THE EFFECTS OF EXPANDING THE GENEROSITY OF THE STATUTORY SICKNESS INSURANCE SYSTEM

This article evaluates an expansion of employer‐mandated sick leave from 80% to 100% of forgone gross wages in Germany. We employ and compare parametric difference‐in‐difference (DID), matching DID and mixed approaches. Overall workplace absences increased by at least 10% or 1 day per worker per year. We show that taking partial compliance into account increases coefficient estimates. Further, heterogeneity in response behavior was of great importance. There is no evidence that the increase in sick leave improved employee health, a finding that supports a shirking explanation. Finally, we provide evidence on potential labor market adjustments to the reform. Copyright © 2013 John Wiley & Sons, Ltd.     

Combining Matching and Nonparametric Instrumental Variable Estimation: Theory and An Application to the Evaluation of Active Labour Market Policies

We show how instrumental variable and matching estimators can be combined in order to identify a broader array of treatment effects. Instrumental variable (IV) estimators are known to estimate effects only for the compliers, representing a subset of the entire population. By combining IV with matching, we can estimate the treatment effects for the always‐ and never‐takers as well. Since in many cases these groups are the (endogenous) outcome of some assignment process, such estimates also help in judging the implications of such a selection process. In our application to the effects of participation in active labour market programmes in Switzerland, we find large and lasting positive employment effects for the compliers, whereas the effects for the always‐ and never‐participants are small. In addition, the compliers have worse employment outcomes without treatment than those who participate in the programme with or without the intervention under investigation. This suggests that the earlier assignment policy of the caseworkers was inefficient in that the always‐participants were neither those unemployed who would experience the highest expected treatment effects nor those unemployed who had the largest need for assistance. Copyright © 2014 John Wiley & Sons, Ltd.     

A semiparametric model for binary response and continuous outcomes under index heteroscedasticity

This paper formulates a likelihood‐based estimator for a double‐index, semiparametric binary response equation. A novel feature of this estimator is that it is based on density estimation under local smoothing. While the proofs differ from those based on alternative density estimators, the finite sample performance of the estimator is significantly improved. As binary responses often appear as endogenous regressors in continuous outcome equations, we also develop an optimal instrumental variables estimator in this context. For this purpose, we specialize the double‐index model for binary response to one with heteroscedasticity that depends on an index different from that underlying the ‘mean response’. We show that such (multiplicative) heteroscedasticity, whose form is not parametrically specified, effectively induces exclusion restrictions on the outcomes equation. The estimator developed exploits such identifying information. We provide simulation evidence on the favorable performance of the estimators and illustrate their use through an empirical application on the determinants, and affect, of attendance at a government‐financed school. Copyright © 2009 John Wiley & Sons, Ltd.     

Estimating within‐cluster spillover effects using a cluster randomization with application to knowledge diffusion in rural India

Spillover effects within randomized clusters pose a challenge for identifying impacts of an individualized treatment. The paper proposes a solution. Longitudinal and intra‐household observations are combined in estimating the direct knowledge gain from watching an info‐movie in rural India, while randomized village assignment identifies knowledge sharing. Simulations on synthetic data and econometric tests provide support for the estimation method. We find evidence of information sharing, but far less so for disadvantaged groups, such as illiterate and lower‐caste individuals; these groups rely more on actually seeing the movie. Our results are consistent with sizeable biases in ordinary least squares, matching or instrumental variable impact estimators that ignore within‐cluster spillovers.     

CAUSAL INFERENCE ON EDUCATION POLICIES: A SURVEY OF EMPIRICAL STUDIES USING PISA,TIMSS AND PIRLS

The identification of the causal effects of educational policies is the top priority in recent education economics literature. As a result, a shift can be observed in the strategies of empirical studies. They have moved from the use of standard multivariate statistical methods, which identify correlations or associations between variables only, to more complex econometric strategies, which can help to identify causal relationships. However, exogenous variations in databases have to be identified in order to apply causal inference techniques. This is a far from straightforward task. For this reason, this paper provides an extensive and comprehensive overview of the literature using quasi‐experimental techniques applied to three well‐known international large‐scale comparative assessments, such as PISA, PIRLS or TIMSS, over the period 2004–2016. In particular, we review empirical studies employing instrumental variables, regression discontinuity designs, difference in differences and propensity score matching to the above databases. Additionally, we provide a detailed summary of estimation strategies, issues treated and profitability in terms of the quality of publications to encourage further potential evaluations. The paper concludes with some operational recommendations for prospective researchers in the field.     

