Bayesian Model Averaging in the Instrumental Variable Regression Model

This paper considers the instrumental variable regression model when there is uncertainty about the set of instruments, exogeneity restrictions, the validity of identifying restrictions and the set of exogenous regressors. This uncertainty can result in a huge number of models. To avoid statistical problems associated with standard model selection procedures, we develop a reversible jump Markov chain Monte Carlo algorithm that allows us to do Bayesian model averaging. The algorithm is very exible and can be easily adapted to analyze any of the di¤erent priors that have been proposed in the Bayesian instrumental variables literature. We show how to calculate the probability of any relevant restriction (e.g. the posterior probability that over-identifying restrictions hold) and discuss diagnostic checking using the posterior distribution of discrepancy vectors. We illustrate our methods in a returns-to-schooling application.

The Good, the Bad and the Populist: A Model of Political Agency with Emotional Voters

This paper attempts to extend existing models of political agency to an environment in which voting may be divided between informed and instrumental, informed and ‘expressive’ (Brennan and Lomasky (1993)) and uninformed due to ‘rational irrationality’ (Caplan (2007)). It constructs a model where politicians may be good, bad or populist. Populists are more willing than good politicians to pander to voters who may choose inferior policies in a large-group electoral setting because their vote is insignificant compared with those that voters would choose were their vote decisive in determining the electoral outcome. Bad politicians would ideally like to extract tax revenue for their own ends. Initially we assume the existence of only good and populist politicians. The paper investigates the incentives for good politicians to pool with or separate from populists and focuses on three key issues – (1) how far the majority of voter’s preferences are from those held by the better informed incumbent politician (2) the extent to which the population exhibits rational irrationality and expressiveness (jointly labelled as emotional) and (3) the cost involved in persuading uninformed voters to change their views in terms of composing messages and spreading them. This paper goes on to consider how the inclusion of bad politicians may affect the behaviour of good politicians and suggests that a small amount of potential corruption may be socially useful. It is also argued that where bad politicians have an incentive to mimic the behaviour of good and populist politicians, the latter types of politician may have an incentive to separate from bad politicians by investing in costly public education signals. The paper also discusses the implications of the model for whether fiscal restraints should be soft or hard.

Expressive Political Behaviour: Foundations, Scope and Implications

A growing literature has focussed attention on ‘expressive’ rather than ‘instrumental’ behaviour in political settings - particularly voting A common criticism of the expressive idea is that its myriad possibilities make it rather ad hoc and lacking in both predictive and normative bite. We agree that no single clear definition of expressive behaviour has emerged to date, and no detailed foundations of specific expressive motivations have been provided, so that there are rather few specific implications drawn from the analysis of expressive behaviour. In response, we provide a foundational discussion and definition of expressive behaviour that accounts for a range of factors. We also discuss the content of expressive choice distinguishing between moral, social and emotional cases, and relate this more general account to the specific theories of expressive choice in the literature. Finally, we discuss the normative and institutional implications of expressive behaviour.

Colonialism, European Descendants and Democracy

This paper advances that the share of European descendants in the population is a major determinant of democracy in former colonial countries. We test this hypothesis using cross-section and panel regressions with 60 developing and developed countries that were once colonies. We find that the share of European descendants can explain more than half of the difference in measures of democracy between the least and the most democratic countries in our sample. We control for other potential determinants of democracy and test for endogeneity bias using instrumental variables.

(Don’t) Make My Vote Count

Proponents of proportional electoral rules often argue that majority rule depresses turnout and may lower welfare due to the 'tyranny of the majority' problem. The present paper studies the impact of electoral rules on turnout and social welfare. We analyze a model of instrumental voting where citizens have private information over their individual cost of voting and over the alternative they prefer. The electoral rule used to select the winning alternative is a combination of majority rule and proportional rule. Results show that the above arguments against majority rule do not hold in this set up. Social welfare and turnout increase with the weight that the electoral rule gives to majority rule when the electorate is expected to be split, and they are independent of the electoral rule employed when the expected size of the minority group tends to zero. However, more proportional rules can increase turnout within the minority group. This effect is stronger the smaller the minority group. We then conclude that majority rule fosters overall turnout and increases social welfare, whereas proportional rule fosters the participation of minorities.

