Vice Chair — John Lipsky These emerging scholars are highly produc- Treasurer — Robert Mednick tive, and they keep the program in a con- DIRECTORS AT LARGE stant state of rejuvenation and intellectual Peter Aldrich Diana Farrell Karen Mills excitement. Third, the program is unusually Elizabeth E. Bailey Jacob A. Frenkel Michael H. Moskow John H. Biggs Robert S. Hamada Alicia H. Munnell diverse and inclusive because education is so John S. Clarkeson Peter Blair Henry Robert T. Parry interesting to so many scholars. The partici- Kathleen B. Cooper Karen N. Horn James M. Poterba pants in program meetings represent a wide Charles H. Dallara Lisa Jordan John S. Reed George C. Eads John Lipsky Marina v. N. Whitman array of institutions, demographic back- Jessica P. Einhorn Laurence H. Meyer Martin B. Zimmerman grounds, national origins, and policy views. Mohamed El-Erian This energy and diversity make writing a report like this one a challenge: I cannot DIRECTORS BY UNIVERSITY APPOINTMENT possibly do justice to all of the research. So, Timothy Bresnahan, Stanford Samuel Kortum, Yale Pierre-André Chiappori, Columbia George Mailath, Pennsylvania in this report, I emphasize a few key topics Alan V. Deardorff, Michigan Marjorie B. McElroy, Duke that have received considerable attention in Edward Foster, Minnesota Joel Mokyr, Northwestern the past few years. In my conclusion, I dis- John P. Gould, Chicago Cecilia Elena Rouse, Princeton Mark Grinblatt, California, Los Angeles Richard L. Schmalensee, MIT cuss some up-and-coming topics as well. Bruce Hansen, Wisconsin-Madison Ing o Wa l t er, New York Benjamin Hermalin, California, Berkeley David B. Yoffie, Harvard Productivity in Higher Education DIRECTORS BY APPOINTMENT OF OTHER ORGANIZATIONS Jean-Paul Chavas, Agricultural and Applied Economics Association Our most important project, since my Martin Gruber, American Finance Association last report, is our initiative to analyze pro- Philip Hoffman, Economic History Association ductivity in higher education. Institutions Arthur Kennickell, American Statistical Association Jack Kleinhenz, National Association for Business Economics of higher education — from large elite Robert Mednick, American Institute of Certified Public Accountants research universities to small private colleges Peter L. Rousseau, American Economic Association and for-profit institutions — have never Gregor W. Smith, Canadian Economics Association William Spriggs, American Federation of Labor and been under greater scrutiny. Policymakers, Congress of Industrial Organizations families, philanthropists, and the media Bart van Ark, The Conference Board question whether the benefits of higher The NBER depends on funding from individuals, corporations, and private education justify the costs. These questions foundations to maintain its independence and its flexibility in choosing its are fundamentally about the productivity research activities. Inquiries concerning contributions may be addressed to James of the sector. M. Poterba, President & CEO, NBER, 1050 Massachusetts Avenue, Cambridge, MA 02138-5398. All contributions to the NBER are tax-deductible. To answer these questions, the NBER, with support from the Alfred P. Sloan The Reporter is issued for informational purposes and has not been reviewed by Foundation and the Spencer Foundation, the Board of Directors of the NBER. It is not copyrighted and can be freely repro- commissioned nine papers to be the focus duced with appropriate attribution of source. Please provide the NBER’s Public Information Department with copies of anything reproduced. of a conference that brought together researchers, university leaders, policymak- Requests for subscriptions, changes of address, and cancellations should be sent ers, and journalists. The papers are available to Reporter, NBER, 1050 Massachusetts Avenue, Cambridge, MA 02138-5398 (please include the current mailing label), or by email to [email protected]. Print as NBER working papers and as chapters copies of the Reporter are only mailed to subscribers in the U.S. and Canada; in a forthcoming volume, Productivity in those in other nations may request electronic subscriptions at www.nber.org/ Higher Education. The studies use rich and drsubscribe/.

2 NBER Reporter • No. 1, March 2019 novel administrative data, employ States.1 While the study emphasizes students are poorly informed investors cogent economic reasoning, deploy productivity results based on life- or rely so greatly on third parties to pay the latest econometric methods, and time earnings because these matter their tuition. evince deep institutional understand- disproportionately for the financial Evan Riehl, Juan Saavedra, and ing. In combination, the papers are stability of the postsecondary sector, Miguel Urquiola draw upon data from fairly comprehensive: They include it also shows results based on public Colombia, a country with a vigorous studies of the returns to undergrad- service and innovation. The study’s market for higher education that is uate education, how costs differ by most important advance, however, is not dissimilar to that of the United major, the productivity of for-profit addressing the aforementioned selec- States.2 Importantly, Colombian stu- schools, the productivity of various tion problem by employing all of the dents’ learning is assessed by examina- types of instructors, and how online possible quasi-experiments in which a tions not only before they enter uni- education has affected the market. student “flips a coin” between schools versities but also when they exit. The Analyses of productivity in that have nearly identical selectiv- researchers demonstrate that college higher education productivity based must confront sig- on learning is quite nificant challenges. Educational Productivity By Institutional Selectivity different from pro- Higher education ductivity based on affects many out- Average productivity of a dollar earnings, especially of core educational spending comes: students’ 8 initial earnings. In skills, employment, particular, learning- innovativeness, and 7 based measures are public service, to 6 much more corre- name a few. Higher 5 lated with long-term education insti- 4 earnings than they tutions conduct a 3 are with initial earn- bewildering array 2 ings. This suggests of activities across 1 that learning reflects many domains, 0 long-term value- from undergraduate 700 800 900 1000 1100 1200 1300 1400 1500 added, while initial teaching to medi- Math + Verbal SAT score of institution's average student earnings more heav- cal research. Even if ily reflect skills that we focus on a single Source: “The Productivity of U.S. Postsecondary Institutions” by Caroline Hoxby, chapter in forthcoming NBER book from depreciate quickly or outcome — the earn- University of Chicago Press, “Productivity in Higher Education”, edited by Caroline M. Hoxby and Kevin Stange students’ pre-college ings-based returns characteristics. to undergradu- Figure 1 Joseph Altonji ate education, for and Seth Zimmerman instance — assessing the contribution ity or in which admission staff “flip analyze whether productivity dif- of an individual institution must over- a coin” between students with nearly fers by college major.3 There are at come the fact that students select into identical achievement. Thus, the least three reasons why such analy- schools based on their aptitude and study exemplifies how having com- ses are hard. First, there is substantial often attend more than one school. prehensive data allows one to pursue selection into majors: Students with Finally, some benefits of higher edu- ambitious econometric strategies. higher aptitude tend to major in cer- cation are inherently public in nature The study’s most important find- tain fields. Second, the relationship and difficult to measure or attribute ing, illustrated in Figure 1, is that when between initial earnings and lifetime to any one institution. This project’s earnings are used to measure benefits, earnings varies by major. Engineering studies demonstrate that these five the productivity of a dollar is fairly majors, for instance, have high ini- challenges — multiple outcomes, the similar across a wide array of selec- tial earnings but subsequently experi- multi-product nature of institutions, tive postsecondary institutions. This ence unusually slow earnings growth. selection, attribution, and public ben- result suggests that market forces com- Third, different majors cost different efits — are surmountable. pel some amount of efficiency among amounts to produce. For example, I attempt to com- selective institutions. However, I also Using administrative data for all pute the productivity of the vast find that market forces appear to exert Florida public institutions, Altonji majority of undergraduate programs little productivity discipline on non- and Zimmerman show that majors (more than 6,000) in the United selective schools, possibly because their that are intensive in equipment,

NBER Reporter • No. 1, March 2019 3 space, or highly paid faculty are dra- This brings us to a key take- nomics comes in: Do economic logic matically more costly on a per-stu- away from Productivity in Higher and evidence suggest that we could have dent basis. If we consider costs, the Education: Higher education institu- better teachers? This question is one on productivity findings are very dif- tions do respond to market forces, but which research in the last few years has ferent than what we might conclude institutions’ investors (the students), made substantial progress. from a naive look at initial earnings. constraints, production functions, Barbara Biasi investigates how Strikingly, as Figure 2 demonstrates, and revenue sources vary. Thus, the teachers responded when a change in the ratio of earnings to costs is simi- extent to which market pressure dis- Wisconsin law allowed school districts lar in majors with high earnings and ciplines productivity differs greatly, to pay teachers in a flexible way rather high costs, like engi- than in the rigid, neering, and modest Earnings Relative to Instructional Cost by Major “lockstep” manner earnings and mod- based almost entirely est costs, like public Ratio of the present discount value on seniority that of earnings to cost of instruction administration. 1 characterizes nearly Pieter De all U.S. schools.5 She Library Science Vlieger, Brian Jacob, Business finds that districts Computer Science and Kevin Stange 0.5 that adopted flex- Mass Communication Law Enforcement estimate instruc- Engineering ible pay ended up Social Science Law Psychology tor productivity in English with teachers whose Philosophy Home Ec. Public Admin. standardized courses 0 Gen. Studies Health Sciences value-added was Language Math Fitness Nat. Resources Education at the University Life Sciences higher. Part of the Art 4 Physical Sciences of Phoenix. Agri. Business Architecture improvement arose

Employing data on -0.5 because young, high

more than 300,000 -0.4 -0.2 0 0.2 0.4 value-added teachers, students and 2,000 Logarithm of present who were systemi- instructors, they discount value of earnings cally underpaid given make use of the Values are relative to education majors their productivity, fact that the assign- Source: J. G. Altonji and S. G. Zimmerman, NBER Working Paper No. 23029 left the rigid-pay dis- ment of students to tricts and joined flex- teachers is virtually Figure 2 ible-pay ones where random. They show they could be paid that instructors’ productivity var- with selective institutions probably more in proportion to their contribu- ies greatly and, interestingly, varies being subject to much more disci- tions. The evidence suggests that part much more in person than in online pline than non-selective ones. of the improvement came through courses. Indeed, if students want to teachers improving their effectiveness obtain instruction that has maximum The Quality of Teachers in flexible-pay districts, presumably value-added, they must get it in per- Is Not Fixed because effective teaching could be son, because the online experience rewarded there. suppresses variation in instructional One of the main results of recent Hugh Macartney, Robert value-added. studies using large administrative data- McMillan, and Uros Petronijevic use The study’s most surprising sets on educational inputs and out- data from North Carolina to dem- result, though, is that the University puts in K-12 education is a persuasive onstrate that teachers improve their of Phoenix, despite being a for-profit body of evidence that, as long assumed, value-added when accountability school, pays its highly variant instruc- teachers matter. A student who is for- incentives are strengthened.6 They tors exactly the same amount. This tunate enough to have a series of effec- use rich longitudinal data to separate finding suggests that the sort of stu- tive teachers can end up with substan- the improvements into two parts: the dents who consider non-selective tially better outcomes, not just in terms part that arises because existing teach- for-profit institutions do not make of academic achievement but also in ers raise their effort, and the part that their enrollment choices based on terms of college attainment, later earn- arises because higher-ability people the schools’ record of skill produc- ings, and a variety of social outcomes. join teaching or decide not to leave tion. If that is the case, the University But the knowledge that teachers matter it. The researchers use their estimates of Phoenix is probably not forgoing could be frustrating if there were no pol- to compare the cost-effectiveness of profit by paying all instructors the icies by which we could improve the set alternative education policies and same amount. of people who teach. This is where eco- show that incentive-oriented reforms

4 NBER Reporter • No. 1, March 2019 can outperform policies that only tar- Crucially, he conducts these experi- tives they face. For example, as Figure get the recruitment of higher-ability ments in a district that is actually 3 illustrates, when the unemployment teachers. This is essentially because reconfiguring its entire system of rate for recent college graduates rises, incentive-oriented policies improve teacher compensation. Thus, nearly the quality of teachers, measured by all teachers, including the “stayers,” all teachers participated in the experi- their value added for students, rises. and effectively recruit better teachers. ments, and they had strong incentives Fascinatingly, much that can be Sally Hudson attempts to answer to answer honestly in order to get the said of U.S. education can also be the question most often asked about system they preferred. One important said of other countries that might Teach for America (TFA).7 Even if result is that teachers with high value- be thought to be very different. For TFA teachers, who come from the added prefer merit pay more than instance, Natalie Bau and Jishnu Das nation’s most selective colleges, are those with low value-added. This sug- showed that Pakistani teachers’ value- really much more able than the average gests that, by switching to pay that is added varies about as much as it does incumbent teacher, is it worthwhile more merit-based, a district can dis- in highly industrialized countries, to hire them? After all, they are inex- proportionately pull in higher value- and is, as in those countries, uncor- perienced, and evi- related with teach- dence is strong that ers’ credentials.11 As instructors improve Change in the Unemployment Rate and Teacher E ectiveness in other countries, in their first couple teachers in Pakistan of years. Moreover, Change in experience adjusted improve in their first teacher value-added in mathematics TFA teachers need +0.2 couple of years, but to be replaced every not much after that.

two to three years +0.1 There is no relation- because the vast ship between teacher majority go on to pay and performance 0 careers outside the in Pakistan’s public classrooms. Hudson schools, where com- shows that, in fact, −0.1 pensation is based hard-to-staff schools almost entirely on appear to benefit −0.2 seniority, but there from a succession of −1 −0.5 0 0.5 1 is a meaningful pos- One−year change in unemployment among TFA teachers. This bachelor’s degree holders (percentage points) itive relationship is largely because in private schools,

the TFA teachers’ Source: M. Nagler, M. Piopiunik, and M. R. West, NBER Working Paper No. 21393 where average pay is effectiveness is so lower. These find- much greater, even Figure 3 ings apply in many when they are nov- countries, rich and ices, than that of non- poor. The similarities TFA teachers. Also, the TFA teach- added recruits. Another interesting are so striking that they must reveal ers improve their effectiveness faster result is that, while teachers do need something about first, the underlying than do non-TFA teachers. to be paid more to teach students who production function for instruction, Andrew C. Johnston conducts a are low achievers, all else being equal, and second, the political economy of novel experiment, asking teachers to they do not need to be paid more to public education. choose between pay that is more or teach students from low-income or Bau and Das’ most surprising less merit-based, between systems of racial/ethnic minority backgrounds. finding is that, in Pakistan, compen- merit evaluation, and between com- Two other papers that demon- sation was so out of line with effec- pensation that is salary-focused or strate that greater relative compensa- tiveness that public schools’ actually benefits-focused.8 Much teacher tion allows schools to recruit more recruited better teachers after a pol- compensation is currently in the form effective teachers are by Markus icy change that put teachers on tem- of unusually generous retirement and Nagler, Marc Piopiunik, Martin R. porary contracts — jobs more suscep- health benefits. Johnston asks teach- West, and Owen Thompson.9,10 tible to performance review — even ers how they would trade off better In short, the quality of teachers in though that same policy cut their pay versus students who were easier the United States is not fixed, but average salaries by 35 percent. to teach, students who were lower- depends on how they are recruited This phenomenon — pub- income, longer commutes, and so on. and the compensation-based incen- lic school teachers’ pay being dra-

NBER Reporter • No. 1, March 2019 5 matically out of line with alterna- that for-profit institutions play in the focus on capturing federal govern- tive jobs and with effectiveness — is non-repayment of student loans.14 ment aid at the expense of student fairly common in developing coun- These schools’ students are very dis- outcomes. It is worth noting that fed- tries. In Colombia, for instance, proportionately responsible for non- eral loans, federal aid, and veterans’ Saavedra, Dario Maldonado, Lucre­cia repayment because they take on unusu- GI Bill-based aid often make up close Santibanez, and Luis Omar Herrera ally great student debt and experience to 100 percent of the revenue at for- Prada show that people across all unusually little improvement in earn- profit schools. In other words, a stu- ability levels earn a substantial pre- ings. The researchers ask whether these dent paying tuition from his or her mium if they teach in public schools patterns are causal. In other words, if own pocket is a rare exception, not the rather than take alternative jobs.12 the same students were to attend, say, rule, at such schools. The researchers The researchers demonstrate this public community colleges, would they find that, when their incentives to convincingly by comparing people end up with equal payment problems? maximize profits are intensified, for- who barely pass and barely fail the By comparing enrollment and postsec- profit schools enroll more students, national teacher-screening exam. ondary outcome changes across areas enroll students who are less likely Finally, Isaac Mbiti, Karthik that experience similar labor demand to benefit from higher education, Muralidharan, Mauricio Romero, shocks but have different supplies of increase tuition, and increase student Youdi Schipper, loans. The results Constantine Manda, are worse student and Rakesh Rajani Outcomes at Public and For-Profit Colleges outcomes in terms Students at for-profit colleges are more likely to obtain a federal loan, less likely to be employed 6 years a er find the intriguing employment, and more than three times as likely to default on their student loans as students at public colleges. of graduation rates, result that teacher pay employment, and incentives are more Probability of obtaining Probability of employment Probability of defaulting earnings, and sig- effective when they a Federally subsidized loan 6 years aer enrollment on student loans nificantly lower 100% 100% 100% are combined with 95% repayment rates. additional resources, 88% Other evi- at least in Tanzania.13 77% dence of for- Their findings are profit institutions’ based on an ambi- eagerness to cap- tious randomized con- 35% 16% ture government 5% trolled trial involv- 0% 0% 0% aid comes from ing 350 schools where Matthew Baird, unconditional grants, For-profit college Public college Mike Kofoed, teacher incentives, or Trey Miller, and a combination of both Source: L. Armona, R. Chakrabarti, and M. F. Lovenheim, NBER Working Paper No. 25042 Jennie Wenger.16 are implemented. They show that Figure 4 for-profit schools Student Loans quickly raised for-profit institutions, they are able to tuition to absorb the increases in Student loans have risen greatly show that much of the effect is indeed maximum tuition allowed under the in volume in recent years and are now causal. As Figure 4 shows, enrollment new Forever GI Bill. Given the high by far the largest source of unsecured in for-profits leads to greater loans, tendency of veterans to attend for- debt in the United States. Moreover, increased non-repayment, and worse profits, the bill thus improved profits some students are unlikely to repay, labor market outcomes. but did little for veterans. Arguably, so these loans generate risks for the Why might enrolling at a for- this bill made matters worse for entire economy in a manner not profit cause greater debt and non- non-veterans who attend for-profit unlike the risks generated by mort- repayment? Charlie Eaton, Sabrina schools, since they too were faced gages in the recent financial crisis. Howell, and Constantine Yannelis with somewhat higher tuition and, Fortunately, many economists of edu- answer this question.15 Studying what consequently, greater loans. cation associated with the NBER are happens when a for-profit college is What would happen if for-profit helping everyone to gain a better subject to stronger incentives to max- colleges were to lose some of their understanding of student loans. imize profits as the result of a private access to federal loans and other fed- For instance, Luis Armona, Rajashri equity buyout, they find that institu- eral financial aid? Stephanie Cellini, Chakrabarti, and Michael Lovenheim tions subject to high-powered profit- Rajeev Darolia, and Lesley Turner focus on the extremely important role maximizing incentives intensify their examine what happened when, in