A REVIEW AND COMPARISON OF TESTS OF CROSS-SECTION INDEPENDENCE IN PANELS

Abstract.  In this paper we review and compare diagnostic tests of cross-section independence in the disturbances of panel regression models. We examine tests based on the sample pairwise correlation coefficient or on its transformations, and tests based on the theory of spacings. The ultimate goal is to shed some light on the appropriate use of existing diagnostic tests for cross-equation error correlation. Our discussion is supported by means of a set of Monte Carlo experiments and a small empirical study on health. Results show that tests based on the average of pairwise correlation coefficients work well when the alternative hypothesis is a factor model with non-zero mean loadings. Tests based on spacings are powerful in identifying various forms of strong cross-section dependence, but have low power when they are used to capture spatial correlation.     

Using OLS to Estimate and Test for Structural Changes in Models with Endogenous Regressors

We consider the problem of estimating and testing for multiple breaks in a single‐equation framework with regressors that are endogenous, i.e. correlated with the errors. We show that even in the presence of endogenous regressors it is still preferable, in most cases, to simply estimate the break dates and test for structural change using the usual ordinary least squares (OLS) framework. Except for some knife‐edge cases, it delivers estimates of the break dates with higher precision and tests with higher power compared to those obtained using an instrumental variable (IV) method. Also, the OLS method avoids potential weak identification problems caused by weak instruments. To illustrate the relevance of our theoretical results, we consider the stability of the New Keynesian hybrid Phillips curve. IV‐based methods only provide weak evidence of instability. On the other hand, OLS‐based ones strongly indicate a change in 1991:Q1 and that after this date the model loses all explanatory power. Copyright © 2013 John Wiley & Sons, Ltd.     

The Blocking Lemma for a many-to-one matching model

The Blocking Lemma identifies a particular blocking pair for each non-stable and individually rational matching that is preferred by some agents of one side of the market to their optimal stable matching. Its interest lies in the fact that it has been an instrumental result to prove key results on matching. For instance, the fact that in the college admissions problem the workers-optimal stable mechanism is group strategy-proof for the workers and the strong stability theorem in the marriage model follow directly from the Blocking Lemma. However, it is known that the Blocking Lemma and its consequences do not hold in the general many-to-one matching model in which firms have substitutable preference relations. We show that the Blocking Lemma holds for the many-to-one matching model in which firms’ preference relations are, in addition to substitutable, quota q-separable. We also show that the Blocking Lemma holds on a subset of substitutable preference profiles if and only if the workers-optimal stable mechanism is group strategy-proof for the workers on this subset of profiles.     

Stable matchings and the small core in Nash equilibrium in the college admissions problem

Both rematching proof and strong equilibrium outcomes are stable with respect to the true preferences in the marriage problem. We show that not all rematching proof or strong equilibrium outcomes are stable in the college admissions problem. But we show that both rematching proof and strong equilibrium outcomes in truncations at the match point are all stable in the college admissions problem. Further, all true stable matchings can be achieved in both rematching proof and strong equilibrium in truncations at the match point. We show that any Nash equilibrium in truncations admits one and only one matching, stable or not. Therefore, the core at a Nash equilibrium in truncations must be small. But examples exist such that the set of stable matchings with respect to a Nash equilibrium may contain more than one matching. Nevertheless, each Nash equilibrium can only admit at most one true stable matching. If, indeed, there is a true stable matching at a Nash equilibrium, then the only possible equilibrium outcome will be the true stable matching, no matter how different are players' equilibrium strategies from the true preferences and how many other unstable matchings are there at that Nash equilibrium. Thus, we show that a necessary and sufficient condition for the stable matching rule to be implemented in a subset of Nash equilibria by the direct revelation game induced by a stable mechanism is that every Nash equilibrium profile in that subset admits one and only one true stable matching. Received: 30 December 1998 / Accepted: 12 October 2001 This paper is a revision of the paper “Manipulation and Stability in a College Admissions Problem” circulated since 1994. I thank Rich McLean, Abraham Neyman, Mark Satterthwaite, Sang-Chul Suh, and Tetsuji Yamada for helpful discussions. I thank the associate editor and the two anonymous referees for their helpful comments that have greatly improved the paper. I am grateful to the Kellogg G.S.M. at the Northwestern University for the hospitality for my visit. Any errors are mine.     

