Use Of Instrumental Variables Research Articles

Absenteeism and peer interaction effects: Evidence from an Italian Public Institute

Using microdata for a sample of about 350 workers employed by an Italian public institute, we explain individual absence rates considering both variables that may be related to health conditions and variables that may suggest shirking behaviour. Among these variables we especially focus our attention on the influence produced by the behaviour of randomly assigned peers. In order to handle reflection problems, we combine random assignment with the use of instrumental variables. The proportion of females in the peer group is used as an instrument of peer absence behaviour. From Two-Stage Least Square estimates it emerges that social and group interaction plays an important role in shaping individual absence behaviour.

Blogs and Local-market Movie Box-office Performance

We measure the effects of pre- and post-release blog volume, blog valence and advertising on the performance of 75 movies in 208 geographic markets of the U.S. We attribute the variation in blog effects across markets to differences in demographic characteristics of markets combined with differences across demographic groups in their access and exposure to blogs as well as their responsiveness conditional on access. We study the effects of pre-release factors on opening day box-office performance and of pre- and post-release factors on box-office performance one month after release. Our estimation accounts for confounding factors in the measurement of these effects via the use of instrumental variables. We find considerable heterogeneity in the effects across consumer and firm generated media and across geographic markets, with gender, income, race and age driving across-market differences. Release day performance is impacted most by pre-release blog volume and advertising, whereas post-release performance is influenced by post-release blog valence and advertising. Across markets, there is more variance in advertising and blog valence (post-release) elasticities than there is in blog volume (pre-release) elasticities. We identify the top 20 markets in terms of their elasticities to each of these 3 instruments. Further, we classify markets in terms of their sensitivities across these 3 instruments to identify the most sensitive markets that studios can target with their limited release strategies. Finally, we characterize the extent to which studios could have improved their limited release strategies by identifying the overlap between the actual release markets and the most responsive ones. At the time of first release, we find that studios cover only 53% of the most responsive advertising markets and 44% of the most responsive markets to pre-release blog volume in their limited release strategies, implying considerable room for improvement if these were the only metrics to assess those strategies.

On the use of instrumental variables in accounting research

Instrumental variable (IV) methods are commonly used in accounting research (e.g., earnings management, corporate governance, executive compensation, and disclosure research) when the regressor variables are endogenous. While IV estimation is the standard textbook solution to mitigating endogeneity problems, the appropriateness of IV methods in typical accounting research settings is not obvious. Drawing on recent advances in statistics and econometrics, we identify conditions under which IV methods are preferred to OLS estimates and propose a series of tests for research studies employing IV methods. We illustrate these ideas by examining the relation between corporate disclosure and the cost of capital.

Calculating the concentration index when income is grouped

The problem introduced by grouping income data when measuring socioeconomic inequalities in health (and health care) has been highlighted in a recent study in this journal. We re-examine this issue and show there is a tendency to underestimate the concentration index at an increasing rate when lowering the number of income categories. This tendency arises due to a form of measurement error and we propose two correction methods. Firstly, the use of instrumental variables (IV) can reduce the error within income categories. Secondly, through a simple formula for correction that is based only on the number of groups. We find that the simple correction formula reduces the impact of grouping and always outperforms the IV approach. Use of this correction can substantially improve comparisons of the concentration index both across countries and across time.

Correcting the Bias in the Concentration Index When Income is Grouped

The problem introduced by grouping income data when measuring socioeconomic inequalities in health (and health care) has been highlighted in a recent study in this journal. We re(examine this issue and show there is a tendency to underestimate the concentration index at an increasing rate when lowering the number of income categories. This bias results from a form of measurement error and we propose two correction methods. Firstly, the use of instrumental variables (IV) can reduce the error within income categories. Secondly, through a simple formula for correction that is based only on the number of groups. We compare the performance of these methods using data from 15 European countries and the United States. We find that the simple correction formula reduces the impact of grouping and always outperforms the IV approach. Use of this correction can substantially improve comparisons of the concentration index both across countries and across time.