6 NBER Reporter • No. 1, March 2019 the 1990s, students at more than from India looks promising. Return to Text 1,200 for-profit institutions faced There is new evidence on the 7 S. Hudson, “The Dynamic Effects restricted access to loans because returns to college majors, calling of Teach for America in Hard-to-Staff loan default rates were so high at into question the common impres- Schools,” NBER Summer Institute, the schools.17 Using variation in the sion that the greatest returns are to 2017. timing and restrictiveness of sanc- becoming an engineer. Supported by Return to Text tions, the researchers find that low- rich data and the econometric ambi- 8 A. Johnston, “Teacher Utility, income students were less likely to tiousness of program members, many Separating Equilibria, and Optimal attend for-profit schools but were so program meetings now include pre- Compensation: Evidence from a much more likely to attend public sentations on applied econometric Discrete-Choice Experiment,” Fall community colleges that the effects methods. Indeed, we now host meth- 2018 Economics of Education on their enrollment were about a ods symposia. Increasingly, advances Program. wash. The effects on loan repayment from behavioral economics, brain Return to Text were positive, however, because the science, and psychology are finding 9 M. Nagler, M. Piopiunik, and M. students who went to community their way into papers in the econom- West, “Weak Markets, Strong Teachers: colleges acquired less debt and were ics of education. Recession at Career Start and Teacher more likely to repay the smaller loans Effectiveness,” NBER Working Paper they took on. No. 21393, July 2015. 1 C. Hoxby, “Estimating the Return to Text Other Exciting Developments Productivity of U.S. Postsecondary 10 O. Thompson, “School Desegregation Institutions,” July 2018, chapter in and Black Teacher Employment,” The program is proud to report forthcoming book, C. Hoxby and K. NBER Summer Institute, 2018. that a long-time member, Parag Stange, eds., Productivity in Higher Return to Text Pathak, won the American Economic Education, University of Chicago Press. 11 N. Bau and J. Das, “The Association’s John Bates Clark Medal Return to Text Misallocation of Pay and Productivity in 2018. The award citation recog- 2 E. Riehl, J. Saavedra, and M. in the Public Sector: Evidence from the nized his work using market design Urquiola, “Learning and Earning: Labor Market for Teachers,” NBER theory to analyze systems in which An Approximation to College Value Spring 2017 Program. students are matched to schools. Such Added in Two Dimensions,” chapter Return to Text systems are used in numerous cities. in forthcoming book, C. Hoxby and K. 12 J. Saavedra, D. Maldonado, L. In a study of Taiwan’s matching sys- Stange, eds., Productivity in Higher Santibanez, and L. Herrera Prada, tem, for example, Umut M. Dur, Fei Education, University of Chicago Press. “Premium or Penalty? Labor Market Song, Pathak, and Tayfun Sönmez Return to Text Returns to Novice Public Sector found that school assignment mech- 3 J. Altonji and S. Zimmerman, Teachers,” NBER Working Paper No. anisms which include deduction sys- “The Costs and Net Returns to College 24012, November 2017. tems are manipulable, meaning that Major,” NBER Working Paper No. Return to Text children from families which are stra- 23029, January 2017. 13 I. Mbiti, K. Muralidharan, M. tegic are more likely to receive desir- Return to Text Romero, Y. Schipper, C. Manda, and able placements.18 4 P. De Vlieger, B. Jacob, and K. R. Rajani, “Inputs, Incentives, and As mentioned at the outset, there Stange, “Measuring Instructor Complementarities in Education: is much more exciting research associ- Effectiveness in Higher Education,” Experimental Evidence from ated with the Economics of Education NBER Working Paper No. 22998, Tanzania,” NBER Working Paper No. Program than I can possibly describe December 2016. 24876, July 2018. here. To induce you to explore fur- Return to Text Return to Text ther, let me mention just a few topics 5 B. Biasi, “The Labor Market for 14 L. Armona, R. Chakrabarti, and that are “up and coming.” A number of Teachers under Different Pay Schemes,” M. Lovenheim, “How Does For-Profit recent NBER working papers and con- NBER Working Paper No. 24813, College Attendance Affect Student ference papers evaluate online educa- July 2018. Loans, Defaults, and Labor Market tion, both at the K-12 and college level. Return to Text Outcomes?” NBER Working Paper The use of technology in education has 6 H. Macartney, R. McMillan, and No. 25042, September 2018. also been the subject of recent stud- U. Petronijevic, “Teacher Performance Return to Text ies which explore both developed and and Accountability Incentives,” NBER 15 C. Eaton, S. Howell, and C. developing countries. Some evidence Working Paper No. 24747, June 2018. Yannelis, “When Investor Incentives

NBER Reporter • No. 1, March 2019 7 and Consumer Interests Diverge: Shocks: An Investigation of Changes in December 2016. Private Equity in Higher Education,” Post-9/11 GI Bill Allowed Maximum Return to Text NBER Working Paper No. 24976, Tuitions,” NBER Fall 2018 Program. 18 U. M. Dur, P. Pathak, F. Song, August 2018. Return to Text and Tayfun Sönmez, “Deduction Return to Text 17 S. Cellini, R. Darolia, and L. Dilemmas: The Taiwan Assignment 16 M. Baird, M. Kofoed, T. Miller, Turner, “Where Do Students Go when Mechanism,” NBER Working Paper and J. Wenger, “For-Profit Higher For-Profit Colleges Lose Federal Aid?” No. 25024, September 2018. Education Responsiveness to Price NBER Working Paper No. 22967, Return to Text

8 NBER Reporter • No. 1, March 2019 Research Summaries

The Labor Market for Financial Misconduct

Gregor Matvos and Amit Seru

Financial advisers in the United “matching on misconduct,” in which We find that financial adviser States manage over $30 trillion in advisers with misconduct records misconduct is broader than a few investible assets, and plan the finan- match with firms which specialize in heavily publicized scandals. Roughly cial futures of roughly half of U.S. misconduct, and that the presence of one in 10 financial advisers who work households. At the same time, trust in uninformed consumers may be criti- with clients on a regular basis have a the financial sector remains near all- cal to maintaining this equilibrium. past record of misconduct. Common time lows. The 2018 Edelman Trust We find that similar forces may also misconduct allegations include mis- Barometer ranks financial services as explain gender discrimination in the representation, unauthorized trading, the least trusted sector by consumers, labor market of financial advisers, and outright fraud — all events that finding that only 54 percent of con- leading to a “gender punishment gap.” could be interpreted as a conscious sumers “trust the financial services We conclude by discussing how aca- decision of the adviser. Adviser mis- sector to do what is right.”1 The dis- demic research may help guide evi- conduct results in substantial costs: trust of finance is perhaps not surpris- dence-based policy. In our sample, the median settlement ing in the wake of the recent financial paid to consumers is $40,000, and the crisis and several high-profile scan- Measuring Misconduct mean is $550,000. These settlements dals that have dominated financial cost the financial industry almost half news. While it is clear that some egre- We began our research program a billion dollars per year. A substan- gious fraud occurs in the financial by documenting the extent of miscon- tial number of financial advisers are services sector, it is less clear whether duct in the financial advisory indus- repeat offenders. Past offenders are misconduct is limited to a few scan- try.2 To study misconduct by financial five times more likely to engage in dals or is a pervasive feature of the advisers, we construct a panel data- misconduct than otherwise compa- industry. Moreover, if misconduct is base of the universe of financial advis- rable advisers in the same firm, at pervasive, why can it survive in the ers (about 1.2 million) registered in the the same location, at the same point marketplace, and conversely, which United States from 2005 to 2015, rep- in time. The large presence of repeat mechanisms constrain it from envel- resenting approximately 10 percent of offenders suggests that consumers oping the entire industry? total employment of the finance and could avoid a substantial amount of This summary describes our insurance sector. Our data, which we misconduct by avoiding advisers with research, which is joint work with have made available to other research- misconduct records. Mark Egan, on these questions. We ers, come from the Financial Industry Moreover, misconduct is not ran- start by describing how we measure Regulatory Authority’s (FINRA) domly allocated across firms. We find misconduct among all registered BrokerCheck website. The data con- large differences in misconduct across financial advisers in the U.S. We then tain advisers’ complete employment some of the largest and best-known turn to the role of labor markets in history. Because the industry is heav- financial advisory firms in the U.S. constraining misconduct, document- ily regulated, data on adviser qualifi- Figure 1 [following page] displays the ing that while some firms penalize cations provide a granular view of job top 10 firms with the highest share misconduct through a sharp increase tasks and roles in the industry. Central of advisers that have a record of mis- in job separations, other firms are will- to our analysis, FINRA requires finan- conduct. Almost one in five financial ing to hire these advisers, recycling cial advisers to formally disclose all cus- advisers at Oppenheimer & Co. had a the bad apples in the industry. We tomer complaints, disciplinary events, record of misconduct. Conversely, at then discuss evidence that suggests and financial matters, which we use to USAA Financial Advisors, the ratio this phenomenon arises because of construct a measure of misconduct. was less than one in 36.

NBER Reporter • No. 1, March 2019 9 These simple statistics lead to two in which misconduct is tolerated by half of advisers do not keep their direct questions. First, given the pres- firms. Firms do not fire advisers who job in the subsequent year. The job ence of repeat offenders, why do mar- engage in misconduct, and they toler- turnover rate among advisers with ket forces and regulators not drive ate a misconduct record in employing recent misconduct is 48 percent per these bad apples from the industry? new advisers. In this case, tolerance year — 29 percentage points higher Second, why can advisers and firms of misconduct is extreme. Our results than among advisers without recent that consistently misconduct. Firms engage in miscon- account for the duct coexist in the Financial Advisory Firms with the Highest Incidence of Misconduct severity of miscon- market with firms duct when doling and advisers with Share of advisors with misconduct records out punishments. clean records? 0% 5% 10% 15% 20% Advisers whose mis- Oppenheimer & Co. conduct results in Labor Market First Allied Securities higher monetary Discipline Wells Fargo Advisors Financial Network costs or those with UBS Financial Services more egregious mis-

One would Cetera Advisors conduct such as

expect labor markets Securities America fraud and forgery are to discipline mis- more likely to lose conduct. In fact, one National Planning Corporation their jobs following might expect that Raymond James & Associates misconduct. firms, wanting to Stifel, Nicolaus & Company Although firms protect their reputa- Janney Montgomery Scott are strict in disci- tion for honest deal- plining misconduct, ing, would fire advis- the industry as a Source: Researchers’ calculations using data from the Financial Industry Regulatory Authority ers who engage in whole undoes some misconduct. Other of the discipline firms would have Figure 1 by recycling advis- the same reputation concerns and suggest that labor market behavior ers with past records of misconduct. would not hire such advisers. In this departs from these benchmarks in an Roughly half of advisers who lose scenario, advisers would be purged interesting way: while some firms fire a job after engaging in misconduct from the industry immediately fol- advisers who engage in misconduct, find new employment in the indus- lowing misconduct, and only advis- other firms hire these advisers, recy- try within a year. In total, roughly ers with a clean record would survive cling bad apples in the market. 75 percent of those advisers who in equilibrium. Under this bench- In fact, the average firm is rela- engage in misconduct remain active mark, punishment is extreme. There tively strict in disciplining miscon- and employed in the industry the fol- is an alternative scenario, however, duct. Following misconduct, roughly lowing year. Industry reemployment

​Gregor Matvos is a professor of His papers in these areas have been finance at the University of Texas at published in several journals, including Austin, McCombs School of Business. American Economic Review, Journal of He is a research associate in the NBER’s Political Economy, The Journal of Finance, Corporate Finance group, and serves as an Journal of Financial Economics, and The editor of The Review of Corporate Finance Review of Financial Studies. His research Studies and an associate editor of The has been featured in major media, Journal of Finance. He was previously an including Bloomberg, The Economist, associate editor at Management Science Financial Times, The New York Times, and a faculty member at the University of and The Wall Street Journal. Chicago Booth School of Business. Born and raised in Slovenia, Matvos Matvos is interested in issues related earned a BA in economics and a PhD to financial intermediation, household in business economics from Harvard finance, and corporate finance. University.

10 NBER Reporter • No. 1, March 2019 helps explain why recidivism is so the information on misconduct is populations — those counties below high, even though the average firm is publicly available. the national averages in terms of strict in disciplining misconduct. We find evidence that the pres- household incomes and college edu- Critical to understanding the ence of unsophisticated consumers is cation rates. phenomenon of recycling bad apples one of the central frictions enabling To summarize, unsophisticated is firm and employee “matching on misconduct. Even though miscon- consumers contribute to the exis- misconduct.” Advisers with past duct records are public information, tence and survival of firms that con- records of misconduct tend to move such unsophisticated customers do sistently engage in misconduct. By to firms whose current employees are not know either that such disclo- rehiring advisers with misconduct more likely to be actively engaging in sures exist, or how to interpret them. records, high-misconduct firms blunt misconduct, and which punish mis- Differences in sophistication lead to the market discipline of low-miscon- conduct less severely. The willingness market segmentation. Some firms duct firms. to recycle advisers with past miscon- “specialize” in misconduct and attract duct, and the “matching on miscon- unsophisticated customers, and oth- The Gender Punishment Gap duct,” undermines discipline in the ers cater to more sophisticated cus- financial advisory industry. tomers who recognize misconduct. In addition to finding that the We find evidence consistent with the labor market for financial advisers Misconduct in Equilibrium: idea that markets are segmented along recycles bad apples and results in Why Does Misconduct Survive? dimensions of consumer sophistica- matching on misconduct, we also find tion. Misconduct is more common evidence of a “gender punishment Why can firms that employ advis- among financial advisers who deal gap.”4 Following an incident of mis- ers who engage in misconduct survive with customers who are deemed less conduct, female advisers are 9 per- in equilibrium with firms that do not sophisticated by regulators. The type centage points more likely to lose engage in this activity? One would of compensation firms charge to cli- their jobs than their male counter- expect consumer demand to reflect ents is correlated with misconduct. parts. Figure 2 [following page] dis- the reputation for misconduct in the Advisory firms that charge based plays job turnover in the financial product market. Eugene Fama, in one either on assets under management advisory industry for male and female paper, and Fama and Michael Jensen or commissions tend to have higher advisers with and without miscon- in another, argue that competition rates of misconduct than firms that duct records. After engaging in mis- should lead to career punishments in charge based on performance. conduct, 54 percent of male advis- labor markets.3 Then advisers who The geographic distribution of ers retain their jobs the following engage in misconduct and the firms advisory firms is also consistent with year while only 45 percent of female that employ them, would not survive market segmentation along the lines advisers retain their jobs, despite no in the market for long. Which fric- of investor sophistication. We find differences in turnover rates for male tions prevent the market from achiev- evidence that rates of misconduct and female advisers without miscon- ing this outcome? This fact is even are 19 percent higher, on average, duct records (19 percent). The gen- more puzzling when considering that in regions with the most vulnerable der punishment gap extends beyond

Amit Seru is the Steven and Roberta and financial advisers. He is also interested in Denning Professor of Finance at Stanford interactions of internal organization of firms University’s Graduate School of Business, a with financing and investment, with partic- senior fellow at the Hoover Institution and ular emphasis on technological innovation. the Stanford Institute for Economic Policy His papers in these areas have appeared in Research, and a research associate in the several leading journals. He is a co-editor of NBER’s Corporate Finance and Monetary The Journal of Finance and was previously Economics groups. He was formerly a faculty editor of The Review of Corporate Finance member at the University of Chicago Booth Studies, a department editor of Management School of Business. Science, and an associate editor of Journal of Seru’s primary research interest is in Political Economy. corporate finance; he is interested in issues Seru earned a BE and an MBA from related to financial intermediation and regu- the University of Delhi. Subsequently, lation, with recent work focusing on regula- he received a PhD in finance from the tory aspects of shadow banks, fintech lenders, University of Michigan.

