We revisit the identification argument of Kirkeboen et al. (2016) who showed how one may combine instruments for multiple unordered treatments with information about individuals’ ranking of these treatments to achieve identification while allowing for both observed and unobserved heterogeneity in treatment effects. We show that the key assumptions underlying their identification argument have testable implications, and provide a new characterization of the bias that may arise if these assumptions are violated. We apply these insights in a comparison of earnings payoffs to post-secondary fields of study in Norway and Denmark. In both countries, we find that field of study, not only education levels, is a key determinant of individual earnings. We also find a strong cross-country correlation in the payoffs to field of study, which is robust to removing fields with evidence of violations of the identifying assumptions. The relative rewards to specific types of human capital are therefore found to be comparable in two small open economies that differ distinctly in industry structure, but are broadly similar in education system, wage setting institutions, and tax-transfer system.
PDF |
We examine how the adoption of information communication technology affects bilateral trade. The context is a public program in Norway that rolled out broadband access points leading to plausibly exogenous variation in the availability and adoption of broadband by firms. We find that broadband makes trade patterns more sensitive to distance and economic size. These results are consistent with a model of trade with variable elasticity of demand. The model predicts that adoption of a technology that lowers information frictions enlarges the choice set of exporters and importers. This makes demand more elastic with respect to trade costs and thus distance.
In this paper we investigate the effect of Head Start on long term education and labor market outcomes using data from the NLSY79. The contributions to the existing literature on the effectiveness of Head Start are threefold: (1) we are the first to examine distributional effects of Head Start on long term outcomes (2) we do not rely on quasi-experimental variation in Head Start participation but instead perform a nonparametric bounds analysis that relies on weak stochastic dominance assumptions and (3) we consider education and labor market outcomes observed for individuals in their early 30s. The results show that Head Start has a statistically significant positive effect on years of education, in particular for women, blacks and Hispanics. For wage income we also find evidence that Head Start has beneficial impacts, with effects located at the lower end of the distribution.
How does class size in compulsory school affect people's long run education and earnings? We use maximum class size rules and Norwegian administrative registries allowing us to observe outcomes up to age 48. We do not find any indication of beneficial effects of class size reduction in compulsory school. For a 1 person reduction in class size we reject effects on income as small as 0.12 percent in primary school and 0.15 percent in middle school. Population differences in parental background, school size or competitive pressure do not appear to reconcile our findings with previous studies.
PDF |
We exploit lottery-determined admission to dental school to estimate the payoffs to the study of dentistry in the Netherlands. Using data from up to 22 years after the lottery, we find that in most years after graduation dentists earn around 50,000 Euros more than they would earn in their next-best profession. The payoff is larger for men than for women but does not vary with high school GPA. The large payoffs cannot be attributed to longer working hours, larger investments while studying (opportunity costs and direct costs), or unpleasant aspects of working as a dentist. A plausible explanation is that dentists earn a monopoly rent. Results from regressions of dentists’ earnings on dentists density are consistent with this, as are the facts that the supply of dentists in the Netherlands is low and that the payoff does not vary with high school GPA.
PDF |
This paper estimates peer effects originating from the ability composition of tutorial groups for undergraduate students in economics. We manipulated the composition of groups to achieve a wide range of support, and assigned students - conditional on their ability - randomly. The data support a specification in which the group composition is captured by the mean and standard deviation of prior ability and their squares and interaction. Estimates from this specification imply that students of low and medium ability gain on average 0.2 SD units of achievement from switching from ability mixing to three-way tracking. Their dropout rate is reduced by 15 percentage points (relative to a mean of 0.6). High-ability students are unaffected. Analysis of survey data indicates that in tracked groups, low-ability students have more positive interactions with other students, and are more involved. We find no evidence that teachers adjust their teaching to the composition of groups.