Armed Conflict Exposure, Human Capital Investments and Child Labor: Evidence from Colombia

Using a unique combination of household and violence data sets and a duration analysis methodology, this paper estimates the effect that exposure to armed conflict has on school drop-out decisions of Colombian children between the ages of six and seventeen. After taking into account the possible endogeneity of municipal conflict related events through the use of instrumental variables, we find that armed conflict reduces the average years of schooling in 8.78% for all Colombian children. This estimate increases to 17.03% for children between sixteen and seventeen years old. We provide evidence that such effect may be induced mainly through higher mortality risks, and to lesser extent due to negative economic shocks and lower school quality; all of which induce a trade-off between schooling and child labor.

Foreign Debt and Domestic Savings in Developing Countries

ABSTRACT This paper addresses the question of potential causality between foreign debt and domestic savings within the context of developing countries. The literature provides evidence insofar as foreign debt and development is concerned, but little attention has been given so far to internal potential for capital formation. We provide a theoretical framework and test its relevance by using 1975–2004 data for two groups of countries: sub-Saharan Africa and Latin America with the Caribbean. With the use of instrumental variables we find that foreign debt impacts negatively on domestic savings, especially in the long run. The results are not susceptible to the choice of countries or outliers. However, the relationship between foreign debt and savings seems to depend on debt accumulation gaining significance only after surpassing a country's specific threshold. RESUMEN. Este estudio aborda el tema de la causalidad potencial que existe entre la deuda externa y el ahorro interno en el ámbito de los países en desarrollo. La literatura existente contiene pruebas sobre la deuda externa y el desarrollo pero, hasta el momento, ha dedicado poca atención a todo lo concerniente al potencial doméstico para la formación de capital (capitalización). Hemos trazado un marco teórico y puesto a prueba su relevancia usando datos del período 1975–2004 para dos grupos de países: África subsahariana y Latinoamérica incluyendo El Caribe. Los instrumentos variables utilizados nos permitieron encontrar el impacto negativo que la deuda externa ejerce sobre el ahorro interno, especialmente a largo plazo. Estos resultados no son susceptibles a la selección de países específicos o aislados. No obstante, la relación entre la deuda externa y el ahorro parece depender de la acumulación de la deuda, adquiriendo más significado solo después de pasar el umbral específico de un país. RESUMO. Este estudo aborda a questão da causalidade possível entre o débito estrangeiro e a poupança interna no contexto dos países em desenvolvimento. A literatura apresenta evidências sobre débito externo e desenvolvimento, mas há pouca informação em relação ao potencial interno para a formação de capital. Mostramos uma estrutura teórica e testamos a sua relevância, utilizando dados de 1975 até 2004 para dois grupos de países: a África Subsaariana e a América Latina com o Caribe. Com a ajuda das variáveis instrumentais, descobrimos um impacto negativo do débito estrangeiro sobre a poupança interna, especialmente a longo prazo. Os resultados não são suscetíveis à escolha de países ou de dados atípicos (outliers). A relação entre débito externo e poupança, entretanto, parece depender do acúmulo do débito, tornando-se importante, apenas no caso de ultrapassar um limiar específico do país.

Better LATE Than Nothing: Some Comments on Deaton (2009) and Heckman and Urzua (2009)

Two recent papers, Deaton (2009) and Heckman and Urzua (2009), argue against what they see as an excessive and inappropriate use of experimental and quasi-experimental methods in empirical work in economics in the last decade. They specifically question the increased use of instrumental variables and natural experiments in labor economics and of randomized experiments in development economics. In these comments, I will make the case that this move toward shoring up the internal validity of estimates, and toward clarifying the description of the population these estimates are relevant for, has been important and beneficial in increasing the credibility of empirical work in economics. I also address some other concerns raised by the Deaton and Heckman–Urzua papers. (JEL C21, C31)

Estimating the Return to Endogenous Schooling Decisions via Conditional Second Moments

Abstract This paper employs conditional second moments to identify the impact of education in wage regressions where education is treated as endogenous. This approach avoids the use of instrumental variables in a setting where instruments are frequently not available. We employ this methodology to estimate the returns to schooling for a sample of Australian workers. We find that accounting for the endogeneity of education in this manner increases the estimated return to education from 6 percent to 10 percent.

