for helpful discussions and comments. Special thanks to Stefanie Schmidt for advice on compulsory... more for helpful discussions and comments. Special thanks to Stefanie Schmidt for advice on compulsory-schooling data. 1. Data on output per worker are from Summers and Heston (1991), with the correction due to Hall and Jones (1999). Education data are from Barro and Lee (1993). See Krueger wages. Second, although in principle CSLs may be correlated with omitted factors that also affect schooling and future wages, we provide evidence suggesting this is not a problem. Omitted variables related to family background or tastes would likely induce correlation between CSLs and college attendance as well as secondary and middle schooling. The results below show that CSLs affected schooling exclusively in middle-school and high-school grades, suggesting that omitted factors do not bias estimates using CSLs as instruments. A third consideration is that changing CSLs were part of the 1910-1940 high-school movement that Goldin (1998) has argued was responsible for much of the humancapital accumulation in the United States in the twentieth century. The baseline results in the paper use samples of white men aged 40-49 from the 1960-1980 Censuses, though some results use 1950 and 1990 data and samples of men aged 30-39. We focus on the 1960-1980 Censuses because the Census schooling variable changed in 1990. Also, we show below that it is important to control for private returns correctly by instrumenting for individual schooling when estimating external returns. The 1960-1980 Censuses include information on quarter of birth, which can be used as an instrument for individual schooling as in Angrist and Krueger (1991). We start with men in their 40s because they are on a relatively flat part of the age-earnings profile. This makes it easier to control for the effect of individual education on earnings, and facilitates the use of quarter-of-birth instruments for individual schooling. Finally, blacks are excluded because blacks in these cohorts experienced marked changes in school quality (see, e.g., Welch, 1973; Margo, 1990; or Card and Krueger, 1992a). Ordinary least-squares (OLS) estimates using data from the 1960-1980 Censuses show a large positive relationship between average schooling and individual wages. A one-year increase in average schooling is associated with about a 7% increase in average wages, over and above the roughly equal private returns. In contrast with the OLS estimates, instrumental variables (IV) estimates of external returns for men aged 40-49 in 1960-1980 are typically around 1-2%, and significantly lower than the corresponding OLS estimates. Adding data from the 1950 Census and/or data for men aged 30-39 yields slightly smaller and more precise estimates.3 We therefore conclude there is little evidence for large external returns, though the results are consistent with modest external returns of 1-3%. The confidence intervals typically exclude human capital externalities greater than 5-6% and therefore rule out magnitudes in the Comment MARK BILS University of Rochester
Journal of the American Statistical Association, 2015
In the canonical regression discontinuity (RD) design for applicants who face an award or admissi... more In the canonical regression discontinuity (RD) design for applicants who face an award or admissions cutoff, causal effects are nonparametrically identified for those near the cutoff. The impact of treatment on inframarginal applicants is also of interest, but identification of such effects requires stronger assumptions than are required for identification at the cutoff. This paper discusses RD identification away from the cutoff. Our identification strategy exploits the availability of dependent variable predictors other than the running variable. Conditional on these predictors, the running variable is assumed to be ignorable. This identification strategy is illustrated with data on applicants to Boston exam schools. Functional-form-based extrapolation generates unsatisfying results in this context, either noisy or not very robust. By contrast, identification based on RD-specific conditional independence assumptions produces reasonably precise and surprisingly robust estimates of the effects of exam school attendance on inframarginal applicants. These estimates suggest that the causal effects of exam school attendance for 9th grade applicants with running variable values well away from admissions cutoffs differ little from those for applicants with values that put them on the margin of acceptance. An extension to fuzzy designs is shown to identify causal effects for compliers away from the cutoff.
NBER working papers are circulated for discussion and comment purposes. They have not been peer-r... more NBER working papers are circulated for discussion and comment purposes. They have not been peer-reviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications.