NBER Reporter • No. 1, March 2019 11 the firm at which misconduct took managers decrease the gender pun- Policy Response place to other reemployment oppor- ishment gap. In other words, manag- tunities in the industry. While half ers only alleviate the punishment gap Our results suggest that finan- of male advisers find new employ- within their gender and ethnic group. cial firms and regulators may want ment after losing their jobs following This evidence is important because to pay close attention to “high-risk” misconduct, only a third of female it rules out several potential alterna- financial advisers with misconduct advisers find new records who are recy- employment. Misconduct and Employment of Financial Advisers cled across firms in We find no evi- the industry. Since dence that occu- Advisers with history Advisers with history our research findings pational segrega- of no misconduct of misconduct were first circulated, tion drives this 100% 100% there have been sev- gap. Because of the eral policy initiatives incredible richness 45% in this direction. The of our regulatory 54% Office of Compliance data, we are able to 81% 81% Inspections and compare the career 37% Examinations with outcomes of male 22% the Securities and and female advisers 10% 9% Exchange Com­ 24% 18% who are working 0% 9% 10% 0% mission,­ the at the same firm, Men Women Men Women Massachusetts in the same loca- Securities Division, tion, at the same Join dierent firm Leave the industry Remain with the firm FINRA, and the point in time, and Financial Stability in the same job Source: Researchers’ calculations using data from the Financial Industry Regulatory Authority Board have added role. Differences in hiring and employ- production, or the Figure 2 ment of high-risk and nature of miscon- recidivist brokers to duct, do not explain the gap. If any- tives, under which firms with female their examination priorities. thing, misconduct by female advisers or minority male executives attract Our work also suggests that unso- is on average substantially less costly a pool of individuals with selected phisticated consumers are essential for firms. misconduct propensities. The gen- to the survival of misconduct in this The gender punishment gap der punishment gap we identify is a market. Increasing market transpar- increases in firms with a larger share potentially less salient form of dis- ency and providing consumers with of male managers at the firm and crimination that may limit the careers access to more information could branch levels. For example, we find of women working in a high human reduce the number of unsophisti- no evidence of a punishment gap at capital, well-paid industry. cated consumers. FINRA’s promo- firms with an equal representation Our findings imply that too many tion of its BrokerCheck website is a of men and women at the executive female advisers with untarnished step in this direction. We have also level. Conversely, at firms with no records are purged from the industry constructed a new website (www. female representation at the executive while too many fraudulent male advis- eganmatvosseru.com) to help raise level, women are 32 percentage points ers remain in the market, resulting in awareness of adviser fraud and to more likely to lose their jobs follow- more misconduct. Gary Becker sug- provide resources for policymakers, ing misconduct than are their male gested that markets combat discrim- practitioners, and researchers. counterparts. We extend our analysis ination because discriminators are to men with names that are relatively harmed with lower profits.5 However, common in minority groups and find misconduct-tolerant firms that have that the punishment gap and pat- male advisers with records of miscon- 1 2018 Edelman Trust Barometer, terns of in-group tolerance extend to duct can survive in equilibrium due to Global Report; M. Egan, G. them as well. These results also sug- the presence of unsophisticated con- Matvos, and A. Seru, “The Market gest that the in-group tolerance we sumers. In other words, the product for Financial Adviser Misconduct,” observe is not driven solely by gen- market equilibrium with unsophisti- NBER Working Paper No. 22050, der-specific factors. In addition, we cated consumers may make it easier to February 2016. find no evidence that male minority discriminate against female advisers. Return to Text

12 NBER Reporter • No. 1, March 2019 2 Ibid, NBER Working Paper No. “Agency Problems and Residual Misconduct,” NBER Working Paper 22050. Claims,” the Journal of Law and No. 23242, March 2017. Return to Text Economics, 26(2), 1983, pp. 327–49. Return to Text 3 E. Fama, “Agency Problems and Return to Text 5 G. Becker, The Economics of the Theory of the Firm,” Journal of 4 M. Egan, G. Matvos, and A. Discrimination, Chicago, Illinois: Political Economy, 88(2), 1980, pp. Seru, “When Harry Fired Sally: University of Chicago Press, 1957. 288–307; E. Fama and M. Jensen, The Double Standard in Punishing Return to Text

NBER Reporter • No. 1, March 2019 13 The Return of Survey Expectations

Andrei Shleifer

The use of survey expectations data based predictors from a rational expec- was a key feature of macroeconomics tations model. People literally put their in the 1950s and 1960s, and an impor- money where their mouth is — not tant part of research at the NBER dur- where it ought to be in rational expec- ing that period. Yet this work slowly tations models. Last but not least, the ground to a halt in the aftermath of the evidence shows that forecast errors can rational expectations revolution. Under be predicted from the information that Andrei Shleifer is John L. Loeb rational expectations, economic agents the decision maker has at the time of Professor of Economics at Harvard forecast the future by optimally using making the forecast. This is inconsistent University. He holds an under- the true structure of the economy they with the rational expectations hypoth- graduate degree from Harvard and operate in. This means that the struc- esis, and points to more realistic eco- a PhD from MIT. Before com- ture of the economy itself dictates what nomic models of expectation formation ing to Harvard in 1991, he taught beliefs they should hold. From the view- and actual behavior. In this summary, I at Princeton University and the point of economic research, this implies review some of my research that contrib- University of Chicago Booth that expectations data are redundant as utes to these findings, conducted jointly School of Business. He is a research long as the econometrician knows the with Pedro Bordalo, Nicola Gennaioli, associate in the NBER’s Corporate model that economic agents rely on and Robin Greenwood, Rafael La Porta, and Finance, Public Economics, Asset can compute their statistically optimal Yueran Ma. Many other researchers have Pricing, Economic Fluctuations expectations of future variables from that generated closely related results. and Growth, Law and Economics, model. In financial economics as well as and Political Economy programs. in macroeconomics, the rise of the effi- Expectations of Aggregate Shleifer has worked in the areas cient markets hypothesis rendered expec- Stock Returns of comparative corporate gover- tations data largely irrelevant for address- nance, law and finance, behavioral ing key questions. Perhaps because the movements of finance, and institutional econom- Although the last 30 to 40 years the stock market engage so many peo- ics. He has published seven books, have seen occasional studies using survey ple, from individual investors to manag- including The Grabbing Hand, expectations data, this line of work has ers to professional forecasters, there are with Robert Vishny; Inefficient picked up pace significantly in the last multiple sources of data on expectations Markets: An Introduction to several years. Part of the reason is that we of aggregate stock market returns. Robin Behavioral Finance; and A Crisis now have much better data, and that sur- Greenwood and I put together data on of Beliefs: Investor Psychology and vey expectations are actually quite useful such expectations, some quantitative and Financial Fragility, with Nicola in distinguishing alternative models, but some qualitative, from six sources, with Gennaioli, as well as over a hun- part is undoubtedly the fact that ratio- very diverse surveyed populations and dred articles. nal expectations models in both mac- different survey questions.1 Shleifer is an editor of the roeconomics and finance have increas- The first source is the Gallup survey of Q uarterly Journal of Economics, ingly reached dead ends. As a result, individual investors, with data from 1996 and a fellow of the Econometric survey expectations are staging a remark- to 2012. For most of this period, this sur- Society, the American Academy able comeback. vey asked respondents about their beliefs of Arts and Sciences, and the Rapidly growing evidence shows about stock market returns over the next American Finance Association. In that, far from being random noise, mea- year, with possible answers ranging from 1999, he won the John Bates Clark sured expectations are highly consistent “very optimistic” to “optimistic” to “neu- Medal of the American Economic across surveys that are conducted with tral” to “pessimistic” to “very pessimistic.” Association. According to the different methodologies and using some- One can construct a qualitative indicator organization Research Papers in what different questions. Furthermore, of return expected by Gallup respondents Economics (RePEc), Shleifer is the actual behavior of survey respondents as the difference between the percentages most-cited economist in the world. is predicted more successfully by their of bullish and bearish investors. Between survey responses than by some model- 1998 and 2003, the survey also asked for

14 NBER Reporter • No. 1, March 2019 quantitative estimates of expected stock esis that survey expectations are merely past returns are high, expected future market returns. Because during this over- noise. Indeed, we also find that investor returns are on average higher than realiza- lap period the movements in the qualita- expectations are highly correlated with tions. A quick summary of what the data tive bullishness indicator and the quan- the flow of money into equity mutual say is that when stock market returns have titative expectations of returns are highly funds. When investors are optimistic been high, investors expect high returns correlated, the overlap allows us to map about expected stock returns, they put to continue, but in reality the returns are, the qualitative indicators into quantitative money into equity mutual funds. This if anything, on average low. measures of expected return over the entire evidence shows that investors act on their Greenwood and I also find that period since 1996. beliefs. Survey expectations can be used expected returns dictated by standard The second source is the survey of to predict economically relevant choices. rational expectations asset pricing mod- Chief Financial Officers of large U.S. com- If these expectations are highly cor- els are strongly and statistically signifi- panies, conducted since 1998 by John related across data sources and investor cantly negatively correlated with survey Graham and Campbell Harvey at Duke types, then what do they reflect? Figure expectations of returns. According to University. Here respondents provide 1 plots, on the same graph, Gallup expec- efficient markets models, stock prices quantitative estimates are high when investors of the expected stock are willing to accept low market return over Expectations of Future Stock Returns and S&P 500 Past Returns returns going forward the next year. The because their wealth or Expectations of returns over S&P 500 returns over other four sources the next 12 months the trailing 12 months consumption is high are a member sur- 60% 40% today. With extrapola- vey of investor sen- 50 30 tion, in contrast, stock timent conducted 40 20 prices are high precisely by the American 30 10 because, by extrapolat- Association of 20 0 ing past returns, inves- Individual Investors; 10 -10 tors incorrectly expect 0 -20 a sentiment measure -10 -30 future prices to become constructed by the -20 -40 even higher, bidding up editors of Investors -30 -50 current prices. This evi- Intelligence newsletter -40 -60 dence explains, in the going back to 1963, 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 clearest way, how the summarizing the rational expectations investment outlook Expectations of returns is built from a Gallup survey of individual investors’ expectations models get it wrong. Source: N. Gennaioli, Y. Ma, and A. Shleifer, NBER Working Paper No. 21260 and published as of more than 120 “Expectations and Investment,” NBER Macroeconomics Annual, 30(1), 2015, pp. 379–431 The models need inves- independent financial tors to expect low market newsletters; Figure 1 returns in good times, Robert Shiller’s sur- the opposite of what vey of individual inves- they say they expect. tor confidence in the stock market; and tations of the next 12-month return and the University of Michigan Surveys of the past 12-month return on the S&P Expectations in a Cross- Consumers, which date to 1946. 500. The two series practically lie on top Section of Stock Returns Greenwood and I put these six data of each other. Gallup investors, as well as sources into comparable units and use participants in other surveys, expect high In 1996, La Porta published an average responses in each survey every future returns precisely when the last intriguing finding.2 He compared com- month. We thus have a monthly time year’s stock market return has been high. panies with the most optimistic and most series of expectations of returns on the They appear to extrapolate past returns pessimistic long-term earnings growth stock market from six sources. We find into the future. forecasts made by financial analysts. He that, at any given point in time, differ- Is such extrapolation an effective found that stocks with the most opti- ent financial market participants, ranging return forecasting strategy? The answer mistic analysts earn sharply lower returns from individuals to CFOs, have highly is no. The correlation between the Gallup than those with the most pessimistic correlated expectations or sentiment expected return and the S&P return for ones. It is not just that when analysts about future returns. Despite the con- the next — as opposed to the past — 12 are extremely optimistic, realized long- siderable heterogeneity of data collec- months is negative, though statistically term earnings growth proves to be slower tion procedures across these surveys, this insignificant. At the same time, high past than analysts expect. Their optimism also evidence decisively rejects the hypoth- returns predict expectation errors. When infects prices, perhaps because they influ-

NBER Reporter • No. 1, March 2019 15 ence investors or per- ings data: HLTG haps because investors Evolving Expectations of Long-Term Earnings Growth firms experience pos- hold similar beliefs. This itive earnings sur- leads to overvaluation Expected annualized long-term earnings growth Date when stocks are ranked prises on the way up, in top or bottom decile and to low subsequent 40% yet disappoint look- 35 Portfolio of equities with high returns on the stocks forecasted long-term earnings growth ing forward from the whose growth prospects 30 peak. LLTG firms the analysts find most 25 disappoint on their favorable. 20 way down, but turn Bordalo, Gennaioli, 15 Portfolio of equities with low out not to be as bad forecasted long-term earnings growth La Porta, and I revisited 10 as expected look- La Porta’s finding with 20 5 ing forward from additional years of data 0 the bottom. To put and took a much closer -3 -2 -1 0 1 2 3 this differently, ana- look at how expectations Years before and aer portfolio creation lysts extrapolate are formed, how they are past performance revised, and how earn- In the year of portfolio creation, firms are assigned to portfolios if they are in and become exces- the top or bottom decile of analysts’ expectations for future earnings growth ings, beliefs, and prices Source: P. Bordalo, N. Gennaioli, R. La Porta, and A. Shleifer, NBER Working Paper No. 23863 sively optimistic co-evolve.3 La Porta’s ini- about HLTG firms tial finding still holds: A Figure 2 and excessively pes- portfolio of high long- simistic about LLTG term growth (HLTG) stocks (those with The opposite pattern holds for LLTG firms. And critically, these extreme beliefs the top 10 percent most optimistic earn- firms: Analysts revise their views of these are reflected in valuations as well, lead- ings-growth forecasts) returns 3 percent firms down prior to portfolio formation, ing to sharp differences in returns as ana- on average in the year after formation dur- but subsequently recognize that long- lysts and others correct their mistaken ing the 1981–2015 sample period. In con- term earnings growth will not be as bad as extreme beliefs. trast, a portfolio of low long-term growth they anticipated and correct their beliefs This evidence raises a deeper ques- (LLTG) stocks — those with the bottom upward. Portfolio returns follow fore- tion about how analysts form beliefs in 10 percent most pessimistic earnings- cast revisions: As analysts curb their ear- the first place. Do they mechanically growth forecasts — returns 15 percent on lier enthusiasm, returns are poor; as they extrapolate past earnings growth trends, average in the year after formation. The lat- cheer up, returns are high. or is there more to their belief forma- ter stocks are a much better investment, on We find similar trends in the earn- tion? Our work suggests that belief for- average. mation is not mechanical, Why are HLTG but takes a particular stocks such a bad form of sophisticated investment? First, as Density Estimates of Growth in Earnings for Long-Term Growth Portfolios yet not entirely rational Figure 2 shows, analyst learning. We find that Distribution density expectations of long- 2.5 among HLTG firms, term earnings growth Mean = 1.07 Mean = 1.12 there are in fact some are at their peak at 2.0 whose earnings con- the time of portfolio tinue to grow spectac- Equities not in the top decile 1.5 formation for HLTG of analysts’ expectations for ularly — they are the firms. These expecta- growth in long-term earnings future Googles. It is just Equities in the top decile 1.0 tions have been rising, of analysts’ expectations for that most HLTG firms along with earnings, growth in long-term earnings do not turn out to be for several years prior 0.5 Googles, but slow down to portfolio forma- instead. Analysts form tion and decline in the 0.0 average expectations years immediately fol- 0 0.5 1 1.5 2 for HLTG firms as if lowing. Analysts learn Annual growth rate of earnings per share there are more future that they are too opti- Googles among them mistic about HLTG Source: P. Bordalo, N. Gennaioli, R. La Porta, and A. Shleifer, NBER Working Paper No. 23863 than in reality there are. firms and revise their Figure 3 illustrates beliefs accordingly. Figure 3 this finding. It shows

16 NBER Reporter • No. 1, March 2019 the distribution of long-term earnings These CFOs also report their expecta- hypothesis, and the evidence again points growth rates from HLTG firms (solid tions of earnings growth and investment to extrapolation: The CFOs expect the curve) and other firms (dashed curve). plans for their own companies. profitability of their firms to be more per- HLTG firms have a fat right tail of per- Figure 4 shows CFO earnings growth sistent than it turns out to be, on average. formance outcomes relative to all firms. expectations and investment plans for the In sum, recent research points to There are indeed more Googles among next year, aggregated over the firms in the many new opportunities presented by them, but not nearly as many as ana- Duke sample. It shows that the two fol- survey expectations data. These include lysts think. Analysts tests of rationality of use information about CFOs’ Earnings Growth Expectations and Investment Plans expectations, but more past performance but importantly suggest

overreact by predict- 30% avenues for construct- ing too many high ing new models in 25 Expectation of earnings growth performers. over the next 12 months macroeconomics and So we see that, in 20 finance. These models the cross-section of 15 could seek to fit not stocks, there are some 10 just data on quanti- very similar phenom- 5 ties and prices, but also ena to those we saw 0 data on the beliefs of in the aggregate. Good -5 Planned capital expenditure households and firms. performance leads to -10 growth over the next 12 months Research on these predictions of good -15 issues could represent performance in the -20 a return to issues that future, and the oppo- 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 engaged many NBER site for bad perfor- researchers in the past, mance. Forecasters Source: N. Gennaioli, Y. Ma, and A. Shleifer, NBER Working Paper No. 21260, and published as this time with better extrapolate, but not “Expectations and Investment” in NBER Macroeconomics Annual, 30(1), 2015, pp. 379–431 data, evidence, and mechanically. They use theory. some forward-looking Figure 4 logic. Analysts under- stand correctly that extremely high earn- low each other closely. The collapse of 1 R. Greenwood and A. Shleifer, ings growth portends extremely high con- earnings growth expectations following “Expectations of Returns and Expected tinued growth for some firms; there is the Lehman bankruptcy, for example, is Returns,” NBER Working Paper No. indeed information in high past earn- accompanied by the collapse of invest- 18686, January 2013, and The Review ings growth for future earnings growth. ment plans. The two recover together in of Financial Studies, 27(3), 2014, pp. Among the HLTG firms, there actually 2009–10 as well. Actual investment is 714–46. are some remarkable performers going very close to investment plans, so CFO Return to Text forward, many more than among LLTG beliefs go hand in hand with economic 2 R. La Porta, “Expectations and the firms. Unfortunately, such performers activity. The study also shows that CFO Cross-Section of Stock Returns,” The tend to be relatively few, and the forecasts earnings growth expectations are a much Journal of Finance, 51(5), 1996, pp. become overly optimistic by expecting stronger predictor of investment plans 1715–42. too many of them. Analysts react exces- (as well as realizations) than Tobin’s q, Return to Text sively to the good news of high earnings the ratio of a company’s market value to 3 P. Bordalo, N. Gennaioli, R. La Porta, growth. Extrapolation is caused by over- the replacement cost of its capital stock, and A. Shleifer, “Diagnostic Expectations reaction to information. which is the preferred indicator of invest- and Stock Returns,” NBER Working ment opportunities in a standard rational Paper No. 23863, September 2017, and Expectations and Investment expectations investment framework. forthcoming in The Journal of Finance. But are CFO expectations ratio- Return to Text Do survey expectations shape corpo- nal? It appears that they are not. Rather, 4 N. Gennaioli, Y. Ma, and A. Shleifer, rate decisions as well? Gennaioli, Ma, and the evidence shows that CFOs are exces- “Expectations and Investment,” NBER I have explored this question for corpo- sively optimistic about their firms’ earn- Working Paper No. 21260, June 2015, rate investment.4 We use the Duke CFO ings growth when past profitability has and the NBER Macroeconomics Global Business Outlook survey of large been high, and conversely when it has been Annual 2015, 30(1), 2015, pp. 379– U.S. companies already mentioned as a low. The predictability of forecast errors is 431 source of data on expected stock returns. inconsistent with the rational expectations Return to Text.