This paper examines the labor-market payoffs to different types of post-secondary education, including field and institution of study. Instrumental variables (IV) estimation of the payoff to choosing one type of education as compared to another is made particularly challenging by individuals choosing between several types of education. Not only does identification require one instrument per alternative, but it is also necessary to deal with the issue that individuals who choose the same education may have different next-best alternatives. We address these difficulties using rich administrative data for Norway’s post-secondary education system. A centralized admission process creates credible instruments from discontinuities that effectively randomize applicants near unpredictable admission cutoffs into different institutions and fields of study. The admission process also provides information on preferred and next-best alternatives from strategy-proof measures of individuals ranking of institutions and fields. The results from our IV approach may be summarized with three broad conclusions. First, different fields of study have substantially different labor-market payoffs, even after accounting for institution and peer quality. Second, the effect on earnings from attending a more selective institution tends to be relatively small as compared to payoffs to field of study. Third, the estimated payoffs to field of study are consistent with individuals choosing fields in which they have a comparative advantage. Comparing our estimates to those obtained from other approaches highlights the importance of using instruments to correct for selection bias and information on individuals ranking of institutions and fields to measure their preferred and next-best alternatives.
We exploit admission lotteries to estimate the returns to medical school in the Netherlands. Using data from up to 22 years after the lottery, we find that in every single year after graduation doctors earn at least 20 percent more than people who end up in their next-best occupation. Twenty-two years after the lottery the earnings difference is almost 50 percent. Only a small fraction of this difference can be attributed to differences in working hours and human capital investments. The returns do not vary with gender or ability, and shift the entire earnings distribution.
What is the effect of school consolidation on student achievement? Theory gives little guidance because possibly positive effects from larger school size can be offset by negative effects from reduced choice and competition. We investigate these issues empirically by analyzing the effects on students’ achievement of a consolidation reform that took place in Dutch primary education in the mid 1990s. The reform was implemented by increasing the minimum required school size, leading to an increase in actual school size and a reduction in the number of schools. For identification we exploit variation between municipalities. We find that an increase in the minimum required school size of 10% has a small positive effect on student achievement of 0.72% of a standard deviation. Further analysis indicates that this effect can be mainly attributed to the increase in actual school size; reduced competition and choice do not seem to have harmed student achievement. We also find no evidence that the consolidation effect is driven by reduced school segregation or the elimination of small schools that were – given their size – underperforming.
Job characteristics can affect worker turnover through their effect on utility and through their effect on outside job opportunities. We separately identify and estimate the roles of these two channels. Our method exploits information on job changes and relies on an augmented sample selection correction. Taking our approach to an exhaustive register of Dutch primary school teachers, and using arguably plausible exclusion restrictions, we show a detailed picture of preferences for school characteristics. We also study how preference estimates may be biased when ignoring information on job opportunities and discuss the implications for the analysis of teacher turnover.
There exists a strong educational gradient in cancer risk, which has been documented in a wide range of populations. Yet relatively little is known about the extent to which education is causally linked to cancer incidence and mortality. This paper exploits a large social experiment where an education reform expanded compulsory schooling during the 1960s in Norway. The reform led to a discontinuous increase in educational attainment, which we exploit to estimate the effect of the reform on various cancer outcomes. Our main finding is that education has little if any impact on cancer risk. This holds for all cancer sites together as well as the most common cancer sites in isolation, with two exceptions. The compulsory school reform lowered the risk of lung cancer for men, but increased the risk of colorectal cancer for women.
PDF |
To estimate the impact of grade composition on pupil achievement we exploit grade mixing rules in Norwegian junior high schools in an instrumental variable framework. We find that pupils in mixed grade classrooms outperform pupils in single grade classrooms on high stake central exit tests and teacher set and graded tests. This effect is driven by pupils benefiting from sharing the classroom with more mature peers from higher grades. We also find that the presence of lower grade peers is detrimental for achievement. Children can therefore benefit from de-tracking by grade, but the effects depend crucially on how the classroom is balanced in terms of lower and higher grades.
This paper estimates the effects of attending medical school on health outcomes by exploiting that admission to medical school in the Netherlands is determined by a lottery. Among the applicants for medical school, people who attended medical school have on average 1.5 more years of completed education than people who did not. They are also more likely to have been exposed to a health-related education curriculum. The results show only modest impacts on health outcomes. Attending medical school reduces alcohol consumption and being underweight somewhat. It has, however, no significant impact on self-reported health status, smoking, physical exercise and being overweight or obese. Attending medical school does have a large positive impact on the probability of being registered for donations of organs.
