Identification Of Causal Effects Research Articles

How income and tax rates provoke cheating – An experimental investigation of tax morale

Anecdotal evidence suggests that tax morale diminishes with income and with levels of taxation. We designed an experiment that enables identification of causal effects. Subjects carry out a sequence of real effort tasks. The income earned varies with individual differences in effort (combined with variances related to skill or luck) and with the exogenously given length of the tasks. For each task, subjects privately roll a die, whose value determines the tax rate, which is then reported by subjects. This provides subjects with an incentive to cheat. Tax morale diminishes with higher effort, which might find ethical justification, but also with longer tasks, which would not. We implement treatments that vary the range of taxation. Contrary to widespread belief, participants’ tax morale is invariant to these treatments. Our findings are best explained by a psychological force that tempts rich people to cheat more. This force does not seem to be related to fairness ideals that are prominent in theories of distributional justice nor to absolute levels of taxation.

CSR Disclosure and Investor Behavior: A Proposed Framework and Research Agenda

ABSTRACTHow market participants respond to corporate disclosure forms an important cornerstone in many areas of accounting and finance research. This article synthesizes behavioral research on how an increasingly important type of corporate disclosure, namely corporate social responsibility (CSR) disclosure, affects investor behavior. Structuring the extant accounting and finance literature, we derive a holistic framework that consolidates observed drivers of investors' information processing and resulting investment decision making when confronted with CSR disclosure. We identify both disclosure and individual investor characteristics that determine investor behavioral response to disclosed CSR-related information. Drawing on the proposed framework, we pinpoint knowledge gaps in the literature and provide guidelines for further research. Fundamental issues for future work include the decoupling of specific behavioral drivers through innovative measurement, the reliable identification of causal effects, and the incorporation of investors' social interactions.

Weighted estimation for confounded binary outcomes subject to misclassification.

In the presence of confounding, the consistency assumption required for identification of causal effects may be violated due to misclassification of the outcome variable. We introduce an inverse probability weighted approach to rebalance covariates across treatment groups while mitigating the influence of differential misclassification bias. First, using a simplified example taken from an administrative health care dataset, we introduce the approach for estimation of the marginal causal odds ratio in a simple setting with the use of internal validation information. We then extend this to the presence of additional covariates and use simulated data to investigate the finite sample properties of the proposed weighted estimators. Estimation of the weights is done using logistic regression with misclassified outcomes, and a bootstrap approach is used for variance estimation.

Identification in Macroeconomics

This paper discusses empirical approaches macroeconomists use to answer questions like: What does monetary policy do? How large are the effects of fiscal stimulus? What caused the Great Recession? Why do some countries grow faster than others? Identification of causal effects plays two roles in this process. In certain cases, progress can be made using the direct approach of identifying plausibly exogenous variation in a policy and using this variation to assess the effect of the policy. However, external validity concerns limit what can be learned in this way. Carefully identified causal effects estimates can also be used as moments in a structural moment matching exercise. We use the term “identified moments” as a short-hand for “estimates of responses to identified structural shocks,” or what applied microeconomists would call “causal effects.” We argue that such identified moments are often powerful diagnostic tools for distinguishing between important classes of models (and thereby learning about the effects of policy). To illustrate these notions we discuss the growing use of cross-sectional evidence in macroeconomics and consider what the best existing evidence is on the effects of monetary policy.

Identification and generalizability in accounting research: A discussion of Christensen, Floyd, Liu, and Maffett (2017)

Christensen et al. (2017) provide evidence that the dissemination of mine safety information in SEC filings has real effects on mine safety. We discuss the extent to which Christensen et al.’s results generalize to a research question that we consider of broader interest to accounting researchers, specifically where and when mandated disclosure in SEC filings can increase the dissemination of information. We also discuss identification of causal effects and generalizability concerns more broadly in the context of large sample studies and quasi-natural experiments, as well as potential ways authors might address these concerns in accounting research.