NBER Reporter • No. 1, March 2019 17 What Inventions Are We Missing?

Heidi Williams

Academics and policymakers have long would have occurred under counterfactual recognized that competitive markets may patent policies. In this piece, I summarize under-incentivize innovation. This concern some of the main findings that have emerged has motivated the design of public policies from my research in this area. such as the patent system, which aims to encourage research investments into new Measuring Innovation Heidi Williams is a technologies by allowing inventors to cap- research associate in the ture a higher share of the social returns to Traditionally, economists who study NBER’s Aging, Health their research investments. innovation have relied on patent counts (or Care, and Productivity A well-developed theoretical literature citation-weighted patent counts) as a mea- Programs. She is an asso- has analyzed optimal patent policy, with a sure of innovation, often leveraging the data ciate professor of econom- focus on the trade-off between providing constructed by Bronwyn Hall, Adam Jaffe, ics at MIT, and is currently incentives for the development of new tech- and Manuel Trajtenberg.1 Although this Trione Visiting Associate nologies and tolerating higher prices during approach has been useful in many settings, Professor at the Stanford the life of the patent. Although such theoret- it encounters two major limitations. First, Institute for Economic ical models — and, importantly, public poli- in many cases it is difficult or impossible to Policy Research. She is cies — typically assume that stronger (e.g. match patents with the specific products co-editor of the Journal of longer or broader) patents will induce addi- they protect, or to identify specific groups Economic Perspectives. tional research investments, there is remark- of consumers that might benefit from those Her research focuses on ably little empirical evidence on how patents products. For example, the text in a patent investigating the causes and affect research investments in practice. protecting a delivery method for a breast consequences of technolog- This question has been difficult to tackle cancer drug may have no information sug- ical change, with a partic- empirically for at least two reasons. First, gesting the patent is relevant to breast can- ular focus on health care measuring research investments can be quite cer patients. Hence, it can be very difficult to markets. Williams received challenging. Second, finding variation in use patents to measure research investments her AB in mathematics patent protection that can be leveraged in in a way that can be linked to product-mar- from Dartmouth College an empirical study is difficult. On paper, the ket or consumer-level outcomes. Second, by in 2003, her MSc in devel- U.S. patent system is uniform, providing a construction, patent data can only measure opment economics from 20-year period of protection for all inven- patented inventions. Because many tech- Oxford University in 2004, tions. While historically some cross-country nologies are not patented, changes in patent and her PhD in econom- variation in patent laws has existed, because counts may in some settings reflect chang- ics from Harvard in 2010. innovations are generally developed for a ing levels of inventors’ willingness to file for Her research has received global market, country-specific patent law patents on their research investments, rather support from the National changes may often induce only a relatively than changes in the underlying research Institutes of Health and small change in global research incentives. investments themselves. an NSF CAREER grant, My research agenda attempts to over- A major focus of my research agenda among others, and she has come both of these challenges in order has been to attempt to overcome these two received a Sloan Research to develop empirical estimates of the key challenges by compiling “real,” non-patent Fellowship and a MacArthur parameters needed to inform optimal patent measures of innovation. For example, Eric Foundation Fellowship. policy. By combining detailed new measures Budish, Ben Roin, and I aimed to develop Williams grew up in of research investments with novel sources measures of research investments in can- North Dakota, but has of variation in the effective patent terms pro- cer drugs.2 The core of our data construc- most recently been living in vided to otherwise similar inventions, my tion was to take advantage of a National Cambridge and Palo Alto work aims to construct frameworks within Cancer Institute (NCI) clinical trial reg- with her husband and two which we can infer the volume, type, and istry that includes an explicit listing of the young sons. value of “missing” research investments that patient groups eligible to enroll in each

18 NBER Reporter • No. 1, March 2019 clinical trial. Cancer treatment tends to cal trials that relate to specific genetic Pharmaceutical firms face strong incen- be specific to an organ of origin, such as variations. These types of curated scien- tives to file patent applications at the prostate, and stage of disease, for exam- tific identifiers can thus provide unique time of invention. However, prior to ple, metastatic. As such, the organ-stage opportunities to trace out meaningful marketing a new drug, firms must sub- classification tends to be used both to links between basic scientific discoveries mit clinical trial results to the U.S. Food label clinical trial-eligible groups (as in and commercialized products. and Drug Administration (FDA) docu- the NCI data, where such classifications Our goal in constructing these data is menting that their product meets a set are used to describe which patients can to apply them in order to test and evalu- of safety and efficacy standards. These enroll in any given clinical trial) and to ate theories about economic factors that clinical trials generate a lag between the label patients in standard clinical datas- may be encouraging or hindering innova- time of patenting and the time of com- ets (e.g. the Surveillance, Epidemiology, tion. Let me now summarize some of the mercialization, which reduces an inven- and End Results cancer registry data, substantive findings of our research. tion’s effective patent life — more so for which provides data on the survival drugs that require longer clinical trials. outcomes of U.S. cancer patients). Incentives to Develop New Drugs A key determinant of clinical trial length Leveraging this organ-stage classifica- is patient survival time: Because clinical tion hence allowed us to use clinical tri- Why don’t we have a cure for can- trials generally must show evidence that als as a real measure of research invest- cer? Informal discussions with doctors the treatment improves mortality-related ments in new cancer drugs, and to link and scientists usually provide a variety of outcomes, clinical trials tend to be lon- those research investments to measures answers to this question, many of which ger — and hence, effective patent terms of patient health outcomes — namely, boil down to some version of “the basic shorter — when enrolling patients with survival outcomes for different groups science is hard.” But from an economic longer survival times. All else being equal, of cancer patients over time. perspective, we are of course inclined firms are thus awarded longer terms of As a second example, two of my to think that how “hard” any given sci- market exclusivity for successfully devel- papers — one joint with Bhaven entific problem seems might reflect not oping drugs to treat patients with shorter Sampat — have constructed data on non- only the innate difficulty of the problem, survival times (such as patients with patent measures of innovation related to but also the level of past research efforts late stage, metastatic cancers) relative to the human genome.3 The set of human to solve it. drugs to treat patients with longer sur- genes is curated by scientists in a way that Budish, Roin, and I investigated vival times (such as patients with early assigns the equivalent of Social Security one aspect of this question — namely, stage, localized cancers). This motivates numbers — unique identifiers for each whether private firms underinvest in the following empirical test: If longer gene — called Entrez Gene IDs. These cancer drug development projects patent terms encourage more research Entrez Gene IDs are in turn linkable to that require a long time to complete.4 investments, we should see higher lev- various databases which catalog scientific The basic idea of our study was this: els of private research investments on papers that have doc- treatments for patient umented evidence for groups that require Share of Clinical Trials that are Privately Financed links between genes shorter clinical trials. and diseases; this Using the cancer Share of trials that are privately financed “paper trail” provides a 50% clinical trial registry consistent measure of data described above, scientific publications 40 we document several related to a given gene. sources of empirical Entrez Gene IDs can evidence that together also be linked to prod- 30 are consistent with uct market databases private research invest- such as the GeneTests. 20 ments being distorted org database, which away from long-term provides information 10 projects such as drugs on the use of genes 0% 20 40 60 80 100 to prevent or treat in gene-based diag- Five-year patient survival rate early-stage cancers. nostic tests, and the For example, Figure 1,

Pharmaprojects data- E. Budish, B. Roin, and H. Williams, NBER Working Paper No. 19430, and published as “Do firms underinvest in long-term reproduced from our base, which provides research? Evidence from cancer clinical trials” in American Economic Review, 105(7), 2015, pp. 2044-2085 paper, documents that information on drug research on cancers compounds in clini- Figure 1 which require longer

NBER Reporter • No. 1, March 2019 19 clinical trials because they have higher follow-on innovation. A well-developed rounded the question of whether Celera’s five-year survival rates is more likely to be literature has documented theoretically data not being open access would hinder publicly funded. A back-of-the-envelope ambiguous predictions on how patents subsequent scientific research and prod- calculation we present suggests that this affect follow-on innovation, but there is uct development. distortion has quantitatively important little available empirical evidence.6 I here I collected records of when each implications for the survival outcomes of describe two papers that aim to shed human gene was sequenced by Celera U.S. cancer patients. light on this question in the context of and by the Human Genome Project,9 Unfortunately, our estimates cannot the human genome. which indicates whether a gene was ever speak directly to the potential role of pat- In 2001, the journals Nature and held with Celera’s intellectual property. ents, since excess impatience of private Science published the initial sequences Genes sequenced first by Celera were firms may also under-incentivize long- of the human genome: Nature pub- held with Celera’s intellectual property term research.5 Our empirical evidence lished a version completed by the pub- until the gene was re-sequenced by the is nonetheless directly relevant to at least licly funded Human Genome Project, Human Genome Project, at which time two policy levers. First, to the extent that and Science published a version com- it moved into the public domain. I linked valid surrogate endpoints — non-sur- pleted by a private firm, Celera.7 As these records to the types of gene-level vival endpoints that innovation measures are known to be caus- Number of Clinical Trials for Patented and Non-Patented Genes described above. Using ally linked with subse- these data, I docu-

quent survival improve- Average log of active gene-related mented that Celera’s ments but which can clinical trials per gene property rights generally be observed 0.12 reduced subsequent more quickly than can 0.10 scientific research and changes in survival out- Genes claimed in a patent product development comes — are available, 0.08 application that was not granted by approximately 30 Genes claimed in a patent our evidence suggests 0.06 application that was granted percent. that allowing clinical Taken at face trials to rely on valid 0.04 value, these results surrogate endpoints Genes never claimed clearly suggest that 0.02 in a patent application can meaningfully data access rules can increase research invest- 0.00 have quantitatively ments and substantially 1995 1997 1999 2001 2003 2005 2007 2009 2011 important effects improve patient health on the rate of inno- outcomes. Second, our B. Sampat and H. Williams, NBER Working Paper No. 21666, and published as “How do patents aect follow-on vation. Because the results suggest a ratio- innovation? Evidence from the human genome,” in American Economic Review, 109(1), 2019, pp. 203-236. timing of that study nale for targeting pub- coincided with a high- lic research subsidies Figure 2 profile U.S. Supreme toward projects that are Court case (AMP v. expected to have long commercialization was described in an accompanying edi- Myriad) that ruled on whether human lags, such as many Alzheimer’s drugs. torial in Nature, the rules surround- genes should qualify as patentable sub- ing data access differed across the two ject matter, a question naturally arose of Human Genomes: efforts.8 The Human Genome Project whether the negative effects of Celera’s Public and Private placed all of its sequenced data in an non-patent intellectual property would open-access database within 24 hours of be similar to the effects of gene patents. The prediction that stronger patent being sequenced, with the stated goal of Sampat and I directly investigated this protection induces additional research maximizing the data’s benefit to society. question.10 investments emerges unambiguously Celera instead chose to protect its data We proposed two quasi-experimen- from a class of theoretical models that with a contract law-based form of intel- tal approaches for investigating how treat innovations as isolated discoveries. lectual property, which allowed free use gene patents affect follow-on innovation. In practice, however, innovation is often of the data for academic research, but First, we compared follow-on innova- “cumulative,” that is, any given discov- placed restrictions on redistribution of tion on genes claimed in accepted ver- ery is also an input into later, follow-on the data and required licensing agree- sus rejected patent applications. Figure discoveries. In such cases, optimal pat- ments to be negotiated for any down- 2 above, reproduced from our paper, ent policy will depend in part on how stream discoveries. As emphasized in the shows that genes included in accepted patents on existing technologies affect Nature editorial, much public debate sur- and rejected patent applications were the

20 NBER Reporter • No. 1, March 2019 subject of a similar number of clinical tri- 2 E. Budish, B. Roin, and H. Williams, Sequencing and Analysis of the Human als both before and after the relevant pat- “Do Firms Underinvest in Long-Term Genome,” Nature, 409(6822), 2001, pp. ent applications were filed. The earliest Research? Evidence From Cancer 860–21; J. Venter, et al., “The Sequence patent application filing date in our sam- Clinical Trials,” NBER Working Paper of the Human Genome,” Science, ple is denoted by the vertical line.11 No. 19430, September 2013, and 291(5507), 2001, pp. 1304–51. Second, we constructed an instru- American Economic Review, 105(7), Return to Text mental variable for whether a patent 2015, pp. 2044–85. 8 “Human Genomes, Public and application is granted a patent, based Return to Text Private,” Nature, 409(6822), 2001, p. on the “leniency” of the conditionally 3 H. Williams, “Intellectual Property 745. randomly assigned patent examiner. In Rights and Innovation: Evidence From Return to Text contrast with the effects I observed with the Human Genome,” NBER Working 9 Ibid, NBER Working Paper No. Celera’s intellectual property, both of Paper No. 16213, July 2010, and 16213. these approaches suggest that gene pat- Journal of Political Economy, 121(1), Return to Text ents have not reduced follow-on inno- 2013, pp. 1–27; B. Sampat and H. 10 Ibid, NBER Working Paper No. vation; in particular, we can statistically Williams, “How Do Patents Affect 21666, and American Economic reject declines in follow-on innovation Follow-On Innovation? Evidence From Review, 109(1), 2019, pp. 203–36. on the order of my earlier estimates for the Human Genome,” NBER Working Return to Text Celera’s intellectual property. Paper No. 21666, October 2015, and 11 The lower line in Figure 2 shows that Both of the quasi-experimental American Economic Review, 109(1), genes that were never included in any approaches developed in this paper have 2019, pp. 203-36. patent applications were the subject of already been re-applied in a number of Return to Text many fewer clinical trials over this entire other papers, by myself as well as other 4 Ibid, NBER Working Paper No. period, suggesting that — perhaps unsur- authors, and I have been pleased to see 19430. prisingly — there was positive selection refinements and improvements of them Return to Text of genes into patent applications. being developed as well, as in the work of 5 See, e.g., J. Stein, “Agency, Return to Text Josh Feng and Xavier Jaravel.12 Information, and Corporate 12 J. Farre-Mensa, D. Hegde, and Taken together, the evidence from Investment,” NBER Working Paper A. Ljungqvist, “What Is a Patent this set of papers suggests that the tra- No. 8342, June 2001, and in G. Worth? Evidence From the U.S. Patent ditional patent trade-off ex( ante incen- Constantinides, M. Harris, and ‘Lottery’,” NBER Working Paper No. tives versus deadweight loss) may be R. Stulz, eds., Handbook of the 23268, December 2018; P. Kline, sufficient to analyze optimal patent pol- Economics of Finance, Amsterdam, N. Petkova, H. Williams, and O. icy design at least in some markets, The Netherlands: Elsevier B.V., (1A), Zidar, “Who Profits from Patents? but that non-patent policies governing 2003, pp. 111–65; J. Stein, Quarterly Rent-Sharing at Innovative Firms,” access to materials — such as data exclu- Journal of Economics, 104(4), 1989, NBER Working Paper No. 25245, sivity — may have important effects on pp. 655–69. November 2018, and forthcoming in follow-on innovation. Return to Text the Quarterly Journal of Economics; 6 For one recent paper that is an excep- P. Gaule, “Patents and the Success of tion to this statement, see A. Galasso Venture-Capital Backed Startups: Using 1 B. Hall, A. Jaffe, and M. Trajtenberg, and M. Schankerman, “Patents and Examiner Assignment to Estimate “The NBER Patent Citation Data File: Cumulative Innovation: Causal Causal Effects,” 2015, unpublished Lessons, Insights, and Methodological Evidence From the Courts,” NBER CERGE-EI Working Paper, Series Tools,” NBER Working Paper No. Working Paper No. 20269, July 2014, No. 546; J. Feng and X. Jaravel, 8498, October 2001, and in A. Jaffe and Quarterly Journal of Economics, “Crafting Intellectual Property Rights: and M. Trajtenberg, eds., Patents, 130 (1), 2015, pp. 317–69. Implications for Patent Assertion Citations, and Innovations, Cambridge, Return to Text Entities, Litigation, and Innovation,” Massachusetts, The MIT Press, 2002. 7 International Human Genome 2016, unpublished Harvard mimeo. Return to Text Sequencing Consortium, “Initial Return to Text

NBER Reporter • No. 1, March 2019 21 Do Government Grants to Charities Crowd Private Donations Out or In?

