Instrumental Variable Design Research Articles

Global capital markets, housing prices, and partisan fiscal policies

In recent years, global imbalances have channeled the excess savings of surplus countries toward the real estate markets of deficit countries. By consequence, the deficit countries that attracted lots of foreign capital experienced large run‐ups in house prices, whereas most surplus countries that exported capital exhibited flat or slow house price growth. We first use new house price data and a novel instrumental variable design to show the causal relationship between housing prices and capital inflows, particularly through debt bonanzas. We then argue that international capital flows affect the fiscal policy preferences of both voters and political parties by way of their impact on housing prices. Where capital inflows are large and housing prices are rising, we expect voters to respond by demanding both lower taxes and less publicly‐provided social insurance because rising house prices allow homeowners to self‐insure against income loss. In contrast, declining house prices produce greater demands for social insurance, particularly among those most exposed to housing market risk. We present evidence from two cross‐national surveys that supports these claims, as well as a “before and after” analysis of the housing crash in Eastern Europe. We also show that the connection between house prices and social policy also manifests itself in government spending outcomes, mediated by partisan control.

The Effect of Immigration from Mexico on Social Capital in the United States

Has mass migration from Mexico since the 1980s contributed to a well-documented decline in US social capital? Theories linking ethnic diversity to lower social cohesion and participation (e.g., Putnam 2007, 30, 137) would strongly predict this effect. Yet the impact of immigration in particular, rather than ethno-racial diversity generally, on US social capital has not been examined. Assessing the impact of immigration is important because some have speculated that associations between measures of diversity and social capital found in the United States are a byproduct of the country's distinctively fraught history of black–white relations. This scope condition would greatly limit the applicability of Putnam's thesis. To assess the impact of Mexican immigration, this study leverages a dynamic measure of social capital and an instrumental variables design. The results address an important recent methodological critique of the broader literature and strongly corroborate the hypothesis that immigration erodes social capital.

The performance of the computer software and services sector in Argentina: Microeconometric evidence on public-sector support programmes

This article analyses the impact of public-sector support programmes on the recent performance of the computer software and services sector in Argentina. First, the effect of these programmes on firms’ innovation performance is studied, with a propensity score matching technique used to calculate the average treatment effect on treated firms. The results confirm that receiving public funds had a positive impact on the ratio between research and development (R&D) spending and sales, employment in R&D and the propensity to introduce new products or processes. The effect of policy intervention on firms’ economic performance is then analysed, with an instrumental variables design being used in this case. The results show a positive impact on the propensity to export, export intensity and employment growth.

A general instrumental variable framework for regression analysis with outcome missing not at random.

The instrumental variable (IV) design is a well-known approach for unbiased evaluation of causal effects in the presence of unobserved confounding. In this article, we study the IV approach to account for selection bias in regression analysis with outcome missing not at random. In such a setting, a valid IV is a variable which (i) predicts the nonresponse process, and (ii) is independent of the outcome in the underlying population. We show that under the additional assumption (iii) that the IV is independent of the magnitude of selection bias due to nonresponse, the population regression in view is nonparametrically identified. For point estimation under (i)-(iii), we propose a simple complete-case analysis which modifies the regression of primary interest by carefully incorporating the IV to account for selection bias. The approach is developed for the identity, log and logit link functions. For inferences about the marginal mean of a binary outcome assuming (i) and (ii) only, we describe novel and approximately sharp bounds which unlike Robins-Manski bounds, are smooth in model parameters, therefore allowing for a straightforward approach to account for uncertainty due to sampling variability. These bounds provide a more honest account of uncertainty and allows one to assess the extent to which a violation of the key identifying condition (iii) might affect inferences. For illustration, the methods are used to account for selection bias induced by HIV testing nonparticipation in the evaluation of HIV prevalence in the Zambian Demographic and Health Surveys.

Intangible Capital and Leverage

This study investigates the causal effect of intangible capital on leverage. We use court invalidations of patents by the U.S. Court of Appeals for the Federal Circuit as a proxy for reductions in intangible capital and mitigate endogeneity concerns by exploiting random assignment of judges to court cases. Using an instrumental variable research design, we show that a loss of intangible capital causes a reduction in a firm’s leverage. The effect is stronger for firms with low interest coverage, high proximity to default and small firm size, and if an invalidated patent is pledged as collateral.