Does internet use trigger sex crime? We use unique Norwegian data on crime and internet adoption to shed light on this question. A public program with limited funding rolled out broadband access points in 2000–2008, and provides plausibly exogenous variation in internet use. Our instrumental variables estimates show that internet use is associated with a substantial increase in both reports, charges and convictions of rape and other sex crimes. We present a conceptual framework that highlights three mechanisms for how internet use may affect reported sex crime, namely a reporting effect, a matching effect on potential offenders and victims, and a direct effect on sex crime propensity. To investigate the importance of these mechanisms, we use data on individual reporting behavior, police investigations, and criminal charges and convictions. None of the analyses we perform suggest that the positive relationship between internet use and sex crime is driven by changes in reporting behavior. Our findings suggest, however, that the direct effect on sex crime propensity is positive and non-negligible, possibly as a result of increased consumption of pornography.
We study student loan behavior in the Netherlands where i) higher education students know little about the conditions of the government's financial aid program and ii) take-up rates are low. In a field experiment we manipulated the amount of information students have about these conditions. The treatment has no impact on loan take-up, which is not due to students already having decided to take a loan or students not absorbing the information. We conclude that a lack of knowledge about specific policy parameters does not necessarily imply a binding information constraint.
This chapter surveys the economics literature on overeducation. The original motivation to study this topic were reports that the strong increase in the number of college graduates in the early 1970s in the US led to a decrease in the returns to college education. We argue that Duncan and Hoffman's augmented wage equation -- the workhorse model in the overeducation literature -- in which wages are regressed on years of overschooling, years of required schooling and years of underschooling is at best loosely related to this original motivation. We discuss measurement and estimation issues and give an overview of the main empirical findings in this literature. Finally we given an appraisal of the economic lessons learned.
PDF |
A vast body of empirical studies lends support to the incentive effects of rank-order tournaments. Evidence comes from experiments in laboratories and non-experimental studies exploiting sports or firm data. Selection of competitors across tournaments may bias these non-experimental studies, whereas short task duration or lack of distracters may limit the external validity of results obtained in lab experiments or from sports data. To address these concerns we conducted a field experiment where students selected themselves into tournaments with different prizes. Within each tournament the best performing student on the final exam of a standard introductory microeconomics course could win a substantial financial reward. A standard non-experimental analysis exploiting across tournament variation in reward size and competitiveness confirms earlier findings. We find however no evidence for effects of tournament participation on study effort and exam results when we exploit our experimental design, indicating that the non-experimental results are completely due to sorting. Treatment only affects attendance of the first workgroup meeting following the announcement of treatment status, suggesting a difference between short-run and long-run decision making.
In a randomized field experiment where first year university students could earn financial rewards for passing the first year requirements within one year we find small and non-significant average effects of financial incentives on the pass rate and the numbers of collected credit points. There is however evidence for heterogeneous treatment effects. In particular, in the first year high ability students have higher pass rates and collect significantly more credit points when assigned to (higher) reward groups. Low ability students collect less credit points when assigned to higher reward groups. After three years these effects have increased, suggesting dynamic spillovers. The small average effect is therefore the sum of a positive effect of high ability students and a (partly) off-setting negative effect for low ability students. A negative effect of financial incentives for less ?capitalized? individuals is in line with research from psychology and recent economic laboratory experiments which shows that external rewards may be detrimental for intrinsic motivation.
This study presents quasi-experimental estimates of the effect of expanding early schooling enrollment possibilities on early achievement. It exploits two features of the school system in Holland. The first is rolling admissions; children are allowed start school immediately after their 4th birthday instead of at the beginning of the school year. The second is that children having their birthday before, during and after the summer holiday are placed in the same class. These features generate sufficient exogenous variation in children's maximum length of schooling to identify its effects on test scores. Making available one additional month of time in school increases language scores of disadvantaged pupils by 0.06 of a standard deviation and their math scores by 0.05 of a standard deviation. For non-disadvantaged pupils we find no effect.
