Economics of education - EconStor by jianghongl



Der Open-Access-Publikationsserver der ZBW – Leibniz-Informationszentrum Wirtschaft
The Open Access Publication Server of the ZBW – Leibniz Information Centre for Economics

   Hoxby, Caroline M.

   Economics of education

   NBER Reporter Online

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

   Suggested Citation: Hoxby, Caroline M. (2011) : Economics of education, NBER Reporter
   Online, National Bureau of Economic Research (NBER), Cambridge, Mass., Iss. 1, pp. 1-5,

   Nutzungsbedingungen:                                                 Terms of use:
   Die ZBW räumt Ihnen als Nutzerin/Nutzer das unentgeltliche,          The ZBW grants you, the user, the non-exclusive right to use
   räumlich unbeschränkte und zeitlich auf die Dauer des Schutzrechts   the selected work free of charge, territorially unrestricted and
   beschränkte einfache Recht ein, das ausgewählte Werk im Rahmen       within the time limit of the term of the property rights according
   der unter                                                            to the terms specified at
   →                  →
   nachzulesenden vollständigen Nutzungsbedingungen zu                  By the first use of the selected work the user agrees and
   vervielfältigen, mit denen die Nutzerin/der Nutzer sich durch die    declares to comply with these terms of use.
   erste Nutzung einverstanden erklärt.

                 Leibniz-Informationszentrum Wirtschaft
                 Leibniz Information Centre for Economics
                                               NATIONAL BUREAU OF ECONOMIC RESEARCH

Reporter OnLine at:                                                                  2011 Number 1

                                               Program Report
                           IN THIS ISSUE

                            Program Report     Economics of Education
                 Economics of Education 1

                          Research Summaries   Caroline M. Hoxby*
      Urban Growth and Climate Change 6
       Reducing the Risks of Catastrophes 9	
Trade Agreements as Incomplete Contracts 12         The Economics of Education Program is both exciting and productive,
The Impact of Employee Pension Promises 15     currently adding new Working Papers at the rate of 7.5 per month — a 50
                                               percent increase from the rate at the time of my last program report in fall
                        NBER Profiles    18    2006. The number of papers submitted for a typical Program Meeting is
                          Conferences    20    often ten times the number of available slots, and attendance at those meet-
                          NBER News      21    ings is high.
   Program and Working Group Meetings    25         I am particularly proud of three aspects of the Program. The first is
                         Bureau Books    30    the quality of the research being produced and the methods used by mem-
                                               bers, including some of the latest, most rigorous methods in applied micro-
                                               econometrics. The second is the fact that members use some of the richest,
                                               most comprehensive datasets in economics — many of these datasets were
                                               initially compiled by schools or school-related organizations, and program
                                               members deserve enormous credit for their resourcefulness in making them
                                               useful for economic research by establishing strong, collegial relationships
                                               with data providers, convincing schools to conduct randomized and other
                                               policy experiments, matching data from diverse sources, and themselves
                                               surveying or testing people when data otherwise would be missing. Third,
                                               program members produce research that is policy relevant, credible to
                                               policymakers, and grounded in economic logic.
                                                    The NBER’s Higher Education Working Group was integrated into
                                               the Economics of Education Program in 2009. We made the integration
                                               an occasion to celebrate the leadership of Charles T. Clotfelter, director of
                                               that working group, who oversaw an immense improvement in the quality
                                               of research on the economics of higher education. Although the practical
                                               policy questions differ across the two levels of education, all of the meth-
                                               ods, much of the data, and much of the deep economic logic are shared.

                                               *Hoxby is the Director of the NBER’s Program on Economics of Education
                                               and the Scott and Donya Bommer Professor of Economics at Stanford
                                               University. The numbers in parends throughout this report refer to NBER
                                               Working Papers. A complete list of NBER Education Working Papers can be
                                               found at:

                                                                                             NBER Reporter • 2011 Number 1
                                                                                         Areas of Continuing Interest
         NBER Reporter                                                                   and New Interest
                                                                                               In my last review, I focused on three areas in
                                                                                         which research was advancing particularly rap-
                                                                                         idly: the analysis of peer effects; the estimation
                                                                                         of teachers’ effects on achievement; and mak-
    The National Bureau of Economic Research is a private, nonprofit research orga-
    nization founded in 1920 and devoted to objective quantitative analysis of the       ing sense of students’ college choices (not just
    American economy. Its officers and board of directors are:                           whether to attend college in the first place, but
    President and Chief Executive Officer — James M. Poterba                             which schools to attend and whether to persist
    Controller — Kelly Horak                                                             at each school). These three areas continue to be
                                                                                         highly productive. For instance, Elias Bruegmann
                                                                                         and C. Kirabo Jackson (15202) demonstrate that,
    Chairman — John S. Clarkeson                                                         when a teacher whose own effect on achievement
    Vice Chairman — Kathleen B. Cooper
    Treasurer — Robert Mednick                                                           is strongly positive moves into a new school, her
                                                                                         new colleagues improve. They further show that
    DIRECTORS AT LARGE                                                                   the colleagues’ improved ability to raise achieve-
    Peter Aldrich              Jessica P. Einhorn          Michael H. Moskow             ment is attributable to their changing, not merely
    Elizabeth E. Bailey        Mohamed El-Erian            Alicia H. Munnell             to selection. That is, incumbent teachers in the
    Richard Berner             Jacob A. Frenkel            Robert T. Parry
    John Herron Biggs          Judith M. Gueron            James M. Poterba              new school raise their performance. For another
    John S. Clarkeson          Robert S. Hamada            John S. Reed                  example, we now have substantial evidence on
    Don R. Conlan              Peter Blair Henry           Marina v. N. Whitman          what happens to a student who goes to a school
    Kathleen B. Cooper         Karen N. Horn               Martin B. Zimmerman
    Charles H. Dallara         John Lipsky                                               where other students are high-achieving: his own
    George C. Eads             Laurence H. Meyer                                         achievement rises. This evidence relies on regres-
                                                                                         sion discontinuity methods, that is, on compar-
                                                                                         ing the later achievement of students who are just
    George Akerlof, California, Berkeley     Mark Grinblatt, California, Los Angeles     above and just below some admissions thresh-
    Jagdish W. Bhagwati, Columbia            Marjorie B. McElroy, Duke
    Glen G. Cain, Wisconsin                  Joel Mokyr, Northwestern                    old, where the threshold is not known to stu-