Rideshare drivers pay a proportion of their fares to a ride-hailing platform operator, a commissi... more Rideshare drivers pay a proportion of their fares to a ride-hailing platform operator, a commission-based compensation model used by many service providers. To Uber drivers, this commission is known as the Uber fee. By contrast, traditional taxi drivers in most US cities make a fixed payment independent of their earnings, usually a weekly or daily medallion lease, keeping every fare dollar net of lease costs and other expenses. We assess these compensation models using an experiment that offered random samples of Boston Uber drivers opportunities to lease a virtual taxi medallion that eliminates the Uber fee. Some drivers were offered a negative fee. Drivers' labor supply response to our offers reveals a large intertemporal substitution elasticity, on the order of 1.2, and higher for those who accept lease contracts. At the same time, our virtual lease program was under- subscribed: many drivers who would have benefitted from buying an inexpensive lease chose to sit out. We use ...
This paper presents evidence on the child-quantity/child-quality trade-off using quasi-experiment... more This paper presents evidence on the child-quantity/child-quality trade-off using quasi-experimental variation due to twin births and preferences for a mixed sibling-sex composition, as well as ethnic differences in the effects of these variables. Our sample includes groups with very high fertility. An innovation in our econometric approach is the juxtaposition of results from multiple instrumental variables (IV) strategies, capturing the effects of fertility over different ranges for different sorts of people. To increase precision, we develop an estimator that combines different instrument sets across partially-overlapping parity-specific sub-samples. Our results are remarkably consistent in showing no
ust ust over a quarter century ago, Edward Leamer (1983) reflected on the state of empirical work... more ust ust over a quarter century ago, Edward Leamer (1983) reflected on the state of empirical work in economics. He urged empirical researchers to “take the con out of econometrics ” and memorably observed (p. 37): “Hardly anyone takes data analysis seriously. Or perhaps more accurately, hardly anyone takes anyone else’s data analysis seriously. ” Leamer was not alone; Hendry (1980), Sims (1980), and others writing at about the same time were similarly disparaging of empirical practice. Reading these commentaries as late-1980s Ph.D. students, we wondered about the prospects for a satisfying career doing applied work. Perhaps credible empirical work in economics is a pipe dream. Here we address the questions of whether the quality and the credibility of empirical work have increased since Leamer’s pessimistic assessment. Our views are necessarily colored by the areas of applied microeconomics in which we are active, but we look over the fence at other areas as well. Leamer (1983) diag...
This section discusses the construction of st , the indicator of the risk-adjusted expected chang... more This section discusses the construction of st , the indicator of the risk-adjusted expected change in the federal funds rate target based on futures prices for months in which there was an FOMC meeting scheduled. Notice that this variable is a key ingredient of the vector zt when we construct the ordered propensity score model p(zt, ψ). The discussion in this section borrows many elements from Hamilton (2008). Using the same notation as in the previous section, the effective federal funds rate observed on day k of month t can be described as rt,k = rt,k + ut,k
Two-stage least squares estimates in heavily over-identified instrumental variables (IV) models c... more Two-stage least squares estimates in heavily over-identified instrumental variables (IV) models can be misleadingly close to the corresponding ordinary least squares (OLS) estimates when many instruments are weak. Just-identified (just-ID) IV estimates using a single instrument are also biased, but the importance of weak-instrument bias in just-ID IV applications remains contentious. We argue that in microeconometric applications, just-ID IV estimators can typically be treated as all but unbiased and that the usual inference strategies are likely to be adequate. The argument begins with contour plots for confidence interval coverage as a function of instrument strength and explanatory variable endogeneity. These show undercoverage in excess of 5% only for endogeneity beyond that seen even when IV and OLS estimates differ by an order of magnitude. Three widelycited microeconometric applications are used to explain why endogeneity is likely low enough for IV estimates to be reliable. We then show that an estimator that's unbiased given a population first-stage sign restriction has bias exceeding that of IV when the restriction is imposed on the data. But screening on the sign of the estimated first stage is shown to halve the median bias of conventional IV without reducing coverage. To the extent that sign-screening is already part of empirical workflows, reported IV estimates enjoy the minimal bias of sign-screened just-ID IV.