Identification and Generalizability in Accounting Research: A Discussion of Christensen, Floyd, Liu, and Maffett (2017)

Christensen et al. (2017) provide evidence that the dissemination of mine safety information in SEC filings has real effects on mine safety. We discuss the extent to which Christensen et al.’s results generalize to a research question that we consider of broader interest to accounting researchers, specifically where and when mandated disclosure in SEC filings can increase the dissemination of information. We also discuss identification of causal effects and generalizability concerns more broadly in the context of large sample studies and quasi-natural experiments, as well as potential ways authors might address these concerns in accounting research.

Behind the Scenes: International Evidence on the Monitoring Role of the Norwegian Pension Fund

This paper revisits the monitoring role of international institutional ownership. Based on a novel setting that improves the identification of causal effects, it examines whether the presence of institutional ownership improves firm-level monitoring and in turn increases firm value. In essence, the paper tracks in detail the investment positions of the Norwegian Pension Fund between 2010 and 2014 and exploits the funds’ entries (initial investment in firms) as well as exits (subsequent divesture from firms) in a two-fold differences-in-differences (DiD) design with two-natural control groups and a comprehensive fixed-effect structure. The respective DiD results document significantly positive monitoring as well as performance effects in the course of the funds’ entries. Corroborating these results, exits of the funds are subsequently followed by significant declines in monitoring and performance. A variety of sensitivity tests, for example, addressing parallel trends assumptions or reverse causality concerns, ensure the validity of these results. Overall, the paper’s findings highlight the beneficial role of institutional ownership.

Causal effect identification in acyclic directed mixed graphs and gated models

We introduce a new family of graphical models that consists of graphs with possibly directed, undirected and bidirected edges but without directed cycles. We show that these models are suitable for representing causal models with additive error terms. We provide a set of sufficient graphical criteria for the identification of arbitrary causal effects when the new models contain directed and undirected edges but no bidirected edge. We also provide a necessary and sufficient graphical criterion for the identification of the causal effect of a single variable on the rest of the variables. Moreover, we develop an exact algorithm for learning the new models from observational and interventional data via answer set programming. Finally, we introduce gated models for causal effect identification, a new family of graphical models that exploits context specific independences to identify additional causal effects.

Ignorability for general longitudinal data.

SummaryLikelihood factors that can be disregarded for inference are termed ignorable. We demonstrate that close ties exist between ignorability and identification of causal effects by covariate adjustment. A graphical condition, stability, plays a role analogous to that of missingness at random, but is applicable to general longitudinal data. Our formulation of ignorability does not depend on any notion of missing data, so is appealing in situations where missing data may not actually exist. Several examples illustrate how stability may be assessed.

Invited Commentary: Bias Attenuation and Identification of Causal Effects With Multiple Negative Controls.

In this commentary, we describe several extensions to the interesting and important negative control exposure approach for partial confounding adjustment in time-series analysis proposed by Flanders et al. (Am J Epidemiol. 2017;185(10):941-949). Specifically, by leveraging the availability of exposure time series, we show that under certain additional fairly reasonable assumptions, one can incorporate both past and future exposures as multiple negative control exposures to further attenuate confounding bias. We further describe 2 specific settings in which multiple controls can be used to fully account for confounding bias; the first assumes a forward-in-time version of the familiar autoregressive model for the exposure time series, while the second combines a negative control exposure with a negative control outcome for joint indirect adjustment of confounding. We briefly illustrate how one might apply our proposed framework in time-series studies. Both the original method of Flanders et al. and our proposed extensions are particularly well-suited for time-series data such as the air pollution study considered in their paper, and as such should be considered in routine environmental health studies.

Identification of the driving forces of climate change using the longest instrumental temperature record

The identification of causal effects is a fundamental problem in climate change research. Here, a new perspective on climate change causality is presented using the central England temperature (CET) dataset, the longest instrumental temperature record, and a combination of slow feature analysis and wavelet analysis. The driving forces of climate change were investigated and the results showed two independent degrees of freedom —a 3.36-year cycle and a 22.6-year cycle, which seem to be connected to the El Niño–Southern Oscillation cycle and the Hale sunspot cycle, respectively. Moreover, these driving forces were modulated in amplitude by signals with millennial timescales.