    Alan V. Deardorff, Michigan              Andrew Postlewaite, Pennsylvania            dents when they apply. Christian Pop-Eleches
    Ray C. Fair, Yale                        Uwe E. Reinhardt, Princeton                 and Miguel Urquiola (16886) study this situation
    Franklin Fisher, MIT                     Craig Swan, Minnesota
    John P. Gould, Chicago                   David B. Yoffie, Harvard                    in Romania; Damon Clark (“Elite Schools and
                                                                                         Academic Performance”, presented at the spring
    DIRECTORS BY APPOINTMENT OF OTHER ORGANIZATIONS                                      2007 Program Meeting) studies this situation in
    Jean Paul Chavas, Agricultural and Applied Economics Association                     England; and C. Kirabo Jackson (16598) studies
    Martin Gruber, American Finance Association                                          this situation in Trinidad and Tobago. Turning to
    Ellen Hughes-Cromwick, National Association for Business Economics                   college-going behavior, some of the most inter-
    Arthur B. Kennickell, American Statistical Association
    Thea Lee, American Federation of Labor and                                           esting new research provides rigorous evidence
                   Congress of Industrial Organizations                                  on how students respond to scholarships and
    William W. Lewis, Committee for Economic Development                                 other financial aid designed to improve their col-
    Robert Mednick, American Institute of Certified Public Accountants
    Alan L. Olmstead, Economic History Association                                       lege outcomes. Aimee Chin and Chinhui Juhn
    John J. Siegfried, American Economic Association                                     (15932) show that allowing undocumented stu-
    Gregor W. Smith, Canadian Economics Association                                      dents to pay in-state tuition (usually just one-third
    Bart van Ark, The Conference Board
                                                                                         to one-half of out-of-state tuition) has no statis-
    The NBER depends on funding from individuals, corporations, and private foun-        tically significant effect on their college atten-
    dations to maintain its independence and its flexibility in choosing its research
    activities. Inquiries concerning contributions may be addressed to James M.          dance. Stephens Desjardins and Brian McCall
    Poterba, President & CEO, NBER 1050 Massachusetts Avenue, Cambridge, MA              (“The Impact of the Gates Millenium Scholars
    02138-5398. All contributions to the NBER are tax deductible.                        Program”, presented at the spring 2008 Program
    The Reporter is issued for informational purposes and has not been reviewed by       Meeting) show that Gates Scholarships very mod-
    the Board of Directors of the NBER. It is not copyrighted and can be freely repro-   estly improved persistence among the low-income
    duced with appropriate attribution of source. Please provide the NBER’s Public
    Information Department with copies of anything reproduced.                           minority students eligible for them.
                                                                                               Since my last report, several new themes
    Requests for subscriptions, changes of address, and cancellations should be sent     also have emerged in Economics of Education
    to Reporter, National Bureau of Economic Research, Inc., 1050 Massachusetts
    Avenue, Cambridge, MA 02138-5398. Please include the current mailing label.          research. Two notable ones are the importance
                                                                                         of information and the role of incentives for stu-