The Difference, System and ‘Double-D’ GMM Panel Estimators in the Presence of Structural Breaks

The effects of structural breaks in dynamic panels are more complicated than in time series models as the bias can be either negative or positive. This paper focuses on the effects of mean shifts in otherwise stationary processes within an instrumental variable panel estimation framework. We show the sources of the bias and a Monte Carlo analysis calibrated on United States bank lending data demonstrates the size of the bias for a range of auto-regressive parameters. We also propose additional moment conditions that can be used to reduce the biases caused by shifts in the mean of the data.

Chasing Graduate Jobs?

This paper examines empirically the relationship between under-employment and migration amongst five cohorts of graduates of Scottish higher education institutions with micro-data collected by the Higher Education Statistical Agency. The data indicate that there is a strong positive relationship between migration and graduate employment—those graduates who move after graduation from Scotland to the rest of the UK or abroad have a much higher rate of graduate employment. Versions of probit regression are used to estimate migration and graduate employment equations in order to explore the nature of this relationship further. These equations confirm that there is a strong positive relationship between the probability of migrating and the probability of being in graduate employment even after other factors are controlled for. Instrumental variables estimation is used to examine the causal nature of the relationship by attempting to deal with the potential endogeneity of migration decisions. Overall the analysis is consistent with the hypotheses that a sizeable fraction of higher education graduates are leaving Scotland for employment reasons. In turn this finding suggests the over-education/under-employment nexus is a serious problem in Scotland.

A Consistent Nonparametric Bootstrap Test of Exogeneity

This paper proposes a novel way of testing exogeneity of an explanatory variable without any parametric assumptions in the presence of a "conditional" instrumental variable. A testable implication is derived that if an explanatory variable is endogenous, the conditional distribution of the outcome given the endogenous variable is not independent of its instrumental variable(s). The test rejects the null hypothesis with probability one if the explanatory variable is endogenous and it detects alternatives converging to the null at a rate n..1=2:We propose a consistent nonparametric bootstrap test to implement this testable implication. We show that the proposed bootstrap test can be asymptotically justi.ed in the sense that it produces asymptotically correct size under the null of exogeneity, and it has unit power asymptotically. Our nonparametric test can be applied to the cases in which the outcome is generated by an additively non-separable structural relation or in which the outcome is discrete, which has not been studied in the literature.

Coordination in Public Good Provision: How Individual Volunteering is Impacted by the Volunteering of Others

In this analysis, we examine the relationship between an individual's decision to volunteer and the average level of volunteering in the community where the individual resides. Our theoretical model is based on a coordination game , in which volunteering by others is informative regarding the benefit from volunteering. We demonstrate that the interaction between this information and one's private information makes it more likely that he or she will volunteer, given a higher level of contributions by his or her peers. We complement this theoretical work with an empirical analysis using Census 2000 Summary File 3 and Current Population Survey (CPS) 2004-2007 September supplement file data. We control for various individual and community characteristics, and employ robustness checks to verify the results of the baseline analysis. We additionally use an innovative instrumental variables strategy to account for reflection bias and endogeneity caused by selective sorting by individuals into neighborhoods, which allows us to argue for a causal interpretation. The empirical results in the baseline, as well as all robustness analyses, verify the main result of our theoretical model, and we employ a more general structure to further strengthen our results.

Coordination in Public Good Provision: How Individual Volunteering is Impacted by the Volunteering of Others

In this analysis, we examine the relationship between an individual’s decision to volunteer and the average level of volunteering in the community where the individual resides. Our theoretical model is based on a coordination game , in which volunteering by others is informative regarding the benefit from volunteering. We demonstrate that the interaction between this information and one’s private information makes it more likely that he or she will volunteer, given a higher level of contributions by his or her peers. We complement this theoretical work with an empirical analysis using Census 2000 Summary File 3 and Current Population Survey (CPS) 2004-2007 September supplement file data. We control for various individual and community characteristics, and employ robustness checks to verify the results of the baseline analysis. We additionally use an innovative instrumental variables strategy to account for reflection bias and endogeneity caused by selective sorting by individuals into neighbourhoods, which allows us to argue for a causal interpretation. The empirical results in the baseline, as well as all robustness analyses, verify the main result of our theoretical model, and we employ a more general structure to further strengthen our results.