We examine the evolution of economics research using a machine-learning-based classification of p... more We examine the evolution of economics research using a machine-learning-based classification of publications into fields and styles. The changing field distribution of publications would not seem to favor empirical papers. But economics' empirical shift is a within-field phenomenon; even fields that traditionally emphasize theory have gotten more empirical. Empirical work has also come to be more cited than theoretical work. The citation shift is sharpened when citations are weighted by journal importance. Regression analyses of citations per paper show empirical publications reaching citation parity with theoretical publications around 2000. Within fields and journals, however, empirical work is now cited more.
This appendix provides supplementary materials including alternate style plots, full journal list... more This appendix provides supplementary materials including alternate style plots, full journal list, and an alternate economics leaderboard. Figures W1 and W2 replicate the styles publication (Figure 13) and citation (Figure 14) plots, but split the empirical papers into marginally empirical and clearly Note: This figure shows weighted citation shares of economics papers to economics styles. Citations are weighted by importance of the citing journal in the left (AER weights) and right panels (Top-6 weights). Plots are smoothed with five-year moving averages. Papers cited were published between 1970 and 2015. Empirical+ are clearly empirical papers with a confidence score between 75 and 100. Plots are smoothed with five-year moving averages. Shares for each group appear in italics.
The average effect of social programs on outcomes such as earnings is a parameter of primary inte... more The average effect of social programs on outcomes such as earnings is a parameter of primary interest in econometric evaluations studies. New results on using exclusion restrictions to identify and estimate average treatment effects are presented. Identification is achieved given a minimum of parametric assumptions initially without reference to a latent index framework. Most econometric analyses of evaluation models motivate identifying assumptions using models of individual behavior. Our technical conditions do not fit easily into a conventional discrete choice framework, rather they fit into a framework where the source of identifyimg infonnation is institutional knowledge regarding program administration. This framework also suggests an attractive experimental design for research using human subjects, in which eligible participants need not be denied treatment. We present a simple instrumental variables estimator for the average effect of treatment on program participants, and show that the estimator attains Chamberlain's semi-parametric efficiency bound. The bias of estimators that satisfy only exclusion restrictions is also considered.
The effect of government programs on the distribution of participants' earnings is important for ... more The effect of government programs on the distribution of participants' earnings is important for program evaluation and welfare comparisons. This paper reports estimates of the effects of JTPA training programs on the distribution of earnings. The estimation uses a new instrumental variable (IV) method that measures program impacts on the quantiles of outcome variables. This quantile treatment effects (QTE) estimator accommodates exogenous covariates and reduces to quantile regression when selection for treatment is exogenously determined. The QTE estimator can be computed as the solution to a convex linear programming problem, although this requires first-step estimation of a nuisance function. We develop distribution theory for the case where the first step is estimated nonparametrically. For women, the empirical results show that the JTPA program had the largest proportional impact at low quantiles. Perhaps surprisingly, however, JTPA training raised the quantiles of earnings for men only in the upper half of the trainee earnings distribution.
Any opinions expressed here are those of the author(s) and not those of IZA. Research published i... more Any opinions expressed here are those of the author(s) and not those of IZA. Research published in this series may include views on policy, but the institute itself takes no institutional policy positions. The Institute for the Study of Labor (IZA) in Bonn is a local and virtual international research center and a place of communication between science, politics and business. IZA is an independent nonprofit organization supported by Deutsche Post World Net. The center is associated with the University of Bonn and offers a stimulating research environment through its international network, workshops and conferences, data service, project support, research visits and doctoral program. IZA engages in (i) original and internationally competitive research in all fields of labor economics, (ii) development of policy concepts, and (iii) dissemination of research results and concepts to the interested public. IZA Discussion Papers often represent preliminary work and are circulated to encourage discussion. Citation of such a paper should account for its provisional character. A revised version may be available directly from the author.