Estimating the Return to Endogenous Schooling Decisions via Conditional Second Moments

This paper employs conditional second moments to identify the impact of education in wage regressions where education is treated as endogenous. This approach avoids the use of instrumental variables in a setting where instruments are frequently not available. We employ this methodology to estimate the returns to schooling for a sample of Australian workers. We find that accounting for the endogeneity of education in this manner increases the estimated return to education from 6 percent to 10 percent.

Some Old and Some New Statistical Tools for Outcomes Research

Outcomes research “seeks to understand the end results of particular health care practices and interventions”1 to inform the development of clinical practice guidelines, evaluate the quality of medical care, and foster effective interventions to improve the quality of care.2 Although randomized trial designs have been used to assess quality of care and to identify effective interventions in the real world,3,4 the empirical basis of outcomes research largely rests on data collected in the observational setting (eg, in the routine setting of everyday practice). With more emphasis placed on increasing the value of health care in terms of lives saved and morbidity avoided, outcomes researchers are making unprecedented demands of observational databases. This is evidenced by the increasing number of and participation in national registries. These include the National Cardiac Data Registry of the American College of Cardiology; the Implantable Cardioverter Defibrillator Registry launched jointly by the American College of Cardiology and the Heart Rhythm Society; the National Cardiac Surgery Database of the Society of Thoracic Surgeons; and the Interagency Registry of Mechanically Assisted Circulatory Support Devices funded by the National Heart, Lung, and Blood Institute, the Centers for Medicare and Medicaid Services, and others. Empirical analyses of these databases require statistical tools that can handle the complexity of the data: observational, sometimes hierarchical, often with multiple outcomes, and always with some missing data. The purpose of the present article is to review key statistical methods important to outcomes research and to introduce newer methodology. The article describes 4 methodological issues commonly present when observational data are analyzed; summarizes the primary assumptions associated with strategies to handle these common problems; demonstrates methods to assess the plausibility of the assumptions associated with each strategy; and illustrates these concepts using examples from cardiovascular outcomes research. Although the intent of the …

New spatial econometric techniques and applications in regional science

The papers appearing in this special issue of Papers in Regional Science, which is devoted to spatial econometrics, come from the First International Conference of the Spatial Econometrics Association held in Cambridge (UK) 12-14 July 2008. This conference was the first official meeting of the new association, which was established in May 2006 in Rome and which has already attracted more than 150 members from around the world. At the Cambridge conference there were close to 120 delegates and more than 100 papers were presented. With regard to the eight papers appearing in this special issue, we would particularly like to thank the authors and the referees for their contribution to what we believe is an interesting and lively selection. Recent years have seen a real explosion in the application of spatial statistical models in all branches of social sciences and in particular in economics. Spatial econometrics models have been used to analyse different topics (see for example Anselin et al. 2004 for a review) and as a matter of fact spatial regression techniques are now becoming an established component in the applied econometrics toolbox, as witnessed by the increasing attention given to this topic in standard econometrics textbooks (Maddala 2001; Woolridge 2002; Gujarati 2003; Kennedy 2003; Baltagi 2008).

Use of instrumental variables in the presence of heterogeneity and self‐selection: an application to treatments of breast cancer patients

Instrumental variable (IV) methods are widely used in the health economics literature to adjust for hidden selection biases in observational studies when estimating treatment effects. Less attention has been paid in the applied literature to the proper use of IVs if treatment effects are heterogeneous across subjects and individuals select treatments based on expected idiosyncratic gains or losses from treatments. In this paper we compare conventional IV analysis with alternative approaches that use IVs to estimate treatment effects in models with response heterogeneity and self-selection. Instead of interpreting IV estimates as the effect of treatment at an unknown margin of patients, we identify the marginal patients and we apply the method of local IVs to estimate the average treatment effect and the effect on the treated on 5-year direct costs of breast-conserving surgery and radiation therapy compared with mastectomy in breast cancer patients. We use a sample from the Outcomes and Preferences in Older Women, Nationwide Survey which is designed to be representative of all female Medicare beneficiaries (aged 67 or older) with newly diagnosed breast cancer between 1992 and 1994. Our results reveal some of the advantages and limitations of conventional and alternative IV methods in estimating mean treatment effect parameters.