James Andreoni

The basic hypothesis of crowding out biasing estimates toward crowding in. is simple: Suppose individuals care only In an important and early paper that about their own private consumption and recognized this bias, Abigail Payne esti- the charity’s capacity to spend. In partic- mated crowding out of about 50 percent; ular, they are indifferent to whether their private charitable giving to an organi- James Andreoni is a own gift is voluntary or is involuntarily zation fell by about half the amount of Distinguished Professor of paid through taxes. As a result, any effort government transfers to it.3 Her paper, Economics at the University by policymakers to support this charity like the previous literature, did not treat of California, San Diego. through more involuntary taxes would charities as active participants in the mar- Before joining UCSD in 2006, be met with equal reductions of volun- ket for donations. My recent empirical Andreoni was a professor at the tary gifts as donors work to reestablish work on crowding out, much of it done University of Wisconsin from their optimal total contributions. The net jointly with Payne, aims to look directly 1986 to 2005. He is a fellow effect is that the government support for at the mechanism of crowding out. We of the Econometric Society, a the charity completely crowds out pri- do this by including charities as strate- research associate in the NBER’s vate giving. gic players in a game with donors and the Public Economics program, When confronted with more realis- government. past president of the Economic tic theory, the basic hypothesis of crowd- Science Association, former co- ing out quickly falls apart. First, com- Strategic Charities editor of the Journal of Public plete crowding requires that donors be Economics, and co-founder of “pure altruists,” that is, they engage in One key question related to the link the Association for the Study of consequentialist reasoning.1 If individu- between receipt of a government grant Generosity in Economics. als have other motives for giving, such and a charity’s total resources is how Over his career, Andreoni as private benefits of a “warm glow,” the charity’s fundraising activities will has published widely in the or social motives like image or pride, respond. In particular, are charitable fun- fields of public finance, exper- then complete crowd-out may not hold.2 draisers net revenue maximizers? It is imental and behavioral eco- Notice too that without impurely altru- useful here to follow the distinction that nomics, and economic deci- istic motives for giving there is also lit- non-profits make betweencontinuing sion-making. He is perhaps best tle use for fundraisers. Yet charities have campaigns and capital campaigns. known for his extensive work on very sophisticated and active fundrais- The goal of continuing campaigns is charitable giving, studying what ing operations. This leads us to questions typically to raise enough money to con- motivates people to give, and such as: How does fundraising attract tinue meeting the ongoing needs of the how donors, fundraisers, and donations? What are the objectives of charity. This means that most charities policymakers interact. He has fundraisers? How do government grants will set a funding goal for the year and, also made contributions to the affect both donors and fundraisers? roughly speaking, stop raising money fields of tax compliance, social To examine these questions empiri- when the goal is reached. Managers preferences, delegated enforce- cally, we first need to observe how donors of such charities are said to be satisfi- ment, decision-making by juries, respond to changes in government grants cers rather than maximizers. The conse- revealed preference, and social to the charities they support. In doing so, quence of this is that charities will stop interactions. In recent years, he we must recognize an important source actively raising money even though the has turned to issues of measur- of bias. If there is a natural disaster, marginal return of the last dollar of fun- ing time preferences, and study- for example, then both the donors and draising effort is still greater than a dol- ing the causes of apparent time the government will want to give more lar.4 Moreover, a common measure of inconsistencies in consumption, money for the Red Cross. This will make the quality of a charity used by watchdog labor supply, and moral and eth- it appear that donations and grants to the groups like Charity Navigator is “fund- ical behavior. Red Cross are positively correlated, thus raising efficiency,” defined as the fund-

22 NBER Reporter • No. 1, March 2019 raising expenses divided by total con- ficiently large gift. Winning a govern- ties over more than 15 years.9 In this tributions. This may further discourage ment grant can have the same effect.6 dataset, we are able to measure whether charitable organizations from pursuing individuals gave directly, or through revenue maximization. New Estimates of Crowding Out participation in fundraising events such Capital campaigns, by contrast, are as gala dinners. The results are reported typically about expanding the size or Payne and I have explored the stra- in Figure 1. For overall private giv- scope of the charity. They often involve tegic role of charities in several ways. ing, we measure crowding out of close significant fixed costs, such as new First, we ask whether government to 100 percent of government grant buildings or offices, and may create sit- grants crowd out giving directly, or amounts. Similar to our study using uations where the managers, and some work indirectly by causing a reduction U.S. data, 64 percent of this crowding donors, have more information than in fundraising. 7 We focus our analy- out is attributable to changes in fun- others about the quality of the planned sis on social services organizations and draising efforts of charities. However, expansions. It is difficult for the charity arts organizations. Social services rely the Canadian data reveal a surpris- or an informed donor to credibly con- heavily on government grants, while ing new finding: Individuals who give vey this quality, as both have an incen- the arts do not. We demonstrate that directly are crowded in by govern- tive to mislead others to believe the increases in government funding sig- ment grants, not crowded out. This is charity’s quality is high. nificantly decrease fundraising efforts, consistent with government funding If no single being a signal of qual- donor can pay all ity. Crowding out of the fixed costs asso- How Charitable Donations Respond to Government Grants individual giving is ciated with the cap- entirely attributable Change in charitable donations ital campaign, then +40% to a decline in rev- there is always a zero- enues from fundrais- +20 equilibrium in which +34% “Crowding in” ing events. no one contributes. 0 Finally, in joint Interestingly, a rule -20 research with Sarah -41% -38% -37% -35% “Crowding out” of thumb for fund- -40 Smith, we study the raising in a capital -16% -13% UK lottery grant -60 10 campaign is to raise Direct crowding in/out program. The UK about 30 percent -80 -64% due to government grants requires that 28 per- of the ultimate goal Indirect crowding out due to -100 reduced fundraising expenditures cent of all revenues of from a handful of -120% the National Lottery large donors before be set aside for dis- All sources Non-governmental Fundraising Individuals even announcing the organizations events tribution to UK campaign. By coor- charities. Eighty per- dinating gifts among Source: Researcher’s calculations using data from the Canada Revenue Agency cent of all the money a limited number of available is distrib- large donors — typi- Figure 1 uted through a pro- cally called “leader- gram called Grants for ship givers” — fundraisers can pro- especially for organizations that rely on Large Projects. Large projects are those vide the assurance that the fixed costs grants more heavily. requiring over £60,000 (about $90,000 will be met, and thus eliminate the In a second study, we analyze data at the time). We analyze over 5,000 zero-equilibrium. Moreover, a gov- on more than 8,000 charities operating applications to this program made ernment grant — often called a “seed in the United States.8 We measure an between 2002 and 2005. Importantly, grant’’ — can double as a leadership overall level of crowding out of about all applications are reviewed by a panel gift. These large gifts and grants can 75 percent: private donations fall by of citizens who first assign a score rule out the zero-equilibrium. The about three quarters of the amount to each qualified applicant. The panel effect could be either tocrowd out or of government grants. The bulk of the then meets publicly to discuss the pro- crowd in private donations, depending crowding out, 70 percent, is due to a posals and select the grant recipients. on the scale of the capital investment.5 change in fundraising. In fact, donors Our data include cases where two char- Leadership givers can also signal infor- may be crowded in, as predicted by the ities have similar scores, but the char- mation about the quality of a charity. A informational role of grants as signals ity with the inferior score receives the leadership giver can provide a credible of quality. In a related study, we ana- grant while the other does not. signal of high quality by giving a suf- lyze more than 13,000 Canadian chari- We use a difference-in-differences

NBER Reporter • No. 1, March 2019 23 approach to identify the effect of grant ple stick within these lines, even if for time-inconsistent preferences as funding on donations to the charity, in they are directly asked to give? Or, described above. With a series of fol- which we compare the change in dona- will they find more giving too tempt- low-up experiments, we learned more tions before and after the funding deci- ing? A clever self-control strategy for about the tension underlying individ- sion across successful and unsuccessful such a donor is simply to avoid being ual behavior, focusing on the inter- charities. We find that receiving a grant asked — if the charity will let them. nal struggle between appearing gener- has a positive and significant effect on a A recent field experiment explored ous to others and watching one’s own charity’s total income. In other words, these questions by using the familiar budget. This finding raises deep and these grants do not completely crowd Salvation Army Red Kettle campaigns. interesting questions about the welfare out other funding sources. Indeed, the The Salvation Army allowed Hanna consequences of fundraising and more data again point to crowding in. We Trachtman, Justin Rao, and me to place generally of a shift away from govern- then analyze the effects separately for bell ringers experimentally at one of the ment, and toward private charities, as a different-sized charities. We find smaller two main entrances of a suburban gro- means of solving social problems. charities, those with incomes less than cery store, making them easy to avoid, £1 million per year, show the strongest or at both entrances, making avoid- evidence of crowding in. Moreover, the ance difficult. The bell ringers were 1 J. Andreoni, D. Aydin, B. Barton, positive effect of a grant persists well either silent as people passed (although D. Bernheim, and J. Naecker, “When beyond the year of the award. Finally, we the ringing bells clearly signaled their Fair Isn’t Fair: Understanding Choice observe that grant applications typically receptivity to a gift), or simply said Reversals Involving Social Preferences,” request funds for distinct, well-defined to those who passed by, “Please give NBER Working Paper No. 25257, activities that extend the current mis- today.”11 When the bell ringers were November 2018, and forthcoming sion of the charity. This is consistent silent, and only one door covered, we in the Journal of Political Economy, with the idea that seed funding can found no effect on traffic in and out of shows that non-consequentialism in crowd in other income. the store. When the bell ringers asked social preferences is essential. shoppers to please give today, however, Return to Text Is Fundraising Making traffic through theother door rose by 2 J. Andreoni, “Impure Altruism Us Worse Off ? 30 percent. When both doors were and Donations to Public Goods: A covered, the verbal ask nearly doubled Theory of Warm-Glow Giving,” The Inherent in many studies of fun- giving. This means that those avoiding Economic Journal, 100(401), 1990, draising is an often-unstated premise the bell ringers were not avoiding say- pp. 464–77. The term “warm-glow” that more funds raised means a better ing no, but rather were avoiding say- is slightly pejorative. The intent is society. A revealed preference argu- ing yes and giving. This is an important to remind readers that, although ment supports this: If donors were distinction. It suggests that the fund- the reduced form may help answer not made better off by giving, then raisers in this case were causing people many questions, there is still much they wouldn’t give. But this ignores an to give by making a social interaction to learn by unlocking the ways important social aspect of fundraising: with them difficult to avoid. warm-glow comes about. See also Individuals rarely give without having Another way to see the givers’ J. Andreoni, “Giving with Impure been asked first. Fundraisers often refer dilemma is that they face contradic- Altruism: Applications to Charity and to this as the power of asking — asking tory desires: a temptation to say yes to Ricardian Equivalence,” Journal of is often a powerful percipient of a gift. a fundraiser, and a personal preference Political Economy, 97(6), 1989, pp. What does this mean for the behavior to avoid actual giving. This opens a 1447–58. On social-image in unself- of both charities and donors? new strategy for fundraisers to exploit: ish behavior, see J. Andreoni and One can imagine two ways of Ask people to decide now to give later. B.D. Bernheim, “Social Image and framing this question. First, people If maintaining social image is more the 50–50 Norm: A Theoretical and like to give. Being asked simply reduces tempting than saving money, even a Experimental Analysis of Audience the transaction costs, making it easier very short delay in paying for the gift Effects,” Econometrica 77(5), 2009, to realize the joy of giving. Second, could have significant effects. pp.1607–36. in a more subtle analysis, people with Marta Serra-Garcia and I explored Return to Text big hearts must exert self-control on this question in a lab experiment.12 3 A. Payne, “Does the Government their giving. There are far too many We found that asking people now to Crowd Out Private Donations? New good causes in the world than a single commit to make a donation within Evidence From a Sample of Non- person could possibly give to, mean- one week resulted in a 50 percent Profit Firms,” Journal of Public ing lines must be drawn somewhere. increase in donations over asking for Economics, 69(3), 1998, pp. 323–45. Then the question becomes: Can peo- a gift today. This provides support Return to Text

24 NBER Reporter • No. 1, March 2019 4 This conclusion can also be reached Fund-raising,” Journal of Political Charities,” NBER Working Paper No. by invoking the non-distribution con- Economy, 106(6), 1998, pp. 1186– 16372, September 2010, and Journal straint imposed on non-profits: Charity 213. The theory was tested and con- of Public Economics, 95(5–6), 2011, managers are accepting as part of their firmed in a field experiment by J. List pp. 334–43. compensation the personal satisfaction and D. Lucking-Reiley, “The Effects of Return to Text of doing good works, not simply engag- Seed Money and Refunds on Charitable 9 J. Andreoni and A. Payne, ing in onerous fundraising. Another way Giving: Experimental Evidence from a “Crowding-Out Charitable to motivate fundraising is to assume University Capital Campaign,” Journal Contributions in Canada: New that charities vary qualitatively — two of Political Economy, 110(1), 2002, Knowledge From the North,” NBER disaster relief organizations can be pp. 215–33. Working Paper No. 17635, December imperfect substitutes if one provides Return to Text 2011. mostly food aid while the other special- 6 J. Andreoni, “Leadership Giving in Return to Text izes in medical aid. Suppose donors Charitable Fund-Raising,” Journal of 10 J. Andreoni, A. Payne, and S. Smith, differ by how much they favor food over Public Economic Theory, 8(1), 2006, “Do Grants to Charities Crowd Out medical aid. If both charities approach pp. 1–22. For a related model, see L. Other Income? Evidence from the UK,” a donor, the donor will give to the char- Vesterlund, “The Informational Value NBER Working Paper No. 18998, ity that represents her preferences best. of Sequential Fundraising,” Journal April 2013, and Journal of Public If only one charity calls to ask for a of Public Economics, 87(3- Economics, 114, 2014, pp. 75–86. donation, then the donor will give, but 4), 2003, pp. 627–57, and tests Return to Text will give less the further the charity is of these ideas by A. Bracha, M. 11 J. Andreoni, J. Rao, and H. from the donor’s ideal quality. Thus, Menietti, and L. Vesterlund, “Seeds Trachtman, “Avoiding the Ask: A Field charities setting fundraising expendi- to Succeed: Sequential Giving to Experiment on Altruism, Empathy, tures face both extensive and intensive Public Projects,” Journal of Public and Charitable Giving,” NBER motivations for seeking donations. This Economics, 95(5-6), 2011, pp. 416– Working Paper No. 17648, December model is developed in J. Andreoni and 27. 2011, and Journal of Political A. Payne, “Do Government Grants to Return to Text Economy 125(3), 2017, pp. 625–53. Private Charities Crowd Out Giving or 7 Ibid, J. Andreoni and A. Payne, Return to Text Fund-raising?” American Economic American Economic Review, 2003. 12 J. Andreoni and M. Serra-Garcia, Review, 93(3), 2003, pp. 792–812. Return to Text “Time-Inconsistent Charitable Giving,” Return to Text 8 J. Andreoni and A. Payne, “Is NBER Working Paper No. 22824, 5 This result is due to J. Andreoni, Crowding Out Due Entirely to November 2016. “Toward a Theory of Charitable Fundraising? Evidence From a Panel of Return to Text