Quasi-Experimental Designs for Causal Inference

When randomized experiments are infeasible, quasi-experimental designs can be exploited to evaluate causal treatment effects. The strongest quasi-experimental designs for causal inference are regression discontinuity designs, instrumental variable designs, matching and propensity score designs, and comparative interrupted time series designs. This article introduces for each design the basic rationale, discusses the assumptions required for identifying a causal effect, outlines methods for estimating the effect, and highlights potential validity threats and strategies for dealing with them. Causal estimands and identification results are formalized with the potential outcomes notations of the Rubin causal model.

Do Independent Directors Monitor Financial Reporting? Evidence from Accounting Misstatements Following Stock Exchange Listing Requirements

A wave of accounting failures at the beginning of the century motivated the national stock exchanges to require firms to maintain a majority independent board of directors. Approximately 24% of firms were not already in compliance with the new listing requirements and so were forced to change. We investigate the causal effect of this forced increase in board independence (outside directors) on the probability of an accounting misstatement. Existing empirical studies find a negative relation between board independence and the probability of a misstatement (Beasley 1996, Dechow et al. 1996), but recent theory suggests that independent directors might be less effective monitors (Adams and Ferreira 2007, Harris and Raviv 2008). Using an instrumental variables design, we find that the forced increase in board independence increased the probability of misstatements. The findings suggest that the listing requirements had the opposite of their intended effect.

Instrumental variable estimation in a survival context.

Bias due to unobserved confounding can seldom be ruled out with certainty when estimating the causal effect of a nonrandomized treatment. The instrumental variable (IV) design offers, under certain assumptions, the opportunity to tame confounding bias, without directly observing all confounders. The IV approach is very well developed in the context of linear regression and also for certain generalized linear models with a nonlinear link function. However, IV methods are not as well developed for regression analysis with a censored survival outcome. In this article, we develop the IV approach for regression analysis in a survival context, primarily under an additive hazards model, for which we describe 2 simple methods for estimating causal effects. The first method is a straightforward 2-stage regression approach analogous to 2-stage least squares commonly used for IV analysis in linear regression. In this approach, the fitted value from a first-stage regression of the exposure on the IV is entered in place of the exposure in the second-stage hazard model to recover a valid estimate of the treatment effect of interest. The second method is a so-called control function approach, which entails adding to the additive hazards outcome model, the residual from a first-stage regression of the exposure on the IV. Formal conditions are given justifying each strategy, and the methods are illustrated in a novel application to a Mendelian randomization study to evaluate the effect of diabetes on mortality using data from the Health and Retirement Study. We also establish that analogous strategies can also be used under a proportional hazards model specification, provided the outcome is rare over the entire follow-up.

Graphical Models for Quasi-experimental Designs.

Randomized controlled trials (RCTs) and quasi-experimental designs like regression discontinuity (RD) designs, instrumental variable (IV) designs, and matching and propensity score (PS) designs are frequently used for inferring causal effects. It is well known that the features of these designs facilitate the identification of a causal estimand and, thus, warrant a causal interpretation of the estimated effect. In this article, we discuss and compare the identifying assumptions of quasi-experiments using causal graphs. The increasing complexity of the causal graphs as one switches from an RCT to RD, IV, or PS designs reveals that the assumptions become stronger as the researcher's control over treatment selection diminishes. We introduce limiting graphs for the RD design and conditional graphs for the latent subgroups of com-pliers, always takers, and never takers of the IV design, and argue that the PS is a collider that offsets confounding bias via collider bias.

The effectiveness of a computer‐assisted math learning program

AbstractComputer‐assisted instruction (CAI) programs are considered as a way to improve learning outcomes of students. However, little is known on the schools who implement such programs as well as on the effectiveness of similar information and communication technology programs. We provide a literature review that pays special attention to the existing causal evidence of computer‐assisted programs on learning outcomes. The paper relies on a rich data set consisting of (i) pupil‐level information on the use of a Dutch computer‐assisted program and (ii) detailed school‐level information on, among others, outcomes on national exams. The results suggest that schools with lower educational attainments use more frequently CAI programs. This suggests that they use CAI programs to catch up on learning outcomes. Moreover, using an instrumental variable design, we argue that given the participation in the CAI program, making more exercises leads to higher test results. Working with a CAI program seems therefore effective.

A note on the control function approach with an instrumental variable and a binary outcome.