for helpful discussions and comments. Special thanks to Stefanie Schmidt for advice on compulsory... more for helpful discussions and comments. Special thanks to Stefanie Schmidt for advice on compulsory-schooling data. 1. Data on output per worker are from Summers and Heston (1991), with the correction due to Hall and Jones (1999). Education data are from Barro and Lee (1993). See Krueger wages. Second, although in principle CSLs may be correlated with omitted factors that also affect schooling and future wages, we provide evidence suggesting this is not a problem. Omitted variables related to family background or tastes would likely induce correlation between CSLs and college attendance as well as secondary and middle schooling. The results below show that CSLs affected schooling exclusively in middle-school and high-school grades, suggesting that omitted factors do not bias estimates using CSLs as instruments. A third consideration is that changing CSLs were part of the 1910-1940 high-school movement that Goldin (1998) has argued was responsible for much of the humancapital accumulation in the United States in the twentieth century. The baseline results in the paper use samples of white men aged 40-49 from the 1960-1980 Censuses, though some results use 1950 and 1990 data and samples of men aged 30-39. We focus on the 1960-1980 Censuses because the Census schooling variable changed in 1990. Also, we show below that it is important to control for private returns correctly by instrumenting for individual schooling when estimating external returns. The 1960-1980 Censuses include information on quarter of birth, which can be used as an instrument for individual schooling as in Angrist and Krueger (1991). We start with men in their 40s because they are on a relatively flat part of the age-earnings profile. This makes it easier to control for the effect of individual education on earnings, and facilitates the use of quarter-of-birth instruments for individual schooling. Finally, blacks are excluded because blacks in these cohorts experienced marked changes in school quality (see, e.g., Welch, 1973; Margo, 1990; or Card and Krueger, 1992a). Ordinary least-squares (OLS) estimates using data from the 1960-1980 Censuses show a large positive relationship between average schooling and individual wages. A one-year increase in average schooling is associated with about a 7% increase in average wages, over and above the roughly equal private returns. In contrast with the OLS estimates, instrumental variables (IV) estimates of external returns for men aged 40-49 in 1960-1980 are typically around 1-2%, and significantly lower than the corresponding OLS estimates. Adding data from the 1950 Census and/or data for men aged 30-39 yields slightly smaller and more precise estimates.3 We therefore conclude there is little evidence for large external returns, though the results are consistent with modest external returns of 1-3%. The confidence intervals typically exclude human capital externalities greater than 5-6% and therefore rule out magnitudes in the Comment MARK BILS University of Rochester
Journal of the American Statistical Association, 2015
In the canonical regression discontinuity (RD) design for applicants who face an award or admissi... more In the canonical regression discontinuity (RD) design for applicants who face an award or admissions cutoff, causal effects are nonparametrically identified for those near the cutoff. The impact of treatment on inframarginal applicants is also of interest, but identification of such effects requires stronger assumptions than are required for identification at the cutoff. This paper discusses RD identification away from the cutoff. Our identification strategy exploits the availability of dependent variable predictors other than the running variable. Conditional on these predictors, the running variable is assumed to be ignorable. This identification strategy is illustrated with data on applicants to Boston exam schools. Functional-form-based extrapolation generates unsatisfying results in this context, either noisy or not very robust. By contrast, identification based on RD-specific conditional independence assumptions produces reasonably precise and surprisingly robust estimates of the effects of exam school attendance on inframarginal applicants. These estimates suggest that the causal effects of exam school attendance for 9th grade applicants with running variable values well away from admissions cutoffs differ little from those for applicants with values that put them on the margin of acceptance. An extension to fuzzy designs is shown to identify causal effects for compliers away from the cutoff.
NBER working papers are circulated for discussion and comment purposes. They have not been peer-r... more NBER working papers are circulated for discussion and comment purposes. They have not been peer-reviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications.