2   NBER Reporter • 2011 Number 1
dents, teachers, and schools. Because any         Oreopoulos, and Lisa Sanbonmatsu                 when viewed against the background of
program review is necessarily selective, I        (15361) designed an experiment in coordi-        family income or the potential returns to
focus here mainly on illustrating these new       nation with the tax preparer H&R Block.          college attendance. Yet, the policy change
themes.                                           Some families with college-aged children         caused about 40 percent of students to
                                                  were randomly assigned to be given infor-        send their scores to an additional school.
The Importance of Information                     mation on their child’s eligibility for gov-     This generated some additional informa-
                                                  ernment-based financial aid and on local         tion for students because, when a student
     Much of the existing research on edu-        college-going options. Some families also        who is a plausible applicant sends his scores
cation concerns the change in some con-           were randomly assigned to receive help           to a school, that school responds with bro-
crete resource: a salary increase for teach-      in filing the federal application for finan-     chures and other materials describing its
ers; a reduction in class size; a scholarship     cial aid (“FAFSA”). The results, which           offerings. It is striking that such a modest
or other financial aid for students; the          are highly credible owing to the random-         change in information produced such siz-
extension of compulsory schooling; or             ized design, suggest that the intervention       able effects on behavior.
the opening of a program. Although such           that combined information and FAFSA                   Avery and Turner (“Playing the
resource changes often can be shown to            help actually caused people to be 25 to 30       College Application Game”, presented at
change educational outcomes, their effects        percent more likely to enroll in college.        the fall 2009 Program Meeting) and Avery
typically are much smaller than proponents        These effects are dramatic in size for such a    and Hoxby (“The Missing One-Offs”, pre-
believed they would be. Also, two students        modest intervention — one that, if imple-        sented at the 2010 Summer Institute) dem-
with similar prior achievement often react        mented routinely, would cost only a few          onstrate that low-income students apply to
to resources in very different ways. For          dollars per family.                              fewer and less selective colleges than their
instance, although making financial aid                Todd Stinebrickner and Ralph                more affluent counterparts who have the
more generous causes some students to             Stinebrickner (14810) investigate whether        same test scores and achievement in high
attend college or to persist longer in college,   students learn about their academic ability      school. This fact holds even for low-income
a good share of students do not respond.          in college and make decisions about per-         students whose achievement is so high
Frustratingly for researchers, the students       sisting in a logical way, based on that infor-   that they qualify for free tuition and liv-
who do not respond often look very simi-          mation. To study this question, they com-        ing expenses at the most selective colleges
lar to the students who do. (On this point,       bine rich administrative data from Berea         in the United States. The authors of these
see for instance the Desjardins and McCall        College with data from surveys they con-         papers assemble an array of evidence that
paper mentioned above.) Put another way,          ducted themselves. Thus, they are able to        indicates that low-income students lack
researchers have been unable to show that         observe not just students’ academic behav-       information about college-going. While it
policymakers could control and improve            ior, such as their course-taking patterns        is hard to argue that these students do not
most people’s educational outcomes simply         and the grades they earn, but also stu-          have access to materials (since most col-
by controlling policies that are concerned        dents’ beliefs about their academic aptitude     leges’ materials are readily available online),
with educational resources.                       and expectations about college completion.       they have few contacts with people who
     Responding to the weak explanatory           The authors show that students enter col-        attended selective colleges. They are fre-
power of resource-type policies, research-        lege with beliefs about their academic abil-     quently too isolated geographically to find
ers increasingly have wondered whether            ity that are both optimistic and diffuse.        a critical mass of college-going peers or
differences in students’ and families’ infor-     Moreover, the students update their beliefs      advisors. In fact, the latter paper shows that
mation can account for variation in edu-          in the manner predicted by the Bayesian          it would not even make sense for selective
cational outcomes. Recent findings from           learning model. Students’ learning about         colleges’ staff to visit the schools or cities of
behavioral economics, which often show            their own aptitude explains much of their        most low-income, high-achieving students:
that apparently small differences in the          decision to drop out of college.                 they are simply too isolated for the benefits
content or framing of information can                  Amanda Pallais (“Why Not Apply?”            of such visits to outweigh the costs. The
have large effects, have only intensified         presented at the spring 2008 Program             bottom line is that information interven-
education researchers’ focus on informa-          Meeting) shows that an apparently tiny           tions might be warranted, but they may
tion. There are practical reasons to focus on     change in ACT policy produced a 20 per-          prove hard to design — see Avery (16359).
information as well: information interven-        cent increase in students’ applications to            Informational differences among stu-
tions tend to be very inexpensive compared        colleges. The change was that ACT, one           dents are also important in primary and
to resource-type interventions (so that even      of the two college aptitude testing orga-        secondary education. Parag Pathak and
modest benefits may outweigh costs) and           nization in the United States, gave stu-         Tayfun Sönmez (16783; also “Leveling
often have positive spillovers (useful infor-     dents four free score reports instead of         the Playing Field,” 2008 Summer Institute)
mation given to one person tends to spread        three. Because an additional score report        show that school choice mechanisms
to other people).                                 cost only $6 before and after the pol-           that are susceptible to strategic manipula-
     Eric Bettinger, Bridget Long, Philip         icy change, the intervention was negligible      tion tend to generate better outcomes for