NBER Reporter • No. 1, March 2019 25 NBER News

Annual Report of Awards to NBER Affiliates

Katherine Baicker was elected to the American Risk and Insurance Economics from the American Society the American Academy of Arts and Association Early Career Scholarly of Health Economics. Sciences. Achievement Award, shared the Benjamin Moll received the Lucian Bebchuk received the 2018 ASHEcon Medal for Top Health Germán Bernácer Prize for best IRRC Institute Award for work in Economist Under 40 with Jonathan European economist under 40 work- corporate governance for his research Kolstad, and was awarded the National ing in macroeconomics and finance. He with Scott Hirst on “Index Funds and Institute for Health Care Management also received the Economics in Central the Future of Corporate Governance: Foundation’s Research Award for his Banking Award for his paper with Theory, Evidence, and Policy.” paper “What Does a Deductible Do?” Greg Kaplan and Gianluca Violante Roland J. Benabou was elected as a with Zarek Brot-Goldberg, Amitabh on “Monetary Policy According to corresponding member of the Académie Chandra, and Jonathan Kolstad. HANK.” des Sciences Morales et Politiques, David Hirshleifer served as pres- Ariel Pakes was elected to the one of the five national academies that ident elect of the American Finance National Academy of Sciences, and collectively constitute the Institut de Association and was elected a fellow of shared the BBVA Frontiers Knowledge France. the Finance Theory Group. Award in Economics, Finance and Eli Berman was selected as pres- Seema Jayachandran was awarded Management with Timothy Bresnahan ident of the Economics of National the 2018 Sustainability Science Award and Robert Porter. Security Association. of the Ecological Society of America Lubos Pastor received a Woodrow Howard Bodenhorn was awarded for her paper “Cash for Carbon” with Wilson Award from the Slovak Embassy the Jonathan Hughes Prize for Excellence Joost de Laat, Eric Lambin, Charlotte to the United States for contribu- in Teaching Economic History by the Stanton, Robin Audy, and Nancy tions to Slovakia-U.S. relations, and Economic History Association. Thomas. won the QMA Award for Best Paper John Y. Campbell received an hon- Supreet Kaur received a Sloan on Investment Management from the orary degree from the BI Norwegian Research Fellowship and a National Western Finance Association for “Fund Business School. Science Foundation CAREER award. Tradeoffs” with Robert Stambaugh and David Deming was awarded the Mervyn King was awarded the Paul Lucian Taylor. David N. Kershaw Prize, which honors Volcker Lifetime Achievement Award Parag Pathak received the John persons under the age of 40 who have by the National Association for Business Bates Clark Medal from the American made distinguished contributions to the Economics. Economic Association, honoring the field of public policy. Morris M. Kleiner received a American economist under the age of Amy Finkelstein was awarded Lifetime Achievement Award from 40 judged to have made the most signifi- a MacArthur Fellowship, which is the Labor and Employment Relations cant contribution to economic thought awarded to extraordinarily talented and Association. and knowledge. creative individuals, and was elected to Jin-Tan Liu was awarded a National Stephen Redding was awarded the the National Academy of Sciences. Chair Professorship by the Taiwan Econometric Society’s Frisch Medal for Price V. Fishback served as Ministry of Education. the best empirical paper published in president of the Economic History Matteo Maggiori received the Econometrica in the last five years for Association. AQR Insight Award for his paper “The Economics of Density: Evidence Gita Gopinath was elected a fel- “International Currencies and Capital from the Berlin Wall” with Gabriel low of both the American Academy of Allocation” with Brent Neiman and Ahlfeldt, Daniel Sturm and Nikolaus Arts and Sciences and the Econometric Jesse Schreger, and an Excellence Award Wolf. Society, and was appointed economic in Global Economic Affairs from the Carmen Reinhart was awarded the counsellor and director of the Research Kiel Institute for the World Economy. King Juan Carlos Prize in Economics, Department of the International Thomas McGuire received the which recognizes influential Spanish Monetary Fund. Victor P. Fuchs Award for Lifetime or Latin American economists; the Benjamin Handel received Contributions to the Field of Health Bernhard Harms Prize from the Kiel

26 NBER Reporter • No. 1, March 2019 Institute for the World Economy, for of Alfred Nobel. Prize for the best article published distinguished contributions in interna- Nancy L. Rose was elected to the in Social Science History, “The One tional economics; and the Adam Smith American Academy of Arts and Sciences Percent across Two Centuries: A Award from the National Association and named a distinguished fellow of the Replication of Thomas Piketty’s Data for Business Economics. Industrial Organization Society. on the Concentration of Wealth in the Hugh Rockoff became pres- Joshua L. Rosenbloom was elected United States.” ident-elect of the Economic History a fellow of the Cliometric Society. Robert Shiller won the Global Association. Maya Rossin-Slater received a Economy Prize from the Kiel Institute Dani Rodrik was elected a fellow National Science Foundation CAREER for World Economy. of the American Association for the award. Esteban Rossi-Hansberg won the Advancement of Science and received Raffaella Sadun received the HBR Robert E. Lucas, Jr. Prize for the most the John von Neumann Award from McKinsey Award for the best paper interesting paper in the last two years the Rajk László College for Advanced published in the Harvard Business in the Journal of Political Economy Studies in Budapest. Review for “Why Do We Undervalue for his paper on “The Geography of Christina Romer received the Competent Management?” with Development” with Klaus Desmet and 2018 Robert A. Muh Alumni Award Nicholas Bloom and John Van Reenen. Dávid Krisztián Nagy. in the Humanities, Arts, and Social Joseph Shapiro was named a Sloan Frank Schorfheide was elected a Sciences from MIT. Research Fellow by the Alfred P. Sloan fellow of the Econometric Society. Paul Romer and William Foundation. G. William Schwert was named Nordhaus shared the Sveriges Riksbank Richard Sutch received the Social a fellow of the American Finance Prize in Economic Sciences in Memory Science History Association Founder’s Association.

NBER Reporter • No. 1, March 2019 27 Conferences

Economics of Infrastructure

An NBER conference on the Economics of Infrastructure took place March 1 in Cambridge. Research Associates Edward L. Glaeser of Harvard University and James M. Poterba of MIT organized the meeting, which was sponsored by the Smith Richardson Foundation. These researchers’ papers were presented and discussed:

• Abhishek Nagaraj, University of California, Berkeley, “The Private Impact of Public Information: Landsat Satellite Maps and Gold Exploration”

* Shoshana Vasserman and Valentin Bolotnyy, Harvard University, “Scaling Auctions as Insurance: A Case Study in Infrastructure Procurement”

• Christoph Boehm, University of Texas at Austin, “Government Consumption and Investment: Does the Composition of Purchases Affect the Multiplier?”

• Pablo Fajgelbaum, University of California, Los Angeles and NBER, and Stephen J. Redding, Princeton University and NBER, “Trade, Structural Transformation, and Development: Evidence from Argentina 1869–1914”

• Daniel Leff Yaffe, University of California, San Diego, “The Interstate Multiplier”

• Aleksandar Andonov, University of Amsterdam; Roman Kräussl, University of Luxembourg; and Joshua Rauh, Stanford University and NBER, “The Subsidy to Infrastructure as an Asset Class” (NBER Working Paper No. 25045)

• Christopher Severen, Federal Reserve Bank of Philadelphia, “Commuting, Labor, and Housing Market Effects of Mass Transportation: Welfare and Identification”

• Nicolas Campos, Eduardo Engel, and Ronald Fischer, Universidad de Chile, and Alexander Galetovic, Stanford University, “Renegotiations and Corruption: The Odebrecht Case”

Summaries of these papers are at www.nber.org/conferences/2019/EIs19/summary.html

28 NBER Reporter • No. 1, March 2019 Transforming Rural Africa

An NBER conference on Transforming Rural Africa took place February 28–March 1 in Cambridge. Christopher B. Barrett of Cornell University, Abebe Shimeles and Hanan Morsy, both of the African Development Bank, and Research Associates Michael Carter of the University of California, Davis, Tavneet Suri of MIT, and Christopher R. Udry of Northwestern University, organized the meeting, which was sponsored by the African Development Bank. These researchers’ papers were presented and discussed:

• Fo Kodjo Dzinyefa Aflagah, International Food Policy Research Institute; Tanguy Bernard, International Food Policy Research Institute and Bordeaux IV; and Angelino Viceisza, Spelman College, “Cheap Talk and Coordination in the Lab and in the Field: Collective Commercialization in Senegal”

• Kibrom A. Abay, International Livestock Research Institute; Leah Bevis, Ohio State University; and Christopher B. Barrett, Cornell University, “Measurement Error Mechanisms Matter: Agricultural Intensification with Farmer Misperceptions and Misreporting”

• Tilman Graff, Busara Center for Behavioral Economics, “Spatial Inefficiencies in Africa’s Trade Network”

• Toyin Samuel Olowogbon and Raphael O. Babatunde, University of Ilorin, Nigeria, and Edward Asiedu, University of Ghana/University of Passau, “How Can Inclusive Agricultural Health Policy Intervention Promote Shared Agricultural Productivity in Nigeria? Evidence from a Randomized Control Trial”

• Denise Hörner and Meike Wollni, University of Goettingen; Adrien Bouguen, University of California, Berkeley; and Markus Frölich, University of Mannheim, “The Effects of Decentralized and Video-based Extension on the Adoption of Integrated Soil Fertility Management — Experimental Evidence from Ethiopia”

• Aminou Arouna, AfricaRice; Jeffrey D. Michler, University of Arizona; and Jourdain Lokossou, International Crops Research Institute for the Semi-Arid Tropics, “Contract Farming and Rural Transformation: Evidence from a Field Experiment in Benin”

• Ameet Morjaria and Richard Merton Peck, Northwestern University, “Entry, Growth, and Exit: Theory and Evidence from Ethiopia’s Cut-Flower Exporters”

• Rute Martins Caeiro, Nova School of Business and Economics, “From Learning to Doing: Diffusion of Agricultural Innovations in Guinea-Bissau”

• Esther Delesalle, Institute of Economic and Social Research, Université catholique de Louvain, “The Impact of the Universal Primary Education Program on Labor Market Outcomes: Evidence from Tanzania”

• Michael Carter, University of California, Davis and NBER; Rachid Laajaj, University of Los Andes; and Dean Yang, University of Michigan and NBER, “Temporary Subsidies and the Adoption of Green Revolution Technologies by Mozambican Farmers and Their Social Networks”

• Shilpa Aggarwal, Indian School of Business; Brian J. Giera, Amazon Research; Dahyeon Jeong and Alan Spearot, University of California, Santa Cruz; and Jonathan Robinson, University of California, Santa Cruz and NBER, “Market Access, Trade Costs, and Technology Adoption: Evidence from Northern Tanzania”

• Joshua Deutschmann and Emilia Tjernström, University of Wisconsin-Madison; Maya Duru, Abdul Latif Jameel Poverty Action Lab; and Kim Siegal, One Acre Fund, “Information, Credit, and Inputs: The Impacts and Mechanisms of a Program to Raise Smallholder Productivity”

Summaries of these papers are at www.nber.org/conferences/2019/TRAs19/summary.html

NBER Reporter • No. 1, March 2019 29 Big Data: Long-Term Implications for Financial Markets and Firms

An NBER conference on Big Data: Long-Term Implications for Financial Markets and Firms took place March 8 in Cambridge. Itay Goldstein of the University of Pennsylvania, Research Associate Chester S. Spatt of Carnegie Mellon University, and Faculty Research Fellow Mao Ye of the University of Illinois at Urbana-Champaign organized the meeting, which was supported by the National Science Foundation in conjunction with the The Review of Financial Studies. These researchers’ papers were presented and discussed:

• Zheng Tracy Ke, Harvard University; Bryan T. Kelly, Yale University and NBER; and Dacheng Xiu, University of Chicago, “Predicting Returns with Text Data”

• Amber Anand, Syracuse University; Mehrdad Samadi and Kumar Venkataraman, Southern Methodist University; Jonathan Sokobin, Financial Industry Regulatory Authority, “Institutional Order Handling and Broker-Affiliated Trading Venues”

• Michael Gofman, University of Rochester; Sajjad Jafri, Queen’s University; and James T. Chapman, Bank of Canada, “High-Frequency Analysis of Financial Stability”

• David Easley, Cornell University; Marcos López de Prado, AQR Capital Management; Maureen O’Hara, Cornell University; and Zhibai Zhang, NYU Tandon, “Microstructure in the Machine Age”

• Jura Liaukonyte, Cornell University, and Alminas Zaldokas, Hong Kong University of Science and Technology, “Background Noise? TV Advertising Affects Real Time Investor Behavior”

• Hedi Benamar and Clar a Veg a, Federal Reserve Board, and Thierry Foucault, HEC Paris School of Management, “Demand for Information, Uncertainty, and the Response of U.S. Treasury Securities to News”

• Robert P. Bartlett III, Richard Stanton, and Nancy Wallace, University of California, Berkeley, and Adair Morse, University of California, Berkeley and NBER, “Consumer-Lending Discrimination in the FinTech Era”

• Isil Erel, Ohio State University; Léa H. Stern, University of Washington; Chenhao Tan, University of Colorado Boulder; and Michael S. Weisbach, Ohio State University and NBER, “Selecting Directors Using Machine Learning” (NBER Working Paper No. 24435)

• Bo Cowgill, Columbia University, and Eric Zitzewitz, Dartmouth College and NBER, “Stock Compensation and Employee Attention”

Summaries of these papers are at www.nber.org/conferences/2019/BDFMs19/summary.html

30 NBER Reporter • No. 1, March 2019 Big Data for 21st Century Economic Statistics

An NBER conference on Big Data for 21st Century Economic Statistics met in Washington on March 15–16. Research Associates Katharine G. Abraham of the University of Maryland and Matthew D. Shapiro of the University of Michigan; Ron S. Jarmin of the U.S. Census Bureau; and Brian Moyer of the Bureau of Economic Analysis organized the meeting, which was sponsored by the Alfred P. Sloan Foundation. These researchers’ papers were presented and discussed:

• Carol Robbins, National Science Foundation; Jose Bayoan Santiago Calderon, Claremont Graduate University; Gizem Korkmaz, Daniel Chen, Sallie Keller, Aaron Schroeder, and Stephanie S. Shipp, University of Virginia; Claire Kelling, Pennsylvania State University, “The Scope and Impact of Open Source Software as Intangible Capital: A Framework for Measurement with an Application Based on the Use of R and Python Packages”

• Katharine G. Abraham, University of Maryland and NBER; Margaret Levenstein, University of Michigan; and Matthew D. Shapiro, University of Michigan and NBER, “Securing Commercial Data for Economic Statistics”

• W. Erwin Diewert, University of British Columbia and NBER, and Robert C. Feenstra, University of California, Davis and NBER, “Estimating the Benefits of New Products”

• David Copple, Bradley J. Speigner, and Arthur Turrell, Bank of England, “Transforming Naturally Occurring Text Data into Economic Statistics: The Case of Online Job Vacancy Postings”

• Edward L. Glaeser, Harvard University and NBER, and Hyunjin Kim and Michael Luca, Harvard University, “Nowcasting the Local Economy: Using Yelp Data to Measure Economic Activity” (NBER Working Paper No. 24010)

• Rishab Guha, Harvard University, and Serena Ng, Columbia University and NBER, “A Machine-Learning Analysis of Seasonal and Cyclical Sales in Weekly Scanner Data”

• Gabriel Ehrlich and David Johnson, University of Michigan; John C. Haltiwanger, University of Maryland and NBER; Ron S. Jarmin, U.S. Census Bureau; and Matthew D. Shapiro, University of Michigan and NBER, “Re-Engineering Key National Economic Indicators”

• Andrea Batch, Jeffrey C. Chen, Alexander Driessen, Abe Dunn, and Kyle K. Hood, Bureau of Economic Analysis, “Off to the Races: A Comparison of Machine Learning and Alternative Data for Predicting Economic Indicators”

• Tomaz Cajner, Leland D. Crane, Ryan Decker, Adrian Hamins-Puertolas, and Christopher Kurz, Federal Reserve Board, “Improving the Accuracy of Economic Measurement with Multiple Data Sources: The Case of Payroll Employment Data”

• J. Bradford Jensen, Georgetown University and NBER; Shawn D. Klimek, Andrew L. Baer, and Joseph Staudt, U.S. Census Bureau; and Lisa Singh and Yifang Wei, Georgetown University, “Automating Response Evaluation for Franchising Questions on the 2017 Economic Census”

• Sudip Bhattacharjee and Ugochukwu Etudo, University of Connecticut, and John Cuffe, Justin Smith, and Nevada Basdeo, U.S. Census Bureau, “Using Public Data to Generate Industrial Classification Codes”

• Jeremy Moulton, University of North Carolina, Chapel Hill, and Marina Gindelsky and Scott A. Wentland, Bureau of Economic Analysis, “Valuing Housing Services in the Era of Big Data: A User Cost Approach Leveraging Zillow Microdata”

• Shifrah Aron-Dine, Stanford University, and Aditya Aladangady, Wendy Dunn, Laura Feiveson, Paul Lengermann, and Claudia R. Sahm, Federal Reserve Board, “From Transactions Data to Economic Statistics: Constructing Real-Time, High-Frequency, Geographic Measures of Consumer Spending”

NBER Reporter • No. 1, March 2019 31 • David Friedman, Crystal G. Konny, and Brendan K. Williams, Bureau of Labor Statistics, “Big Data in the U.S. Consumer Price Index: Experiences & Plans”

• Don Fast and Susan Fleck, Bureau of Labor Statistics, “Measuring Export Price Movements with Administrative Trade Data”

• Rebecca J. Hutchinson, U.S. Census Bureau, “Investigating Alternative Data Sources to Reduce Respondent Burden in United States Census Bureau Retail Economic Data Products”

• Abe Dunn, Bureau of Economic Analysis; Dana Goldman and Neeraj Sood, University of Southern California and NBER; and John Romley, University of Southern California, “Quantifying Productivity Growth in Health Care Using Insurance Claims and Administrative Data”

Summaries of these papers are at www.nber.org/conferences/2019/CRIWs19/summary.html