Matching models with a conservation law: The existence and global structure of the set of stationary equilibria

We study random matching models where there is a set of infinitely lived agents, and in each period agents are pairwise matched to each other and play a stage game. We investigate the basic structure of equilibria in such models: the existence of equilibria and the global structure of the set of equilibria. Specifically, we focus on models with a conservation law, which typically holds in economies having some assets, such as money. In such models, under certain regularity conditions the set of equilibria is one-dimensional and each connected component of it is a piecewise smooth one-dimensional manifold being homeomorphic to either the unit circle or the unit interval. Moreover, in an endpoint of an interval all agents have the same amount of assets.     

WHAT IS THE EFFECT OF HOUSEWORK ON THE MARKET WAGE,AND CAN IT EXPLAIN THE GENDER WAGE GAP?

Abstract Does housework reduce the market wage, and if so, does it have a similar impact for males and females? In this paper, we survey and evaluate the recent and growing empirical literature on the linkages between housework and the wage rate. The review is motivated by unexplained gender wage gaps across studies, which consider personal and market‐related factors. We focus on this less‐studied aspect of wage determination. We consider the required modelling framework, and provide standardized estimated effects of housework on the hourly wage across studies. We evaluate how this literature has addressed potential estimation problems, in particular, the endogeneity of housework, concavity of the housework–wage function, threshold effects and work effort effects. We conclude that the evidence across ordinary least squares, instrumental variable, fixed effects and two‐stage least squares results casts serious doubt on the idea that the negative female housework–wage relationship is only driven by endogeneity bias or individual‐specific characteristics. Yet, much more needs to be done to address modelling and data requirements, and we point out likely and promising future research directions.     

Stable many-to-many matchings with contracts

We consider several notions of setwise stability for many-to-many matching markets with contracts and provide an analysis of the relations between the resulting sets of stable allocations for general, substitutable, and strongly substitutable preferences. Apart from obtaining “set inclusion results” on all three domains, we introduce weak setwise stability as a new stability concept and prove that for substitutable preferences the set of pairwise stable matchings is nonempty and coincides with the set of weakly setwise stable matchings. For strongly substitutable preferences the set of pairwise stable matchings coincides with the set of setwise stable matchings.     

Difference‐in‐differences when the treatment status is observed in only one period

This paper considers the difference‐in‐differences (DID) method when the data come from repeated cross‐sections and the treatment status is observed either before or after the implementation of a program. We propose a new method that point‐identifies the average treatment effect on the treated (ATT) via a DID method when there is at least one proxy variable for the latent treatment. Key assumptions are the stationarity of the propensity score conditional on the proxy and an exclusion restriction that the proxy must satisfy with respect to the change in average outcomes over time conditional on the true treatment status. We propose a generalized method of moments estimator for the ATT and we show that the associated overidentification test can be used to test our key assumptions. The method is used to evaluate JUNTOS, a Peruvian conditional cash transfer program. We find that the program significantly increased the demand for health inputs among children and women of reproductive age.     

Utility Theory and reciprocal pairwise comparisons: The Eigenvector Method

The criticisms of Utility Theory focus on either its axioms or the construction of utility functions. Here we present a method which avoids the problems of uniqueness encountered in the construction of utility functions when using either the certainty equivalence method or the probability equivalence method. The method is based on the construction of ratio scale value functions from reciprocal pairwise comparisons and Saaty's Eigenvector Method. We show that under the assumption of cardinal consistency utility functions are a particular case of these ratio scales. Reciprocal pairwise comparisons allow decision makers to relax the transitivity assumption and help to derive a unique scaling of preferences.     

A Semi‐Parametric Analysis of Two‐Sided Markets: An Application to the Local Daily Newspapers in the USA

This paper considers an empirical semiparametric model for two‐sided markets. Contrary to existing empirical literature on two‐sided markets, we specify network effects and probability distribution functions of net benefits of the two sides nonparametrically. We then estimate the model by nonparametric instrumental variables regression for local daily newspapers from the USA. We show that semiparametric specification is supported by the data and the network effects are neither linear nor monotonic. With a numerical illustration we demonstrate that the mark‐up of the newspaper on each side changes drastically with the nonlinearly specified network effects from the case with linear network effects. Copyright © 2015 John Wiley & Sons, Ltd.     