                                                                                                           NBER Reporter • 2011 Number 1          3
families who are more informed. That is,        private schools. The authors interpret these     Incentives for Students,
although all students have the same oppor-      results as showing that report cards gen-        Teachers, and Schools
tunities under these mechanisms, students       erate competitive pressure on schools to
who understand how the mechanisms work          increase price-adjusted quality.                      Even though improving incentives is
and which schools are in demand end up                Jonah Rockoff, Douglas Staiger,            often more expensive than improving infor-
enrolling in schools that are higher in their   Thomas Kane, and Eric Taylor (16240)             mation, incentive-type interventions are
preference rankings. These better informed      study another informational intervention         often much less expensive than resource-
students disproportionately have parents        that appears small yet had big effects. They     type interventions, especially when their
who are affluent and educated. Thus, supe-      evaluate the effect of a program in which        relative efficacy is taken into account. This
rior information is one reason why stu-         New York City school principals were pro-        is shown by an array of recent research
dents’ outcomes are correlated with their       vided with estimates of how much each            done by program members.
family’s socioeconomic circumstances.           of their teachers had raised students’ test           Joshua Angrist, Daniel Lang, and
     Abigail Wozniak and Ofer Malamud           scores. Principals were randomly assigned        Philip Oreopoulos (12790) and Joshua
(16463) explore another reason why stu-         to this program, so the study’s findings         Angrist, Philip Oreopoulos, and Tyler
dents from more educated families have          are highly credible. The authors show that       Williams (16643) explore incentives
better outcomes. They investigate the           principals update their beliefs about teach-     for students to improve their grades in a
long-standing hypothesis that more edu-         ers’ effects in accordance with the Bayesian     Canadian university. In the former paper,
cated people respond more elastically to        learning model: for instance, principals         they study students who are randomly
changes in opportunities. (Theodore W.          update their beliefs more when the esti-         assigned to receive a merit scholarship if
Schultz often is credited with originating      mates provided to them are more precise          they maintain solid grades. In the second
this idea. See Bowman, 1980, cited in end-      and their own prior opinions are less pre-       paper, they study students who are ran-
note.) Specifically, Wozniak and Malamud        cise. More importantly, principals are like-     domly assigned to receive cash for better
investigate people who were induced to          lier to retain their effective teachers (and     grades: $100 for each grade of 70 or better
attend college because they had a higher        not retain their ineffective ones) when they     and an additional $20 for each percentage
risk of being drafted for the Vietnam War.      are provided with the estimated teacher          point above 70 percent. They find that the
They use draft induction risk as an instru-     effects. The change in the sensitivity of        merit scholarship improved the grades and
ment for attending and graduating from          retention to performance improves stu-           persistence of female students, though not
college, and they show that college educa-      dent achievement by a statistically signif-      of males. Interestingly, they also find that
tion makes a person more likely to subse-       icant though small amount. Here, it is           the availability of the merit scholarship
quently choose his labor market experience      worthwhile to remember the cost-bene-            caused female students to seek out more
based on expected earnings, as opposed to       fit ratios typical of information interven-      help with their courses: they were more
the market’s mere proximity to his place of     tions: although the change in achievement        likely to take advantage of supplemental
origin.                                         is small, the cost of the intervention is very   instructional services. In the latter paper,
     School report cards — simple reports       small on an ongoing basis.                       the authors find that the cash rewards
that describe students’ achievement in                Finally, Eric Taylor and John Tyler        improved males’ achievement, though not
absolute terms and relative to other local      (16877) examine a highly reputed teacher         females! The effects on males are modest
schools — are very inexpensive to provide.      evaluation system and find that it improves      overall, but larger for males who under-
Asim Khwaja, Tahir Andrabi, and Jishnu          teachers’ performance, as measured by their      stood the function linking performance
Das (“Report Cards,” spring 2009 Program        effects on student achievement. While the        to rewards.
Meeting) arranged to provide reports in         cost-benefit ratio of the program they study          Judith Scott-Clayton (“On Money
112 randomly selected educational markets       is not as impressive as the results of the       and Motivation”, fall 2008 program meet-
in Pakistan. The intervention was purely        information program in New York City             ing) studies a West Virginia incentive
informational: no explicit rewards or pun-      (16240), the improvement that Taylor and         scheme for college students. The program
ishments were included. The authors find        Tyler see is entirely within teacher. As a       offered free tuition to students who main-
that the report cards improved learning         rule, it has been hard for researchers to        tained a certain minimum course load
by 0.10 standard deviations and increased       produce credible evidence that teachers          and minimum GPA (2.75 in the fresh-
enrollment slightly. Private schools that       improve simply through being evaluated           men year, 3.0 thereafter). Since students
were initially bad — those with below           and then informed about how to improve           were not randomly assigned to the pro-
median scores at baseline — improved            their instruction. Even if such evaluation       gram, Scott-Clayton exploits differences
especially strongly: learning gains were        systems are an expensive means of improv-        in the timing of implementation and dis-
0.34 standard deviations. Private schools       ing achievement relative to some of the          continuities in the eligibility formula to
that were initially good did not improve        informational interventions described            generate credible estimates. Not only does
learning but did cut their fees. Government     above, they remain inexpensive relative to       she find substantial effects on achieve-
schools were somewhat less responsive than      most resource-type interventions.                ment, she also finds that the effects are