Economics of Digitization

An NBER conference on the Economics of Digitization met at Stanford on March 22. Research Associate Shane Greenstein of Harvard University, Program Director Josh Lerner of Harvard Business School, and Research Associate Scott Stern of MIT organized the meeting, which was sponsored by the Alfred P. Sloan Foundation, Amazon, and The Tides Foundation. These researchers’ papers were presented and discussed:

• Alberto Cavallo, Harvard University and NBER, “More Amazon Effects: Online Competition and Pricing Behaviors” (NBER Working Paper No. 25138)

• Charles I. Jones and Christopher Tonetti, Stanford University and NBER, “Nonrivalry and the Economics of Data”

• John M. Barrios, University of Chicago; Yael Hochberg, Rice University and NBER; and Livia Hanyi Yi, Rice University, “The Cost of Convenience: Ridesharing and Traffic Fatalities”

• Abhishek Nagaraj, University of California, Berkeley, and Imke C. Reimers, Northeastern University, “Digitization and the Demand for Physical Works: Evidence from the Google Books Project”

• Timothy J. DeStefano, Organisation for Economic Co-operation and Development, and Richard Kneller and Jonathan D. Timmis, University of Nottingham, “Cloud Computing and Firm Growth”

• Shuang Wang, Boston University; Jacob LaRiviere, Microsoft; and Aadharsh Kannan, Amazon, “Spatial Competition and Missing Data: An Application to Cloud Computing”

• Ben T. Leyden, Cornell University, “There’s an App (Update) for That: Understanding Product Updating under Digitization”

• Hunt Allcott, New York University and NBER; Luca Braghieri and Sarah Eichmeyer, Stanford University; and Matthew Gentzkow, Stanford University and NBER, “The Welfare Effects of Social Media” (NBER Working Paper No. 25514)

Summaries of these papers are at www.nber.org/conferences/2019/EoDs19/summary.htm

32 NBER Reporter • No. 1, March 2019 Program and Working Group Meetings

Industrial Organization

Members of the NBER’s Industrial Organization Program met February 8–9 at Stanford. Research Associates Eric Budish of the University of Chicago and Jean-François Houde of the University of Wisconsin-Madison organized the meeting. These researchers’ papers were presented and discussed:

• Thomas G. Wollmann, University of Chicago and NBER, “How to Get Away with Merger: Stealth Consolidation and Its Effects on U.S. Healthcare”

• Sumit Agarwal, Georgetown University; John Grigsby, University of Chicago; Ali Hortaçsu, University of Chicago and NBER; Gregor Matvos, University of Texas at Austin and NBER; Amit Seru, Stanford University and NBER; and Vincent Yao, Georgia State University, “Searching for Approval”

• Panle Jia Barwick, Cornell University and NBER; Myrto Kalouptsidi, Harvard University and NBER; and Nahim B. Zahur, Cornell University, “China’s Industrial Policy: An Empirical Evaluation”

• Daniel Bjorkegren, Brown University, “Competition in Network Industries: Evidence from Mobile Telecommunications in Rwanda”

• Michael Ostrovsky, Stanford University and NBER, and Michael Schwarz, Microsoft, “Carpooling and the Economics of Self-Driving Cars” (NBER Working Paper No. 24349)

• Pietro Tebaldi and Alexander Torgovitsky, University of Chicago, and Hanbin V. Yang, Harvard University, “Nonparametric Estimates of Demand in the California Health Insurance Exchange”

• Gaston Illanes, Northwestern University, and Manisha Padi, University of Chicago, “Competition, Asymmetric Information, and the Annuity Puzzle: Evidence from a Government-run Exchange in Chile”

• Yuyu Chen, Peking University; Mitsuru Igami and Masayuki Sawada, Yale University; and Mo Xiao, University of Arizona, “Privatization and Productivity in China”

• Guangyu Cao, Peking University and Guanghua-ofo Center for Sharing Economy Research; Ginger Zhe Jin, University of Maryland and NBER; and Xi Weng and Li-An Zhou, Peking University, “Market Expanding or Market Stealing? Competition with Network Effects in Bike-Sharing” (NBER Working Paper No. 24938)

• Keaton S. Miller, University of Oregon; Amil Petrin, University of Minnesota and NBER; Robert Town, University of Texas at Austin and NBER; and Michael Chernew, Harvard University and NBER, “Optimal Managed Competition Subsidies”

• Rebecca Diamond and Petra Persson, Stanford University and NBER; Michael J. Dickstein, New York University and NBER; and Timothy McQuade, Stanford University, “Take-Up, Drop-Out, and Spending in ACA Marketplaces” (NBER Working Paper No. 24668)

• Mark L. Egan, Harvard University; Gregor Matvos, University of Texas at Austin and NBER; and Amit Seru, Stanford University and NBER, “Arbitration with Uninformed Consumers” (NBER Working Paper No. 25150)

Summaries of these papers are at www.nber.org/conferences/2019/IOs19/summary.html

NBER Reporter • No. 1, March 2019 33 Insurance

Members of the NBER’s Insurance Working Group met February 8–9 at Stanford. Research Associates Benjamin R. Handel of the University of California, Berkeley and Motohiro Yogo of Princeton University organized the meeting. These researchers’ papers were presented and discussed:

• Gaston Illanes, Northwestern University, and Manisha Padi, University of Chicago, “Competition, Asymmetric Information, and the Annuity Puzzle: Evidence from a Government-run Exchange in Chile”

• Radek Paluszynski, University of Houston, and Pei Cheng Yu, University of New South Wales, “Pay What Your Dad Paid: Commitment and Price Rigidity in the Market for Life Insurance”

• Robin Greenwood, Harvard University and NBER, and Annette Vissing-Jorgensen, University of California, Berkeley and NBER, “The Impact of Pensions and Insurance on Global Yield Curves”

• Andrew Ellul, Indiana University; Anastasia Kartasheva, Bank for International Settlements; Chotibhak Jotikasthira, Southern Methodist University; Christian Lundblad, University of North Carolina at Chapel Hill; and Wolf Wagner, Erasmus University, “Insurers as Asset Managers and Systemic Risk”

• Juan Pablo Atal, University of Pennsylvania; Hanming Fang, University of Pennsylvania and NBER; Martin Karlsson, University of Duisburg-Essen; and Nicolas R. Ziebarth, Cornell University and NBER, “Long-Term Health Insurance: Theory Meets Evidence”

• David Schoenherr, Princeton University; Janis Skrastins, Washington University in St. Louis; and Bernadus Doornik, Banco Central do Brasil, “Unemployment Insurance, Strategic Unemployment, and Firm-Worker Collusion”

• Michael Geruso, University of Texas at Austin and NBER; Timothy Layton and Mark Shepard, Harvard University and NBER, and Grace McCormack, Harvard University, “The Two Margin Problem in Insurance Markets”

• Keaton S. Miller, University of Oregon; Amil Petrin, University of Minnesota and NBER; Robert Town, University of Texas at Austin and NBER; and Michael Chernew, Harvard University and NBER, “Optimal Managed Competition Subsidies”

• Rebecca Diamond and Petra Persson, Stanford University and NBER; Michael J. Dickstein, New York University and NBER; and Timothy McQuade, Stanford University, “Take-Up, Drop-Out, and Spending in ACA Marketplaces” (NBER Working Paper No. 24668)

Summaries of these papers are at www.nber.org/conferences/2019/INSs19/summary.html

34 NBER Reporter • No. 1, March 2019 Law and Economics

Members of the NBER’s Law and Economics Program met February 15 in Cambridge. Program Director Christine Jolls of Yale University organized the meeting. These researchers’ papers were presented and discussed:

• Andrew Daughety and Jennifer Reinganum, Vanderbilt University, “Reducing Unjust Convictions: Plea Bargaining, Trial, and Evidence Suppression/Disclosure”

• Albert Choi, University of Virginia, and Kathryn E. Spier, Harvard University and NBER, “Class Actions and Private Antitrust Litigation”

• Martijn Cremers, University of Notre Dame; Scott Guernsey, University of Cambridge; and Simone M. Sepe, University of Arizona, “Directors’ Duties Laws and Long-Term Firm Value”

• Huseyin Gulen, Purdue University, and Brett W. Myers, Texas Tech University, “The Selective Enforcement of Government Regulation: Battleground States and the EPA”

• Alexander Dyck, University of Toronto; Adair Morse, University of California, Berkeley and NBER; and Paulo Martins Manoel, University of California, Berkeley, “Outraged by Compensation: Implications for Public Pension Performance”

• Lucian A. Bebchuk, Harvard University and NBER, and Doron Y. Levit, University of Pennsylvania, “Myopic Shareholders”

• Marcella Alsan, Stanford University and NBER, and Crystal Yang, Harvard University and NBER, “Fear and the Safety Net: Evidence from Secure Communities” (NBER Working Paper No. 24731)

Summaries of these papers are at www.nber.org/conferences/2019/LEs19/summary.html

Labor Studies

Members of the NBER’s Labor Studies Program met February 22 in San Francisco. Program Directors David Autor of MIT and Alexandre Mas of Princeton University organized the meeting. These researchers’ papers were presented and discussed:

• Conrad Miller, University of California, Berkeley and NBER; Jennifer Peck, Swarthmore College; and Mehmet Seflek, University of California, Berkeley, “Big Push Policies and Firm-Level Barriers to Employing Women: Evidence from Saudi Arabia”

• Randall Akee, University of California, Los Angeles and NBER, and Maggie R. Jones, U.S. Census Bureau, “Immigrants’ Earnings Growth and Return Migration from the U.S.: Examining Their Determinants Using Linked Survey and Administrative Data”

• François Gerard, Columbia University and NBER; Lorenzo Lagos, Columbia University; Edson R. Severnini, Carnegie Mellon University; and David Card, University of California, Berkeley and NBER, “Assortative Matching

NBER Reporter • No. 1, March 2019 35 or Exclusionary Hiring? The Impact of Firm Policies on Racial Wage Differences in Brazil”

• Shai Bernstein and Rebecca Diamond, Stanford University and NBER, and Timothy McQuade and Beatriz Pousada, Stanford University, “The Contribution of High-Skilled Immigrants to Innovation in the United States”

• David J. Deming, Harvard University and NBER, and Kadeem L. Noray, Harvard University, “STEM Careers and Technological Change” (NBER Working Paper No. 25065)

• Luigi Pistaferri, Stanford University and NBER, and Hamish Low, University of Cambridge, “Disability Insurance and Gender Differences: Evidence from Merged Survey-Administrative Data”

• Alisa Tazhitdinova, University of California, Santa Barbara, “Increasing Hours Worked: Moonlighting Responses to a Large Tax Reform”

• Brigham Frandsen and Emily C. Leslie, Brigham Young University, and Lars Lefgren, Brigham Young University and NBER, “Judging Judge Fixed Effects”

Summaries of these papers are at www.nber.org/conferences/2019/LSs19/summary.html

Economic Fluctuations and Growth subgroup

Members of the NBER’s Economic Fluctuations and Growth program’s subgroup on growth met February 28 in San Francisco. Martí Mestieri of Northwestern University and Faculty Research Fellow Christopher Tonetti of Stanford University organized the meeting. These researchers’ papers were presented and discussed:

• Matthew J. Delventhal, Claremont McKenna College; Jesús Fernández-Villaverde, University of Pennsylvania and NBER; and Nezih Guner, Center for Monetary and Financial Studies, “Demographic Transitions across Time and Space”

• Sebastian Heise, Federal Reserve Bank of New York, and Tommaso Porzio, University of California, San Diego, “Workers’ Home Bias and Spatial Wage Gaps: Lessons from the Enduring Divide between East and West Germany”

• Wyatt Brooks and Terence R. Johnson, University of Notre Dame, and Kevin Donovan, Yale University, “Bringing Data to the Model: Quantitative Implications of an Equilibrium Diffusion Model”

• Victor Couture, University of California, Berkeley; Cecile Gaubert, University of California, Berkeley and NBER; Jessie Handbury, University of Pennsylvania and NBER; and Erik Hurst, University of Chicago and NBER, “Income Growth and the Distributional Effects of Urban Spatial Sorting”

• Hugo Hopenhayn, University of California, Los Angeles and NBER, and Julian Neira and Rish Singhania, University of Exeter, “From Population Growth to Firm Demographics: Implications for Concentration, Entrepreneurship, and the Labor Share” (NBER Working Paper No. 25382)

• Jie Cai, Shanghai University of Finance and Economics; Nan Li, International Monetary Fund; and Ana Maria Santacreu, Federal Reserve Bank of St. Louis, “Knowledge Diffusion, Trade and Innovation across Countries and Sectors”

Summaries of these papers are at www.nber.org/conferences/2019/EGCw19/summary.html

36 NBER Reporter • No. 1, March 2019 Economic Fluctuations and Growth

Members of the NBER’s Economic Fluctuations and Growth Program met March 1 in San Francisco. Research Associates David Lagakos of the University of California, San Diego and Martin Schneider of Stanford University organized the meet- ing. These researchers’ papers were presented and discussed:

• Paolo Martellini, University of Pennsylvania, and Guido Menzio, New York University and NBER, “Declining Search Frictions, Unemployment and Growth” (NBER Working Paper No. 24518)

• Tarek Alexander Hassan, Boston University and NBER; Stephan Hollander, Tilburg University; Laurence van Lent, Frankfurt School of Finance and Management; and Ahmed Tahoun, London Business School, “Firm-Level Political Risk: Measurement and Effects” (NBER Working Paper No. 24029)

• Fernando E. Alvarez, University of Chicago and NBER, and Francesco Lippi, LUISS Guido Carli University and Einaudi Institute for Economics and Finance, “The Analytic Theory of a Monetary Shock”

• Pedro Bordalo, University of Oxford; Nicola Gennaioli, Bocconi University; Yueran Ma, University of Chicago; and Andrei Shleifer, Harvard University and NBER, “Over-Reaction in Macroeconomic Expectations” (NBER Working Paper No. 24932)

• Joel David, University of Southern California, and Venky Venkateswaran, New York University and NBER, “The Sources of Capital Misallocation” (NBER Working Paper No. 23129)

• Juan Morelli and Diego Perez, New York University, and Pablo Ottonello, University of Michigan, “Global Banks and Systemic Debt Crises”

Summaries of these papers are at www.nber.org/conferences/2019/EFGw19/summary.html

International Finance and Macroeconomics

Members of the NBER’s International Finance and Macroeconomics Program met March 8 in Cambridge. Faculty Research Fellow Cristina Arellano of the Federal Reserve Bank of Minneapolis, Research Associate Brent Neiman of the University of Chicago, and Program Director Pierre-Olivier Gourinchas of the University of California, Berkeley organized the meeting. These researchers’ papers were presented and discussed:

• Egemen Eren, Bank for International Settlements, and Semyon Malamud, Swiss Finance Institute, “Dominant Currency Debt”

• Javier Bianchi, Federal Reserve Bank of Minneapolis and NBER, and Jorge Mondragon, “Monetary Independence and Rollover Crises” (NBER Working Paper No. 25340)

• Kalina Manova, University College London; Antoine Berthou, Banque de France; Jong Hyun Chung, Stanford University; and Charlotte Sandoz, International Monetary Fund, “Productivity, (Mis)allocation, and Trade”

• Illenin Kondo, University of Notre Dame; Fabrizio Perri, Federal Reserve Bank of Minneapolis; and Sewon Hur, Federal Reserve Bank of Cleveland, “Real Interest Rates, Inflation, and Default”

NBER Reporter • No. 1, March 2019 37 • Juan Morelli and Diego Perez, New York University, and Pablo Ottonello, University of Michigan, “Global Banks and Systemic Debt Crises”

• Mishita Mehra, Grinnell College, “Skilled Immigration, Firms, and Policy”

• Rodrigo Barbone Gonzalez, Central Bank of Brazil, and Dmitry Khametshin, José-Luis Peydró, and Andrea Polo, Pompeu Fabra University, “Hedger of Last Resort: Evidence from Brazilian FX Interventions, Local Credit, and Global Financial Cycles”

Summaries of these papers are at www.nber.org/conferences/2019/IFMs19/summary.html

Monetary Economics

Members of the NBER’s Monetary Economics Program met March 8 at the Federal Reserve Bank of Chicago. Faculty Research Fellow David W. Berger of Northwestern University, Research Associate Giorgio Primiceri of Northwestern University, and Program Directors Emi Nakamura and Jón Steinsson, both of the University of California, Berkeley organized the meeting. These researchers’ papers were presented and discussed:

• Jasmine Xiao, University of Notre Dame, “Borrowing to Save and Investment Dynamics”

• Gauti B. Eggertsson, Brown University and NBER; Ella Wold, Brown University; Ragnar Juelsrud, BI Norwegian Business School; and Lawrence H. Summers, Harvard University and NBER, “Negative Nominal Interest Rates and the Bank Lending Channel”

• Pierpaolo Benigno, LUISS Guido Carli University, “Monetary Policy in a World of Cryptocurrencies”

• Gabriel Chodorow-Reich and Gita Gopinath, Harvard University and NBER (on leave); Prachi Mishra, Goldman Sachs; and Abhinav Narayanan, Reserve Bank of India, “Cash and the Economy: Evidence from India’s Demonetization” (NBER Working Paper No. 25370)

• Stefano Giglio, Yale University and NBER; Matteo Maggiori, Harvard University and NBER; Johannes Stroebel, New York University and NBER; and Stephen Utkus, Vanguard, “Five Facts about Beliefs and Portfolios”