Causal Effect Identification by Adjustment under Confounding and Selection Biases

Controlling for selection and confounding biases are two of the most challenging problems in the empirical sciences as well as in artificial intelligence tasks. Covariate adjustment (or, Backdoor Adjustment) is the most pervasive technique used for controlling confounding bias, but the same is oblivious to issues of sampling selection. In this paper, we introduce a generalized version of covariate adjustment that simultaneously controls for both confounding and selection biases. We first derive a sufficient and necessary condition for recovering causal effects using covariate adjustment from an observational distribution collected under preferential selection. We then relax this setting to consider cases when additional, unbiased measurements over a set of covariates are available for use (e.g., the age and gender distribution obtained from census data). Finally, we present a complete algorithm with polynomial delay to find all sets of admissible covariates for adjustment when confounding and selection biases are simultaneously present and unbiased data is available.

Public Funding and Corporate Innovation

We review and condense the body of literature on the economic returns of public RD (ii) private RD (iii) RD (v) the mobility of RD (vi) there are many university spin-offs but these are no more successful than non-university spin-offs; (vii) universities constitute important collaboration partners and (viii) clusters enhance collaboration, patents and productivity. Key problems for economic policy advice are that the identification of causal effects is problematic in most studies and that little is known about the optimal design of policy measures.

A Beginner’s Guide to Randomized Evaluations in Development Economics

I. IntroductionRandomized evaluations in development economics are receiving considerable attention; there has been a dramatic increase in randomized evaluations carried out not only by academia but also by governments, international agencies, and non-governmental organizations to test the effectiveness of a specific program or policy (Banerjee, and Duflo 2009).1 The rise in randomized evaluations is in line with the credibility revolution in empirical economics which emphasizes the identification of causal effects (Angrist, and Pischke 2010). Previously, sensitivity analysis, where researchers presented the rigorousness of their results with different specifications or functional forms had been considered a salutary econometric practice (Leamer 1985). However, the sensitivity analysis was limited in identifying the causal effect of a program because it is difficult to know the difference between the outcome a participant experienced under the program and the experienced outcome if the participant did not take part in the program. Although the causal effect of a program for each person is not identifiable (this is the fundamental problem of causal inference), a randomized evaluation makes it possible to create grouplevel counterfactual which is free from selection bias, and to provide internally valid estimates of the average treatment effect (Rubin 1974).In this paper we provide a brief overview of the important steps in planning and implementing a randomized evaluation from start to finish including issues such as research question development, local partner institution, IRB (Institutional Review Board), PAP (Pre-Analysis Plan), sample size calculation, random sampling, randomization, survey implementation, data entry, and data analysis.2II. Research Question DevelopmentNot all research questions need a randomized evaluation to be answered, thus, it is important to ask the right research question that requires a randomized evaluation. If we are interested to know what the needs are in a specific context, then descriptive need assessment can be enough. However, if we want to know whether a program or policy works or not, then we need to conduct a randomized evaluation to answer this question. Furthermore, randomized evaluations should be performed if we aim to answer one of the following research questions. 1) Which elements of the program matter the most? 2) Which of the two different programs produce a better outcome? 3) Are there complementarities among the programs? 4) Can the results from one context be replicated in another context?A good place to start developing research questions is with existing literature reviews. We can quickly summarize the information in existing research regarding issues of interest. The information can also help identify the most important knowledge gaps, where new research questions begin. High quality literature review on randomized evaluations can be found within institutions such as Jameel Poverty Action Lab (J-PAL) at MIT, Innovations for Poverty Action (IPA), the International Initiative for Impact Evaluation (3ie), and the Development Impact Evaluation Initiative (DIME) in the Research Group of the World Bank.3 Many academic journals in economics, which are all reliable resources, also publish up-to-date literature reviews. These journals include The Handbook of Economics series, the Annual Review of Economics, the Journal of Economic Perspectives, and the Journal of Economic Literature.After developing research questions for a randomized evaluation, the next step is to incorporate the questions into the survey questionnaire to analyze the research hypotheses and its possible mechanisms. It is useful to review the existing survey questionnaire of related research, which is often available on its author's personal webpage or on the publication journal's website.4 Another good source for questionnaire development is the government's or multilateral organization's survey questionnaire such as Demographic and Health Surveys (DHS). …