Use of instrumental variables in the analysis of generalized linear models in the presence of unmeasured confounding with applications to epidemiological research

A major, often unstated, concern of researchers carrying out epidemiological studies of medical therapy is the potential impact on validity if estimates of treatment are biased due to unmeasured confounders. One technique for obtaining consistent estimates of treatment effects in the presence of unmeasured confounders is instrumental variables analysis (IVA). This technique has been well developed in the econometrics literature and is being increasingly used in epidemiological studies. However, the approach to IVA that is most commonly used in such studies is based on linear models, while many epidemiological applications make use of non-linear models, specifically generalized linear models (GLMs) such as logistic or Poisson regression. Here we present a simple method for applying IVA within the class of GLMs using the generalized method of moments approach. We explore some of the theoretical properties of the method and illustrate its use within both a simulation example and an epidemiological study where unmeasured confounding is suspected to be present. We estimate the effects of beta-blocker therapy on one-year all-cause mortality after an incident hospitalization for heart failure, in the absence of data describing disease severity, which is believed to be a confounder.

Regression Models for Method Comparison Data

Regression methods for the analysis of paired measurements produced by two fallible assay methods are described and their advantages and pitfalls discussed. The difficulties for the analysis, as in any errors-in-variables problem lies in the lack of identifiability of the model and the need to introduce questionable and often naïve assumptions in order to gain identifiability. Although not a panacea, the use of instrumental variables and associated instrumental variable (IV) regression methods in this area of application has great potential to improve the situation. Large samples are frequently needed and two-phase sampling methods are introduced to improve the efficiency of the IV estimators.

Obesity and the Availability of Fast-Food: An Instrumental Variables Approach

This paper addresses the simultaneous determination of fast-food availability on obesity rates through the use of instrumental variables, specifically the number of interstate exits in the county of residence. Using the 2005 Behavioral Risk Factor Surveillance Survey and self-collected data on the density of various fast-food restaurants in US counties, I find that a ten percent increase in the number of restaurants from the mean would increase BMI by .33 points, roughly 1.05kg for a male 1.78 meters tall and .88kg for a female 1.64 meters tall. The results are robust to the selection criteria for counties and the method yields results comparable to previous work looking at rural counties, though the sample employed here is much more generalizable.

Healthcare-Associated Infection and Antimicrobial Resistance: Moving Beyond Description to Prevention