• Carlo Altavilla, Frank Smets, and Miguel Boucinha, European Central Bank; and José-Luis Peydró, Pompeu Fabra University, “Banking Supervision, Monetary Policy and Risk-Taking: Big Data Evidence from 15 Credit Registers”

Summaries of these papers are at www.nber.org/conferences/2019/MEs19/summary.html

38 NBER Reporter • No. 1, March 2019 Productivity, Innovation, and Entrepreneurship

Members of the NBER’s Productivity, Innovation, and Entrepreneurship Program met March 15 in Cambridge. Program Directors Nicholas Bloom of Stanford University and Josh Lerner of Harvard Business School, Research Associate Serguey Braguinsky of the University of Maryland, and Faculty Research Fellow Sabrina T. Howell of New York University organized the meeting. These researchers’ papers were presented and discussed:

• Shai Bernstein and Rebecca Diamond, Stanford University and NBER, and Timothy McQuade and Beatriz Pousada, Stanford University, “The Contribution of High-Skilled Immigrants to Innovation in the United States”

• Greer K. Gosnell, London School of Economics; John A. List, University of Chicago and NBER; and Robert Metcalfe, Boston University, “The Impact of Management Practices on Employee Productivity: A Field Experiment with Airline Captains” (NBER Working Paper No. 25620)

• Laurent Fresard, University of Lugano and Sustainable Forestry Initiative, and Gerard Hoberg and Donald E. Bowen III, University of Maryland, “Technological Disruptiveness and the Evolution of IPOs and Sell-Outs”

• Timothy J. DeStefano, Organisation for Economic Co-operation and Development, and Richard Kneller and Jonathan D. Timmis, University of Nottingham, “Cloud Computing and Firm Growth”

• Nicolas Crouzet, Apoorv Gupta, and Filippo Mezzanotti, Northwestern University, “Shocks and Technology Adoption: Evidence from Electronic Payment Systems”

• Achyuta Adhvaryu, University of Michigan and NBER; Anant Nyshadham, Boston College and NBER; and Jorge A. Tamayo, Harvard University, “Managerial Quality and Productivity Dynamics”

• George P. Ball, Indiana University; Jeffrey Macher, Georgetown University; and Ariel Dora Stern, Harvard University, “Recalls, Innovation, and Competitor Response: Evidence from Medical Device Firms”

Summaries of these papers are at www.nber.org/conferences/2019/PRs19/summary.html

Environment and Energy Economics

Members of the NBER’s Environment and Energy Economics Program met March 14–15 at Stanford. Research Associates Kelsey Jack of the University of California, Santa Barbara and Ryan Kellogg of the University of Chicago organized the meet- ing. These researchers’ papers were presented and discussed:

• Fiona Burlig and Louis Preonas, University of Chicago, and Akshaya Jha, Carnegie Mellon University, “Out-of- Merit Costs and Blackouts: Evidence from the Indian Electricity Market”

• Peter Christensen and Ignacio Sarmiento, University of Illinois, and Christopher Timmins, Duke University and NBER, “Housing Discrimination and the Pollution Exposure Gap in the United States”

• Koichiro Ito, University of Chicago and NBER, and Shuang Zhang, University of Colorado Boulder, “Setting the Price Right: Evidence from Heating Price Reform in China”

• Derek Lemoine, University of Arizona and NBER, “Estimating the Consequences of Climate Change from Variation in Weather” (NBER Working Paper No. 25008)

NBER Reporter • No. 1, March 2019 39 • David Keiser, Iowa State University, and Joseph S. Shapiro, University of California, Berkeley and NBER, “Burning Waters to Crystal Springs? U.S. Water Pollution Regulation over the Last Half-Century”

• Cloe Garnache, University of Oslo, and Todd Guilfoos, University of Rhode Island, “The Effect of Salience on Risk Perceptions and Asset Prices”

• Jonathan I. Dingel, University of Chicago and NBER; Kyle C. Meng, University of California, Santa Barbara and NBER; and Solomon M. Hsiang, University of California, Berkeley and NBER, “Spatial Correlation, Trade, and Inequality: Evidence from the Global Climate” (NBER Working Paper No. 25447)

• James M. Sallee, University of California, Berkeley and NBER, “Pigou Creates Losers: On the Implausibility of Achieving Pareto Improvements from Pigouvian Taxation”

• Nicholas Ryan, Yale University and NBER, “Contract Enforcement and Productive Efficiency: Evidence from the Bidding and Renegotiation of Power Contracts in India” (NBER Working Paper No. 25547)

• Stephen P. Holland, University of North Carolina at Greensboro and NBER; Erin T. Mansur, Dartmouth College and NBER; Nicholas Muller, Carnegie Mellon University and NBER; and Andrew J. Yates, University of North Carolina at Chapel Hill, “Decompositions and Policy Consequences of an Extraordinary Decline in Air Pollution from Electricity Generation” (NBER Working Paper No. 25339)

• Panle Jia Barwick and Shanjun Li, Cornell University and NBER; Liguo Lin, Shanghai University of Finance and Economics; and Eric Zou, Cornell University, “The Value of Pollution Information: Evidence from China’s Air Quality Disclosure”

Summaries of these papers are at www.nber.org/conferences/2019/EEEs19/summary.html

Chinese Economy

Members of the NBER’s Chinese Economy Working Group met March 21–22 in Shanghai. Research Associates Hanming Fang of the University of Pennsylvania, Shang-Jin Wei of Columbia University, and Wei Xiong of Princeton University organized the meeting jointly with the Fanhai International School of Finance, Fudan University, and the School of Entrepreneurship and Management, ShanghaiTech University. These researchers’ papers were presented and discussed:

• J. Vernon Henderson, London School of Economics; Dongling Su, Boston University; Qinghua Zhang, Peking University; and Siqi Zheng, MIT, “Local Factor Market Distortions in China”

• Tao Chen, Nanyang Technological University; Yi Huang, The Graduate Institute, Geneva; and Chen Lin, University of Hong Kong, “Finance and Volatility”

• Shang-Jin Wei, and Chunliu Yang, Fudan University, “Do Internet Finance Platforms Mitigate Conflicts of Interest? The Case of Mutual Fund Investment”

• Kaiji Chen and Tong Xu, Emory University; Qing Wang, Southwest University of Finance and Economics; and Tao Zha, Emory University and NBER, “Aggregate and Distributional Impacts of Housing Policy: China’s Experiment”

• Ran Duchin, University of Washington; Zhenyu Gao, Chinese University of Hong Kong; and Haibing Shu, Shanghai Jiaotong University, “Involuntary Political Connections and Firm Outcomes”

40 NBER Reporter • No. 1, March 2019 • Yan Bai, University of Rochester and NBER; Keyu Jin, London School of Economics; and Dan Lu, University of Rochester, “Misallocation under Trade Liberalization”

• Guangwei Li, ShanghaiTech University, “The Role of R&D Offshoring in Knowledge Diffusion: New Evidence from China”

• Sumit Agarwal, Georgetown University; Yongheng Deng, University of Wisconsin-Madison; Quanlin Gu, Peking University; Jia He, Nankai University; and Wenlan Qian and Yuan Ren, National University of Singapore, “Mortgage Debt, Hand-to-Mouth Households, and Monetary Policy Transmission”

• Zheng Michael Song, Chinese University of Hong Kong; Duncan Thomas and Daniel Xu, Duke University and NBER; and Miaojun Wang, Zhejiang University, “The Rise of Modern Retail in China: An Anatomy of the Footwear Industry”

• Kevin Lim, University of Toronto; Daniel Trefler, University of Toronto and NBER; and Miaojie Yu, Peking University, “Trade and Innovation: The Role of Scale and Competition Effects”

• Bo Li, Tsinghua University, and Jacopo Ponticelli, Northwestern University, “Going Bankrupt in China”

• Ziying Fan and Hang Zhang, Shanghai University of Finance and Economics, and Xiaxin Wang, Fudan FISF and University of Michigan, “Understanding Misreporting: Responses to a Housing Transaction Tax Notch in China”

Summaries of these papers are at www.nber.org/conferences/2019/CEs19/summary.html

Development of the American Economy

Members of the NBER’s Development of the American Economy Program met March 23 in Cambridge. Program Directors Leah Platt Boustan of Princeton University and William J. Collins of Vanderbilt University organized the meeting. These researchers’ papers were presented and discussed: • Lisa D. Cook, Michigan State University and NBER, “A New National Lynching Data Set and New Explanations for Lynching Behavior in the United States from 1684 to 1983” • Prottoy Akbar and Sijie Li, University of Pittsburgh, and Allison Shertzer and Randall Walsh, University of Pittsburgh and NBER, “Racial Segregation in Housing Markets and the Erosion of Black Wealth” • Ariell Zimran, Vanderbilt University and NBER, “Transportation and Health in a Developing Country: The United States, 1820–1847” • Shawn E. Kantor, Florida State University and NBER, and Alexander T. Whalley, University of Calgary and NBER, “Space Race: Automation Innovation and Labor’s Share” • Andrew Goodman-Bacon, Vanderbilt University and NBER, and Lucie Schmidt, Williams College and NBER, “Federalizing Benefits: The Introduction of Supplemental Security Income and the Size of the Safety Net” • Joshua K. Hausman and Paul Rhode, University of Michigan and NBER; and Johannes Wieland, University of California, San Diego and NBER, “Farm Prices, Redistribution, and the Severity of the Early U.S. Great Depression” Summaries of these papers are at www.nber.org/conferences/2019/DAEs19/summary.html

NBER Reporter • No. 1, March 2019 41 NBER Books edited by ernst r. berndt, r. ernst by edited

Economic Dimensions of Personalizedbooks of related interest p.dana and goldman, Economic Dimensions of

ECONOMIC DIMENSIONS 2/4/19 The Economics of Poverty Traps john w. rowe ernst r. berndt is the Louis E. Seley Pro- response of multinational firms, the conse- Personalized and Precision fessor in Applied Economics and professor of ap- Edited by Christopher B. Barrett, Michael quences for firm-level dynamics of entry and OF PERSONALIZED AND plied economics at the MIT Sloan School of Man- R. Carter, and Jean-Paul Chavas exit, and the nature of comparative advantage Medicine and Precision Medicineagement and a research associate of the NBER. across countries. In contrast to much of the The research in this volume explores the hypoth- PRECISION MEDICINE edited by ernst r. berndt, dana p. existing research on immigration that focuses esis that poverty is self-reinforcing because the goldman, and john w. rowe dana p. goldman holds the Leonard D. on lower-skilled workers, this volume provides equilibrium behaviors of the poor perpetuate Keyline Schaeffer Director’s Chair and is a distinguished a clear view of today’s high-skilled immigration Edited by Ernst R. Berndt, low standards of living. Contributions explore professor of public policy, pharmacy, and eco- and employment environment. Personalized and precision medicine (PPM)—the the dynamic, complex processes by which Dana P. Goldman, and John W. Rowe nomics at the Sol Price School of Public Policy an nber conference report targeting of therapies according to an individual’s households accumulate assets and increase and USC School of Pharmacy at the University of genetic, environmental, or lifestyle characteris- their productivity and earnings potential, as Ernst R. Berndt, Dana P. Goldman,Southern California. He is also founding and director John W. Rowe, editorsU.S. Engineering in a Global Economy tics—is becoming an increasingly important ap- well as the conditions under which some in-

of the Schaeffer Center for Health Policy and Eco- proach in health care treatment and prevention. Berndt / Economic Dimensions dividuals, groups, and economies struggle to Edited by Richard B. Freeman and Hal Personalized of Dimensions Economic nomics and a research associate of the NBER. Salzman The advancement of PPM is a challenge in tra- escape poverty. Investigating the full range of ditional clinical, reimbursement, and regulatory phenomena that combine to generate poverty john w. rowe is the Julius B. Richmond Pro- Since the late 1950s, the engineering job market landscapes because it is costly to develop and traps—gleaned from behavioral, health, and fessor of Health Policy and Aging at the Columbia in the United States has been fraught with fears introduces a wide range of scientific, clinical, ethi-

resource economics as well as the sociology, and Precision Medicine Personalized and precisionUniversity med Mailman School- of Public whetherHealth. the benefits ofof a shortagePPM of engineering willskill and talent. be cal, and economic issues. The economic issues in- psychology, and environmental literatures— U.S. Engineering in a Global Economy brings clude how information on accurate diagnosis and chapters in this volume also present new evi- icine (PPM) — the targetingan nber conferenceof reportconfined to developed countriesclarity to issues of supply and demandor in willthis treatment success will be disseminated and who dence that highlights both the insights and the important market. Following a general overview will bear the cost; changes to physician training limits of a poverty trap lens. The framework of trends in the engineering labor market, the to incorporate genetics, probability and statistics, therapies according to an individ- diffuse tointroduced emerging in this volume provides a robust economiesvolume examines the educational pathwayswith of and economic considerations; whether the bene- platform for studying well-being dynamics in undergraduate engineers and their entry into the fits of PPM will be confined to developed countries ual’s genetic, environmental, or life- less developeddeveloping economies. health laborcare market, the impactsystems; of engineers working or will diffuse to emerging economies with less de- an nber conference report in firms on productivity and innovation, and veloped health care systems; the effects of patient

different dimensions of the changing engineer- heterogeneity on cost-effectiveness analysis; and 4-color style characteristics — is becoming an the effectsHigh-Skilled of Migrationpatient to the United heterogeneity ing labor market, from licensing to changes on opportunities for PPM’s growth beyond treatment States and Its Economic Consequences in demand and guest worker programs. The of acute illness, in particular for prevention and increasingly important approach in cost-effectivenessEdited by Gordon H. Hanson, analysis; William R. volume provides and insights on opporengineering ed- reversal of chronic conditions. Kerr, and Sarah Turner ucation, practice, and careers that can inform educational institutions, funding agencies, and This volume explores recent empirical appli-

No issue is more prominent on the current gloss health care treatment and prevention. tunities for PPM’s growthpolicy makers beyond about the challenges facingthe the cations of PPM as well as the intersection of the policy agenda than immigration. Today’s im- United States in developing its engineering scientific, clinical, and economic factors affecting migrants to the United States—who are of- workforce in the global economy. the development of PPM, including its effects The advancement of PPM is a challenge treatmentten of highly skilledacute and educated—contribute illness, in particular on the drug pipeline, on reimbursement of PPM an nber conference report substantially to the operation and growth of diagnostics and treatments, and on funding of in traditional clinical, reimbursement, to preventiona technology-driven and economy. By bringingreversal to- of chronic the requisite underlying research. gether scholarship from a range of economic the university of chicago press subdisciplines, this volume broadens our un- www.press.uchicago.edu and regulatory landscapes becauseFor information onit books ofis related interestconditions. or derstanding of the interaction between high- for a catalog of new publications, please visit skilled immigration, scientific innovation, and www.press.uchicago.edu. isbn-13: 978-0-226-61106-8 market adjustments. Contributors to the volume isbn-10: 0-226-61106-x costly to develop and introduces a wide This volume explores recent empir9 0 0 0 0 - Jacket illustration: Yulia Drozdova / Alamy Stock Vector examine the effects of high-skilled immigra- tion on innovation and productivity, the impact 9 7 8 0 2 2 6 6 1 1 0 6 8 range of scientific, clinical, ethical,Printed in the USA and ical applicationson overall inequality across skillof groups, PPM the as well as economic issues. The economic issues the intersection of the scientific, clin- Chicago include: how information on accurate ical, and economic factors affecting diagnosis and treatment success will the development of PPM, including be disseminated and who will bear the its effects on the drug pipeline, on cost; changes to physician training to reimbursement of PPM diagnostics and incorporate genetics, probability and treatments, and on funding of the req- statistics, and economic considerations; uisite underlying research.

For more information or to place an order, go to: https://press.uchicago.edu/ucp/books/book/chicago/E/bo38872102.html You may also contact the University of Chicago Press Distribution Center, at Telephone: 1-800-621-2736 Email: [email protected]

42 NBER Reporter • No. 1, March 2019 The Economics of Artificial Intelligence: An Agenda

Ajay Agrawal, Joshua Gans, and Avi Golfarb, editors

Advances in artificial intelli- research is conducted. It explores the gence (AI) highlight the potential of economic influence of machine learn- this technology to affect productiv- ing, the branch of computational sta- ity, growth, inequality, market power, tistics that has driven much of the innovation, and employment. This recent excitement around AI, the eco- volume seeks to set the agenda for eco- nomic impact of robotics and auto- nomic research on the impact of AI. mation, and the potential economic It covers four broad themes: AI as a consequences of a still-hypothetical general purpose technology; the rela- artificial general intelligence. The vol- tionships between AI, growth, jobs, ume provides frameworks for under- and inequality; regulatory responses standing the economic impact of to changes brought on by AI; and the AI, and identifies a number of open effects of AI on the way economic research questions.

For more information or to place an order, go to: https://press.uchicago.edu/ucp/books/book/chicago/E/bo35780726.html You may also contact the University of Chicago Press Distribution Center, at Telephone: 1-800-621-2736 Email: [email protected]

NBER Reporter • No. 1, March 2019 43 Nonprofit Org. NBER U.S. Postage Reporter PAID NATIONAL BUREAU OF ECONOMIC RESEARCH National Bureau of Economic Research 1050 Massachusetts Avenue Cambridge, Massachusetts 02138-5398 (617) 868-3900

Change Service Requested