Proxy variables and nonparametric identification of causal effects

Proxy variables are often used in linear regression models with the aim of removing potential confounding bias. In this paper we formalise proxy variables within the potential outcomes framework, giving conditions under which it can be shown that causal effects are nonparametrically identified. We characterise two types of proxy variables and give concrete examples where the proxy conditions introduced may hold by design.

Causal mechanisms in organization and innovation studies

We outline the guiding ideas behind mechanisms-based theorizing in analytical sociology as a fruitful alternative to economics-inspired research on identification of causal effects, and discuss the potential of mechanisms-based theorizing for further development in organization and innovation studies. We discuss the realist stance on providing broader explanations as an identifying characteristic of the mechanism approach, its focus on the dynamic processes through which outcomes to be explained are brought about, and outline theoretical and methodological implications for organization and innovation studies.

Identifiability and transportability in dynamic causal networks

In this paper, we propose a causal analog to the purely observational dynamic Bayesian networks, which we call dynamic causal networks. We provide a sound and complete algorithm for the identification of causal effects in dynamic causal networks, namely for computing the effect of an intervention or experiment given a dynamic causal network and probability distributions of passive observations of its variables, whenever possible. We note the existence of two types of hidden confounder variables that affect in substantially different ways the identification procedures, a distinction with no analog in either dynamic Bayesian networks or standard causal graphs. We further propose a procedure for the transportability of causal effects in dynamic causal network settings, where the result of causal experiments in a source domain may be used for the identification of causal effects in a target domain.

Does choosing jobs based on income risk lead to higher job satisfaction in the long run? Evidence from the natural experiment of German reunification

Workers make job choices based in part on their income risk preferences. But how important is risk matching for long-run job satisfaction? Empirical research has been limited because job choices are endogenous, hindering identification of causal effects. We circumvent these problems by using data from the natural experiment of German Reunification. Pre-reunification East Germans chose jobs in a socialist environment without market income risk. These jobs were then exposed to market risks post-reunification, providing quasi-experimental variation. We find that workers in pre-reunification jobs report relatively higher job satisfaction, despite being less well-matched in terms of risk preferences. This unexpected result is robust to alternate explanations which might affect job satisfaction such as potential differences in job expectations, self-selection into leaving pre-reunification jobs, and past unemployment spells. We conjecture the results are consistent with projection bias: individuals at the start of their careers may over-estimate the extent to which risk matching matters for their future job satisfaction.

Viewpoint: Estimating the causal effects of policies and programs

AbstractEstimation, inference and interpretation of the causal effects of programs and policies have all advanced dramatically over the past 25 years. We highlight three particularly important intellectual trends: an improved appreciation of the substantive importance of heterogeneous responses and of their methodological implications, a stronger focus on internal validity brought about by the “credibility revolution,” and the scientific value that follows from grounding estimation and interpretation in economic theory. We discuss a menu of commonly employed partial equilibrium approaches to the identification of causal effects, emphasizing that the researcher's central intellectual contribution always consists of making an explicit case for a specific causal interpretation given the relevant economic theory, the data, the institutional context and the economic question of interest. We also touch on the importance of general equilibrium effects and full cost–benefit analyses.

Ballot order effects in direct democracy elections

Many political practitioners believe that voters are more likely to approve propositions listed at the top than the bottom of the ballot, potentially distorting democratic decision making, and this belief influences election laws across the United States. Numerous studies have investigated ballot order effects in candidate elections, but there is little evidence for direct democracy elections, and identification of causal effects is challenging. This paper offers two strategies for identifying the effect of ballot order in proposition elections, using data from California during 1958–2014 and Texas during 1986–2015. The evidence suggests that propositions are not advantaged by being listed at the top compared to the bottom of the ballot. Approval rates are lower with more propositions on the ballot.