Rideshare drivers pay a proportion of their fares to a ride-hailing platform operator, a commissi... more Rideshare drivers pay a proportion of their fares to a ride-hailing platform operator, a commission-based compensation model used by many service providers. To Uber drivers, this commission is known as the Uber fee. By contrast, traditional taxi drivers in most US cities make a fixed payment independent of their earnings, usually a weekly or daily medallion lease, keeping every fare dollar net of lease costs and other expenses. We assess these compensation models using an experiment that offered random samples of Boston Uber drivers opportunities to lease a virtual taxi medallion that eliminates the Uber fee. Some drivers were offered a negative fee. Drivers' labor supply response to our offers reveals a large intertemporal substitution elasticity, on the order of 1.2, and higher for those who accept lease contracts. At the same time, our virtual lease program was under- subscribed: many drivers who would have benefitted from buying an inexpensive lease chose to sit out. We use ...
This paper presents evidence on the child-quantity/child-quality trade-off using quasi-experiment... more This paper presents evidence on the child-quantity/child-quality trade-off using quasi-experimental variation due to twin births and preferences for a mixed sibling-sex composition, as well as ethnic differences in the effects of these variables. Our sample includes groups with very high fertility. An innovation in our econometric approach is the juxtaposition of results from multiple instrumental variables (IV) strategies, capturing the effects of fertility over different ranges for different sorts of people. To increase precision, we develop an estimator that combines different instrument sets across partially-overlapping parity-specific sub-samples. Our results are remarkably consistent in showing no
ust ust over a quarter century ago, Edward Leamer (1983) reflected on the state of empirical work... more ust ust over a quarter century ago, Edward Leamer (1983) reflected on the state of empirical work in economics. He urged empirical researchers to “take the con out of econometrics ” and memorably observed (p. 37): “Hardly anyone takes data analysis seriously. Or perhaps more accurately, hardly anyone takes anyone else’s data analysis seriously. ” Leamer was not alone; Hendry (1980), Sims (1980), and others writing at about the same time were similarly disparaging of empirical practice. Reading these commentaries as late-1980s Ph.D. students, we wondered about the prospects for a satisfying career doing applied work. Perhaps credible empirical work in economics is a pipe dream. Here we address the questions of whether the quality and the credibility of empirical work have increased since Leamer’s pessimistic assessment. Our views are necessarily colored by the areas of applied microeconomics in which we are active, but we look over the fence at other areas as well. Leamer (1983) diag...
This section discusses the construction of st , the indicator of the risk-adjusted expected chang... more This section discusses the construction of st , the indicator of the risk-adjusted expected change in the federal funds rate target based on futures prices for months in which there was an FOMC meeting scheduled. Notice that this variable is a key ingredient of the vector zt when we construct the ordered propensity score model p(zt, ψ). The discussion in this section borrows many elements from Hamilton (2008). Using the same notation as in the previous section, the effective federal funds rate observed on day k of month t can be described as rt,k = rt,k + ut,k
Two-stage least squares estimates in heavily over-identified instrumental variables (IV) models c... more Two-stage least squares estimates in heavily over-identified instrumental variables (IV) models can be misleadingly close to the corresponding ordinary least squares (OLS) estimates when many instruments are weak. Just-identified (just-ID) IV estimates using a single instrument are also biased, but the importance of weak-instrument bias in just-ID IV applications remains contentious. We argue that in microeconometric applications, just-ID IV estimators can typically be treated as all but unbiased and that the usual inference strategies are likely to be adequate. The argument begins with contour plots for confidence interval coverage as a function of instrument strength and explanatory variable endogeneity. These show undercoverage in excess of 5% only for endogeneity beyond that seen even when IV and OLS estimates differ by an order of magnitude. Three widelycited microeconometric applications are used to explain why endogeneity is likely low enough for IV estimates to be reliable. We then show that an estimator that's unbiased given a population first-stage sign restriction has bias exceeding that of IV when the restriction is imposed on the data. But screening on the sign of the estimated first stage is shown to halve the median bias of conventional IV without reducing coverage. To the extent that sign-screening is already part of empirical workflows, reported IV estimates enjoy the minimal bias of sign-screened just-ID IV.