Unobserved confounding is a well known threat to causal inference in non-experimental studies. The instrumental variable design can under certain conditions be used to recover an unbiased estimator of a treatment effect even if unobserved confounding cannot be ruled out with certainty. For continuous outcomes, two stage least squares is the most common instrumental variable estimator used in epidemiologic applications. For a rare binary outcome, an analogous linear-logistic two-stage procedure can be used. Alternatively, a control function approach is sometimes used which entails entering the residual from the first stage linear model as a covariate in a second stage logistic regression of the outcome on the treatment. Both strategies for binary response have previously formally been justified only for continuous exposure, which has impeded widespread use of the approach outside of this setting. In this note, we consider the important setting of binary exposure in the context of a binary outcome. We provide an alternative motivation for the control function approach which is appropriate for binary exposure, thus establishing simple conditions under which the approach may be used for instrumental variable estimation when the outcome is rare. In the proposed approach, the first stage regression involves a logistic model of the exposure conditional on the instrumental variable, and the second stage regression is a logistic regression of the outcome on the exposure adjusting for the first stage residual. In the event of a non-rare outcome, we recommend replacing the second stage logistic model with a risk ratio regression.

A Corporate Culture Channel: How Increased Shareholder Governance Reduces Firm Value

I show corporate culture is an important channel through which shareholder governance affects firm value. I develop a novel data set to measure aspects of corporate culture and use a regression discontinuity strategy to demonstrate stronger shareholder governance significantly changes aspects of culture. I find greater results-orientation but less customer-focus, integrity, and collaboration. Consistent with a positive link between governance and value, shareholders initially realize financial gains: increases in sales, profitability, and payout occur. However, over time, I find intangible assets associated with customer satisfaction and employee integrity deteriorate, which partly reverses the gains from greater results-orientation. These findings are consistent with a model of multitasking where stronger governance incentivizes managers to concentrate on easy-to-observe benchmarks at the expense of the harder-to-measure intangibles, even though such actions are not in the firm's best long-term interest. On average, I find firm value declines 1.4% through this corporate culture channel. I use an instrumental variable design and interventions by activist hedge funds to test the external validity of the inferences. Across these complementary research designs, I consistently find strong support for the economic importance of a corporate culture channel.

The Efficacy of Shareholder Voting: Evidence from Equity Compensation Plans

ABSTRACTThis study examines the effects of shareholder support for equity compensation plans on subsequent CEO compensation. Using cross‐sectional regression, instrumental variable, and regression discontinuity research designs, we find little evidence that either lower shareholder voting support for, or outright rejection of, proposed equity compensation plans leads to decreases in the level or composition of future CEO incentive compensation. We also find that, in cases where the equity compensation plan is rejected by shareholders, firms are more likely to propose, and shareholders are more likely to approve, a plan the following year. Our results suggest that shareholder votes for equity pay plans have little substantive impact on firms’ incentive compensation policies. Thus, recent regulatory efforts aimed at strengthening shareholder voting rights, particularly in the context of executive compensation, may have limited effect on firms’ compensation policies.

Receiver Design for Range and Doppler Sidelobe Suppression Using MIMO and Phased-Array Radar

This paper investigates the receiver instrumental variable (IV) filter design for sidelobe suppression using multiple-input-multiple-output (MIMO) and phased-array radar. First, we analyze the performance of existing integrated sidelobe level (ISL) and the zero sidelobe (ZS) methods for range sidelobe suppression, and show the latter's superiority. The limitation of MIMO radar sidelobe suppression in a large number of range bins is discussed. It is shown that the phased-array radar avoids this limitation. Then we consider joint range and Doppler sidelobe suppression, and the solutions for both MIMO and phased-array radar are derived, classified, and analyzed. The conclusions are drawn through performance trade-off. The use of phased-array radar compromises the waveform diversity of MIMO radar. However, with the objective of range, and joint range and Doppler sidelobe suppression, phased-array radar has better performance in terms of computational complexity, sidelobe suppression level and signal-to-noise ratio (SNR) loss. This will relieve the design workload for large scale radar systems. In addition, we show via simulations that the bounds of clear area of matched filter output (ambiguity function) are inapplicable for the IV filter design.