4   NBER Reporter • 2011 Number 1
highly concentrated around the thresh-            times what we would predict if the dis-        lyzed by David Figlio and Cassandra Hart
olds for annual scholarship renewal, indi-        trict had spent the same amount on class       (16056) and by Winnie Chan and Robert
cating that the program’s effects come via        size reduction. (The class size comparison     McMillan (“School Choice and Public
the incentives it provides, not simply via        is based on Project Star, which generates      School Performance, fall 2009 program
relaxing financial constraints.                   some of the highest credibly estimated         meeting). Although the authors investi-
     C. Kirabo Jackson (15722) studies            effects of class size reduction.)              gate programs in different locations — Fi-
incentives for students and teachers based             Karthik Muralidharan and Venkatesh        glio and Hart analyze a Florida corpo-
on Advanced Placement (AP) scores.                Sundararaman (15323) investigate per-          rate tax credit program and Chan and
The program he analyzes (“APIP”) pays             formance pay for teachers, using a pro-        McMillan analyze a tax credit program
high school students and their teachers           gram in India that they themselves largely     in Ontario — both teams of authors
between $100 and $500 per score of three          designed. Hundreds of schools were ran-        exploit variation in pressure on public
or above on an AP exam. To give a sense           domly assigned to have their teachers          schools that arises through pre-existing
of magnitude of rewards that a person             receive higher pay for higher students’        differences in the local availability of pri-
could earn, the maximum that a teacher            scores. Hundreds of schools were assigned      vate schools. Both teams find that pub-
has ever earned in one year is $11,500,           to an alternative treatment that gave them     lic schools respond to the potential loss
and the maximum that a student has ever           additional resources equal to the value of     of students to private schools by raising
earned in high school is $1,400. Because          the performance pay. Muralidharan and          their students’ achievement. Neither team
the program is not randomly assigned to           Sundararaman find that students in incen-      of authors finds evidence that differen-
schools, Jackson has to use a detrended dif-      tive schools improved their performance        tial student sorting (poor students dispro-
ference-in-differences strategy: essentially,     by 0.28 and 0.16 standard deviations in        portionately leaving the public schools)
the achievement trends of schools that            math and language tests, relative to con-      accounts for the improvement.
adopted the program earlier are compared          trol scores. Students scored significantly
to the achievement trends of schools that         higher on “conceptual” as well as “mechan-     Summing Up
adopted it later. Because the program’s           ical” components of the tests and also per-
sponsors were not able to roll out the            formed better on subjects for which no               New themes emerge in research
program in a single year to every school          incentives were given. These results sug-      because researchers find themselves con-
interested in adoption, the late adopters         gest that the students’ gains in achieve-      vinced by previous studies that some ques-
are fairly idiosyncratically selected from        ment were authentic, not mere “teaching        tions remain answered, thereby exposing
among schools who applied. Thus, the              to the test.” The gains in schools that sim-   other questions as likely to be important.
results are quite credible. Jackson finds         ply received the extra resources were one-     Thus, I think that it is a measure of the suc-
that students who participate in the pro-         third to one-half as large as the incentive-   cess of the NBER’s Economics of Education
gram are more likely to attend college            driven gains.                                  Program that, although some recent
and persist in college beyond their fresh-             Several authors have examined what        research extends and elaborates themes
man year. In addition, Black and Hispanic         happens when schools face incentives.          I identified previously, I did not predict
students are more likely to graduate from         For instance, Jonah Rockoff and Lesley         the themes of much recent research in
college.                                          Turner (14564) and Hanley Chiang               my previous program review. In particu-
     Eric Bettinger (16333) examines cash         (“Accountability Pressure on Failing           lar, it is encouraging that so much current
incentives for students funded by a philan-       Schools,” fall 2008 Program Meeting)           research focuses on issues like informa-
thropist in Coshocton, a poor city in the         use regression discontinuity methods to        tion and incentives that economists have
Appalachian area of Ohio. Schools and             show that schools that “just fail” accord-     long regarded as important. That informa-
grades in the city were randomly assigned         ing to their state’s accountability program    tion and incentive-type interventions also
to have their students get rewards of up          raise their students’ achievement more         tend to have propitious cost-benefit ratios
$75 per year for “proficient” scores and          than schools that “just pass.” In these two    is a bonus. Finally, it is important that
$100 per year for “advanced” scores on            studies, failing schools faced several pos-    NBER researchers continue to pioneer
Ohio’s statewide exams. Bettinger finds           sible consequences: students could trans-      rigorous methodological designs and cre-
that the incentives improve math scores by        fer out, principals could lose their jobs,     ate good data that allow them to analyze
0.15 standard deviations but he does not          and schools could be closed completely         such interventions.
find similar effects on other subject exams.      (though this was rare). Since the pass-
The Coshocton program was highly bene-            ing thresholds were unknown to schools
ficial relative to its costs: the program costs   in advance, the regression discontinuity       M.J. Bowman, “Theodore W. Schultz’s
were only fifteen hundredths of 1 percent         designs produce convincing results.            Contributions to Economics,” The
(0.15 percent) of the district’s per-pupil             A very different source of school         Scandinavian Journal of Economics, Vol.
expenditures. The effects of this inexpen-        incentives — competitive pressures gener-      82, No. 1 (1980), pp. 80–107
sive program on achievement were 250              ated by private school vouchers — is ana-

                                                                                                        NBER Reporter • 2011 Number 1        5

To top