Using a comprehensive administrative database we exploit independent quasi-experimental methods to estimate the effect of class size on student achievement in Norway. The first method is based on a maximum class size rule in the spirit Angrist and Lavy (1999). The second method exploits population variation as first proposed by Hoxby (2000). The results of both methods (and of variations on these methods) are very similar and cannot reject that the class size effect is equal to zero. The estimates are very precise; we can rule out effects as small as 1.5 percent of a standard deviation for a one student change in class size during three years in a row.
PDF |
This paper follows an alternative approach to identify the wage effects of private-sector training. The idea is to narrow down the comparison group by only taking into consideration the workers who wanted to participate in training but did not do so because of some random event. This makes the comparison group increasingly similar to the group of participants in terms of observed individual characteristics and the characteristics of (planned) training events. At the same time the point estimate of the average return to training consistently drops from a large and significant return to a pointrnestimate close to zero.
This paper evaluates the effects of two subsidy schemes targeted at disadvantaged students in the Netherlands. The first scheme gives schools with at least 70 percent minority students extra funding for personnel. The second scheme gives schools with at least 70 percent students from different disadvantaged groups extra funding for computers and for language materials. The cutoffs at 70 percent provide a regression discontinuity design which we exploit in an instrumental variables framework. Estimates of the effects of the Personnel Subsidy on achievement of 8th graders in language, math and information processing are positive but never significantly different from zero. Estimates of the effects of the Computer Subsidy on the same outcome variables are negative and in some cases significantly different from zero.
This paper reviews the existing evidence on workplace training in Europe in different data sources - the CVTS, OECD data and the European Community Household Panel. We outline the differences in training incidence and relate these differences to the private costs and benefits of training, and to institutional factors such as unions, employment protection and product market competition. We ask whether there is a case for under-provision of training in Europe and examine alternative policies aiming both at raising training incidence and at reducing inequalities in the provision of skills.
PDF |
Standard economic theory predicts that firms will not invest in general training and will underinvest in specific training. Empirical evidence, however, indicates that firms do invest in general training of their workers. Evidence from laboratory experiments points to less underinvestment in specific training than theory predicts. We propose a simple model in which a firm invests the socially optimal amounts in general and specific training if the worker is sufficiently motivated by reciprocity. A reciprocal worker may be willing to give the firm a full return on its investment. We present empirical evidence that supports the proposed mechanism. Workers with a high sensitivity to reciprocity have 15% higher training rates than workers with a low sensitivity to reciprocity.
PDF |
PDF |
This paper explores the hypothesis that wage differentials between skill groups across countries are consistent with a demand and supply framework. Using micro data from 15 countries we find that about one third of the variation in relative wages between skill groups across countries is explained by differences in net supply of skill groups. The demand and supply framework does an even better job at explaining relative wages of low skilled workers.
PDF |
Dutch employers can claim an extra tax deduction when they trainrnemployees older than age 40. This discontinuity in a firm’s trainingrncost is exploited to identify the tax deduction’s effects on trainingrnparticipation and of training participation on wages. The results showrnthat the training rate of workers just above age 40 is 15%–20% higherrnthan that of workers just below age 40. This difference mainly resultsrnfrom the postponement of training and is not a stimulating effect ofrnthe measure. The two-stage least squares estimate of the wage effectrnof training is not statistically different from zero.
PDF |
Breach remedies serve an important role in protecting relationship-specific investments. Theory predicts that some common remedies protect too well and induce overinvestment, either though complete insurance against potential separation or the possibility that breach is prevented byrnincreasing the damage payment due through the investment made. In this article we report on an experiment designed to address whether these two motives show up in practice. In line with theoretical predictions, we find that overinvestment does not occur under liquidated damages. In the case of expectation damages, the full-insurance motive indeed appears to be operative. In the case of reliance damages, both motives are at work, as predicted.
PDF |
In a corporatist country, of which the Netherlands is an example, wages should not be distinguished by union membership status, but by the bargaining regime. Four bargaining regimes can be distinguished: (i) company level bargaining, (ii) industry level bargaining, (iii) mandatory extension of an industry agreement, and (iv) no collective bargaining. Acknowledging firms bargaining regime, we find small differences between the four regimes, and certainly no distinctionrnbetween covered and uncovered firms.
PDF |
PDF |
PDF |