We examine the evolution of economics research using a machine-learning-based classification of p... more We examine the evolution of economics research using a machine-learning-based classification of publications into fields and styles. The changing field distribution of publications would not seem to favor empirical papers. But economics' empirical shift is a within-field phenomenon; even fields that traditionally emphasize theory have gotten more empirical. Empirical work has also come to be more cited than theoretical work. The citation shift is sharpened when citations are weighted by journal importance. Regression analyses of citations per paper show empirical publications reaching citation parity with theoretical publications around 2000. Within fields and journals, however, empirical work is now cited more.
This appendix provides supplementary materials including alternate style plots, full journal list... more This appendix provides supplementary materials including alternate style plots, full journal list, and an alternate economics leaderboard. Figures W1 and W2 replicate the styles publication (Figure 13) and citation (Figure 14) plots, but split the empirical papers into marginally empirical and clearly Note: This figure shows weighted citation shares of economics papers to economics styles. Citations are weighted by importance of the citing journal in the left (AER weights) and right panels (Top-6 weights). Plots are smoothed with five-year moving averages. Papers cited were published between 1970 and 2015. Empirical+ are clearly empirical papers with a confidence score between 75 and 100. Plots are smoothed with five-year moving averages. Shares for each group appear in italics.
The average effect of social programs on outcomes such as earnings is a parameter of primary inte... more The average effect of social programs on outcomes such as earnings is a parameter of primary interest in econometric evaluations studies. New results on using exclusion restrictions to identify and estimate average treatment effects are presented. Identification is achieved given a minimum of parametric assumptions initially without reference to a latent index framework. Most econometric analyses of evaluation models motivate identifying assumptions using models of individual behavior. Our technical conditions do not fit easily into a conventional discrete choice framework, rather they fit into a framework where the source of identifyimg infonnation is institutional knowledge regarding program administration. This framework also suggests an attractive experimental design for research using human subjects, in which eligible participants need not be denied treatment. We present a simple instrumental variables estimator for the average effect of treatment on program participants, and show that the estimator attains Chamberlain's semi-parametric efficiency bound. The bias of estimators that satisfy only exclusion restrictions is also considered.
The effect of government programs on the distribution of participants' earnings is important for ... more The effect of government programs on the distribution of participants' earnings is important for program evaluation and welfare comparisons. This paper reports estimates of the effects of JTPA training programs on the distribution of earnings. The estimation uses a new instrumental variable (IV) method that measures program impacts on the quantiles of outcome variables. This quantile treatment effects (QTE) estimator accommodates exogenous covariates and reduces to quantile regression when selection for treatment is exogenously determined. The QTE estimator can be computed as the solution to a convex linear programming problem, although this requires first-step estimation of a nuisance function. We develop distribution theory for the case where the first step is estimated nonparametrically. For women, the empirical results show that the JTPA program had the largest proportional impact at low quantiles. Perhaps surprisingly, however, JTPA training raised the quantiles of earnings for men only in the upper half of the trainee earnings distribution.
Any opinions expressed here are those of the author(s) and not those of IZA. Research published i... more Any opinions expressed here are those of the author(s) and not those of IZA. Research published in this series may include views on policy, but the institute itself takes no institutional policy positions. The Institute for the Study of Labor (IZA) in Bonn is a local and virtual international research center and a place of communication between science, politics and business. IZA is an independent nonprofit organization supported by Deutsche Post World Net. The center is associated with the University of Bonn and offers a stimulating research environment through its international network, workshops and conferences, data service, project support, research visits and doctoral program. IZA engages in (i) original and internationally competitive research in all fields of labor economics, (ii) development of policy concepts, and (iii) dissemination of research results and concepts to the interested public. IZA Discussion Papers often represent preliminary work and are circulated to encourage discussion. Citation of such a paper should account for its provisional character. A revised version may be available directly from the author.
Uploads
Papers by Joshua Angrist