The Effects of Immigrant Concentration on Changes in Neighborhood Crime Rates

This study investigated the extent to which immigrant concentration is associated with reductions in neighborhood crime rates in the City of Los Angeles. A potential outcomes model using two-stage least squares regression was estimated, where immigrant concentration levels in 1990 were used as an instrumental variable to predict immigrant concentration levels in 2000. The instrumental variables design was used to reduce selection bias in estimating the effect of immigrant concentration on changes in official crime rates between 2000 and 2005 for census tracts in the City of Los Angeles, holding constant other demographic variables and area-level fixed effects. Non-parametric smoothers were also employed in a two-stage least squares regression model to control for the potential influence of heterogeneity in immigrant concentration on changes in crime rates. The results indicate that greater predicted concentrations of immigrants in neighborhoods are linked to significant reductions in crime. The results are robust to a number of different model specifications. The findings challenge traditional ecological perspectives that link immigrant settlement to higher rates of crime. Immigration settlement patterns appear to be associated with reducing the social burden of crime. Study conclusions are limited by the potential for omitted variables that may bias the observed relationship between immigrant concentration and neighborhood crime rates, and the use of only official crime data which may under report crimes committed against immigrants. Understanding whether immigrant concentration is an important dynamic of changing neighborhood patterns of crime outside Los Angeles will require replication with data from other U.S. cities.

THE BENEFITS OF MIGRATION

ABSTRACT Assuming that migration and integration have lasting effects on economic and social processes, this paper analyses if a diverse cultural environment has a positive impact on labour market success. We expect that culturally diverse labour markets provide new opportunities through an open and tolerant climate, thus contributing to overall economic growth. We test this assumption by analysing the successful transition from education to work depending on the regional distribution and ethnic mix of the foreign population in Germany (cultural diversity). To account for variation within Germany, cultural diversity is observed at small administrative units. We analyse a cohort of young adults at the time of the successful completion of their apprenticeship in Germany and follow them through the beginning of their career. The concentration on a homogenous group regarding occupational certificates enables us to focus on the effects of the local and social environment as well as individual characteristics such as their national background on finding a job (or not). We apply an instrumental variable design to disentangle the effects of cultural diversity and share of foreigners. The results show that both young foreigners and Germans face significantly lower barriers for employment entry in culturally more diverse German regions.

The Efficacy of Shareholder Voting: Evidence from Equity Compensation Plans

This study examines the effects of shareholder support for equity compensation plans on subsequent chief executive officer (CEO) compensation. Using cross-sectional regression, instrumental variable, and regression discontinuity research designs, we find little evidence that either lower shareholder voting support for, or outright rejection of, proposed equity compensation plans leads to decreases in the level or composition of future CEO incentive-compensation. We also find that in cases where the equity compensation plan is rejected by shareholders, firms are more likely to propose, and shareholders are more likely to approve, a plan the following year. Our results suggest that shareholder votes have little substantive impact on firms’ incentive-compensation policies. Thus, recent regulatory efforts aimed at strengthening shareholder voting rights, particularly in the context of executive compensation, may have limited effect on firms’ compensation policies.

Designs Combining Instrumental Variables with Case-Control: Estimating Principal Strata Causal Effects

The instrumental variables framework is commonly used for the estimation of causal effects from cohort samples. However, the combination of instrumental variables with more efficient designs such as case-control sampling requires new methodological consideration. For example, as the use of Mendelian randomization studies is increasing and the cost of genotyping and gene expression data can be high, the analysis of data gathered from more cost-effective sampling designs is of prime interest. We show that the standard instrumental variables analysis does not appropriately estimate the causal effects of interest when the instrumental variables design is combined with the case-control design. We also propose a method that can estimate the causal effects in such combined designs. We illustrate the method with a study in oncology.

The Efficacy of Shareholder Voting: Evidence from Equity Compensation Plans

This study examines the effects of shareholder support for equity compensation plans on subsequent chief executive officer (CEO) compensation. Using cross-sectional regression, instrumental variable, and regression discontinuity research designs, we find little evidence that either lower shareholder voting support for, or outright rejection of, proposed equity compensation plans leads to decreases in the level or composition of future CEO incentive-compensation. We also find that in cases where the equity compensation plan is rejected by shareholders, firms are more likely to propose, and shareholders are more likely to approve, a plan the following year. Our results suggest that shareholder votes have little substantive impact on firms' incentive-compensation policies. Thus, recent regulatory efforts aimed at strengthening shareholder voting rights, particularly in the context of executive compensation, may have limited effect on firms' compensation policies.