Division of Healthcare Quality Promotion, Centers for Disease Control and Prevention, Atlanta, Georgia, and George W. Merck Fellow, the Institute for Healthcare Improvement, Cambridge, Massachusetts (B.L.Z.); and the Institute for Healthcare Improvement, Cambridge, Massachusetts (D.A.G.) Received February 12, 2007; accepted February 12, 2007; electronically published February 20, 2007. Infect Control Hosp Epidemiol 2007; 28:261-264 2007 by The Society for Healthcare Epidemiology of America. All rights reserved. 0899-823X/2007/2803-0002$15.00. DOI: 10.1086/513722 The burden of healthcare-associated infection (HAI) and infections caused by healthcare-associated multidrug-resistant organisms (HA-MDROs) is substantial. The Centers for Disease Control and Prevention (CDC) estimates the annual number of HAIs in US hospitals at about 1.7 million, resulting in approximately 99,000 deaths, which makes HAI the most common infectious cause of death and one of the 10 leading causes of death overall. Included in these troubling statistics are at least 350,000 infections and 12,000 deaths caused by methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant enterococci, and Clostridium difficile. It is estimated that infections caused by these multidrug-resistant pathogens result in at least $3.5 billion in excess healthcare costs annually (CDC, unpublished data). Three articles in this issue of the journal seek to shed additional light on estimates of the mortality, excess length of stay, and costs attributable to HAI. The first of these, a study from France by Fabbro-Perray et al. measured the attributable mortality of HAI using a large, prospective, matched-cohort study design in a single hospital. The researchers determined that the population-attributable risk of HAI-attributable mortality throughout France lies between 2.1% and 4.0%, on the basis of a 3%-6% range in the incidence of HAI in French hospitals. The second study, by Shurland et al., reports on a 9-year retrospective cohort study of S. aureus infection complicated by bacteremia in a single Veterans Health Administration system, found that patients with MRSA infection, excepting those with pneumonia, had higher mortality than did patients infected with methicillinsusceptible S. aureus. Together these 2 reports emphasize that HAI and HA-MDROs are associated with increased mortality independent of the patient’s underlying illness. The third article, by Graves et al., is a prospective cohort study performed in 2 Australian hospitals that examined excess costs and length of stay attributable to HAI. The authors conclude that HAIs (urinary tract infections, lower respiratory-tract infections, and “other” infections) are associated with little excess cost or length of stay. Specifically, they found only US$17 of extra variable costs and 2.58 extra hospital days per case of lower respiratory-tract infection, whereas urinary tract infection resulted in no excess costs or length of stay, and “other” infection resulted in no excess costs and only 2.61 extra hospital days. The findings of Graves et al. stand in marked contrast to previous higher estimates based on either a matched-cohort approach or the regression analysis methods favored by Fabbro-Perray et al. and Shurland et al. These 3 studies complement a vast and growing literature on the clinical and financial impact of HAI and HA-MDROs. Investigators have studied diverse populations and used a variety of epidemiological and statistical approaches to control for bias and confounding. Not surprisingly, the range of attributable harm and cost varies dramatically. For example, the cost per case of MRSA infection ranges from $7,000 to $32,000 in published studies. In general, but not always, the more aggressive the statistical approach to confounding, the lower the attributable risk. Of course, no amount of statistical manipulation can compensate fully for unmeasured factors, although the use of instrumental variables is a promising partial solution to this problem. At this point in the long history of infection control, it is reasonable to ask why we need more descriptive epidemiological studies of the attributable risk of HAI. Estimating the mortality, excess length of stay, and costs attributable to HAIs would be an interesting academic exercise were there not increasing evidence that most, if not all, of these infections are preventable. Although there are numerous examples of successful smallscale interventions to eliminate adverse events, including HAIs, in hospitals, the rate of overall improvement has been slow. The public is frustrated, and consumer groups are promoting legislation mandating public reporting of rates of infection and other healthcare-associated adverse events. Impatient with the pace of quality improvement nationally, in December 2004 the Institute for Healthcare Improvement (IHI) launched the 100,000 Lives Campaign, a national initiative to prevent 100,000 unnecessary inpatient deaths within 18 months. To help reach this goal, the Campaign proposed

Regression and Weighting Methods for Causal Inference Using Instrumental Variables

Please see the supplementary material for a correction and a new result.Recent researches in econometrics and statistics have gained considerable insights into the use of instrumental variables (IVs) for causal inference. A basic idea is that IVs serve as an experimental handle, the turning of which may change each individual's treatment status and, through and only through this effect, also change observed outcome. The average difference in observed outcome relative to that in treatment status gives the average treatment effect for those whose treatment status is changed in this hypothetical experiment. We build on the modern IV framework and develop two estimation methods in parallel to regression adjustment and propensity score weighting in the case of treatment selection based on covariates. The IV assumptions are made explicitly conditional on covariates to allow for the fact that instruments can be related to these background variables. The regression method focuses on the relationship between responses (observed outcome and treatment status jointly) and instruments adjusted for covariates. The weighting method focuses on the relationship between instruments and covariates to balance different instrument groups with respect to covariates. For both methods, modeling assumptions are made directly on observed data and separated from the IV assumptions, whereas causal effects are inferred by combining observed data models with the IV assumptions through identification results. This approach is straightforward and flexible enough to host various parametric and semiparametric techniques that attempt to learn associational relationships from observed data. We illustrate the methods by an application to estimating returns to education.

Overcoming Apartheid: Can Truth Reconcile a Divided Nation? By James L. Gibson New York: Russell Sage Foundation, 2004. Pp. 488. $47.50 cloth.