Subsidizing low-skilled jobs in a dual labor market

We introduce a matching model that allows for classical and frictional unemployment. The labor market is dual featuring low-skilled and high-skilled workers and simple and complex jobs. Simple jobs pay a minimum wage, while wages in the complex jobs are determined by Nash bargaining. Opportunities for low-skilled workers are limited to simple jobs; while high-skilled unemployed can apply for both types of jobs, and thereby can accept to be downgraded. We analyze the outcomes of simple job subsidy policies assuming that government budget is balanced through taxes on occupied workers. We first give conditions for the existence and uniqueness of a steady-state equilibrium and we then analyze the effects of different fiscal instruments. We show that in this set-up, increasing simple job subsidies does not necessarily reduce low-skilled unemployment or unemployment spells. By introducing heterogeneous skills and possible downgrading of the high-skilled workers, we show that the effectiveness of such policies in reducing the classical unemployment is decreasing. In fact, any additional classical unemployed re-entering the job market is accompanied by an increasing number of high-skilled workers downgrading to low-skilled jobs. We calibrate the model on French labor market data. It is found that for five low-skilled workers leaving classical unemployment, two high-skilled workers are downgraded.     

Treatment effect bounds: An application to Swan–Ganz catheterization

We reanalyze data from the observational study by Connors et al. (1996) on the impact of Swan–Ganz catheterization on mortality outcomes. The study by Connors et al. (1996) assumes that there are no unobserved differences between patients who are catheterized and patients who are not catheterized and finds that catheterization increases patient mortality. We instead allow for such differences between patients by implementing both the instrumental variable bounds of Manski (1990), which only exploits an instrumental variable, and the bounds of Shaikh and Vytlacil (2011), which exploit mild nonparametric, structural assumptions in addition to an instrumental variable. We propose and justify the use of indicators of weekday admission as an instrument for catheterization in this context. We find that in our application, the Manski (1990) bounds do not indicate whether catheterization increases or decreases mortality, where as the Shaikh and Vytlacil (2011) bounds reveal that at least for some diagnoses, Swan–Ganz catheterization reduces mortality at 7 days after catheterization. We show that the bounds of Shaikh and Vytlacil (2011) remain valid under even weaker assumptions than those described in Shaikh and Vytlacil (2011). We also extend the analysis to exploit a further nonparametric, structural assumption–that doctors catheterize individuals with systematically worse latent health–and find that this assumption further narrows these bounds and strengthens our conclusions. In our analysis, we construct confidence regions using the methodology developed in Romano and Shaikh (2008). We show in particular that the confidence regions are uniformly consistent in level over a large class of possible distributions for the observed data that include distributions where the instrument is arbitrarily “weak”.     

Identifiability of mean‐reverting measurement error with instrumental variable

In the context where one main regressor is measured with error and at least one instrumental variable is available for the correction of measurement error, this paper provides, to the best of our knowledge, a first point‐identification result on the variance of measurement error, the variance of latent variable, and their covariance. We show that the parameters are identified if the regression model is not de facto linear. We illustrate the method in an application to identify mean‐reverting measurement error, a typical issue in reported income where the measurement error of income is negatively correlated with the true income.     

Spatial Decentralization and Programme Evaluation: Theory and an Example

This paper proposes an instrumental variable method for programme evaluation that only requires a single cross‐section of data on the spatial intensity of programmes and outcomes. The instruments are derived from a simple theoretical model of government decision‐making in which governments are responsive to the attributes of places, rather than to the attributes of individuals, in making allocation decisions across space, and have a social welfare function that is spatially weakly separable, that is, that the budgeting process behaves as if it is multi‐stage with respect to administrative districts and sub‐districts. The spatial instrumental variables model is then estimated and tested with a single cross‐section of Indonesian census data. The results offer support to the identification strategy proposed but also highlight some issues affecting validity.     

The effects of certification on total factor productivity: A propensity score matching approach

The primary purpose of this study is to investigate the impact of quality certification on total factor productivity, which is unobservable to and difficult to gauge for firms. To do so, we implement various propensity score matching estimators that handle the self‐selection problem. We use firm‐level data on manufacturing companies in Turkey where firms have devoted significant amounts of resources to get certified. Our findings show that certification does not have any positive effect on total factor productivity. We provide potential explanations for these findings.