Overcoming Apartheid: Can Truth Reconcile a Divided Nation? By James L. Gibson. New York: Russell Sage Foundation, 2004. Pp. 488. $47.50 cloth. Reviewed by John Hagan, Northwestern University and American Bar Foundation, and Sanja Kutnjak Ivkovic, Florida State University Gibson's ambitious goal is to assess the success of the South African Truth and Reconciliation Commission (TRC) not only in reducing this nation's racial divide, but also in increasing South Africa's political tolerance, its support for human rights, and the legitimacy of its governing political institutions. He rises to this challenge by analyzing the results of a landmark 2000-2001 survey of nearly 4,000 South Africans. Gibson is cautious but not shy in delivering a bottom-line assessment. He finds that nearly one-half of the South African population entered the new millennium expressing some degree of Given the challenges of overcoming apartheid, the TRC, according to Gibson, has likely delivered as much as could be expected. The reader learns much across the chapters of this book: that reconciliation is a measurable construct, that interpretive is critical in creating a collective memory, that intense contact among members of conflicting subgroups can help achieve reconciliation, that the creation of a human rights culture-among the populace and the government-is both a backward- and forward-looking process, that intolerance is a social and not an individual-level characteristic, that amnesty can be a powerful source of perceived injustice, and that public institutions that serve as the backbone of a democracy must develop a substantial degree of legitimacy. Among the strengths of Gibson's accomplishment are the precision of his measurement approach and the cautiousness of his analytic judgments. For example, subdimensions of reconciliation are scaled in a manner that allows Gibson to estimate levels of reconciliation among national subgroups. When the analysis is taken to this level, he finds that reconciliation characterizes the attitudes of more than one-half of Whites (56%) and Colored people (59%) but only about one-third of Africans. When the variable of language is further introduced, over two-thirds of English-speaking Colored people (75%) and English-speaking Whites (64%) are found somewhat reconciled, compared to less than one-fifth (17%) of Sotho-speaking blacks. The latter nearly four-toone difference in reconciliation provides good reason for the cautious framing of Gibson's conclusions (p. 332). Gibson also develops measures of the he attributes to the TRC. Here he finds a surprising similarity in the acceptance of the TRC's depiction of apartheid across racial groups, and he attributes this to a complementary acknowledgment and moderation of views across these groups. He can further show that those who are more accepting of the TRC's version of the are more likely to be reconciled. He acknowledges that a more convincing demonstration of the causality involved in this relationship would require longitudinal data, which his survey does not provide. Nonetheless, Gibson's use of instrumental variables with his cross-sectional data supports that assumption that the causal flow is indeed from truth to reconciliation. He can then again break this analysis down by subgroups, leading to the finding that the effect is 0.23 among Africans and 0.53 among Whites. This more than two-to-one difference is added reason for guardedness in Gibson's optimism (p. 334). A particularly intriguing part of Gibson's contribution involves his assessment of the contact hypothesis as an explanatory mechanism for the level of reconciliation that is evident in postapartheid South Africa. Gibson provides important evidence that the segregation, indeed isolation, especially social-psychological isolation, of Africans from Whites is pervasive in South Africa. …

Does Ambulatory Process of Care Predict Health‐Related Quality of Life Outcomes for Patients with Chronic Disease?

The validity of quality of care measurement has important implications for practicing clinicians, their patients, and all involved with health care delivery. We used empirical data from managed care patients enrolled in west coast physician organizations to test the hypothesis that observed changes in health-related quality of life across a 2.5-year window reflecting process of care. Patient self-report data as well as clinically detailed medical record review regarding 963 patients with chronic disease associated with managed care from three west coast states. Prospective cohort study of change in health-related quality of life scores across 30 months as measured by change in SF-12 physical component scores. Patient self-report and medical record abstraction. We found a positive relationship between better process scores and higher burden of illness (p<.05). After adjustment for burden of illness, using an instrumental variables approach revealed better process is associated with smaller declines in SF-12 scores across a 30-month observation window (p=.014). The application of the best quartile of process of care to patients currently receiving poor process is associated with a 4.24 increment in delta SF-12-physical component summary scores. The use of instrumental variables allowed us to demonstrate a significant relationship between better ambulatory process of care and better health-related quality of life. This finding underscores the importance of efforts to improve the process of care.