Two-sample Instrumental Variables Research Articles

Effects of e-cigarette use on mental health among youths: quasi-experimental evidence from Canada.

Existing research on mental health comorbidities of youth e-cigarette use is subject to confounding bias and reverse causality. This study aimed to measure the effects of e-cigarette use on youth mental health, using e-cigarette minimum legal age (MLA) law in Canada as a natural experiment. We used difference-in-differences (DD), difference-in-differences-in-differences (DDD) and two-sample instrumental variables (TSIV) methods. Data were from nationally representative Canadian Community Health Surveys 2008-2019 and Canadian Student Tobacco Alcohol and Drugs Surveys 2008-2019. The study sample comprised of respondents aged 15 to 18 (in DD analysis; n = 33 858) and aged 15 to 24 (in DDD analysis; n = 78 689). Primary outcomes were self-reported mood disorders and anxiety disorders. Secondary outcomes were cannabis use, illicit drug use, cigarette use and strength of peer relationships at schools. After the e-cigarette MLA laws, risks of mood disorders declined by 1.9 percentage points (95% CI, 0.0-3.8; P =0.05) in the DD analysis and by 2.6 percentage points (95% CI, 0.2-5.0; P =0.03) in the DDD analysis. For anxiety disorders, while the DD estimate was negative but imprecisely estimated, the MLA law reduced risks of anxiety disorder by 3.6 percentage points (95% CI, 0.9-6.2; P = 0.01) in the DDD analysis. Youths in provinces with MLA laws were also less likely to report cannabis use and illicit drug use and more likely to feel being part of schools. TSIV analysis indicates that youth e-cigarette use increased the likelihood of mood and anxiety disorders by 44% and 37%, respectively. In Canada, the e-cigarette minimum legal age law appears to have reduced risks of mood and anxiety disorders, lowered substance use and improved peer relationships at schools. Combined with previous evidence of lower e-cigarette use following the minimum legal age law, our findings indicate that youth e-cigarette use increases risks of mood and anxiety disorders.

Does a good father now have to be rich? Intergenerational income mobility in rural India

For developing countries, it is difficult to find income mobility studies that rely on datasets linking parents with their children. Most of the available evidence for developing countries, instead, relies on co-resident households and two-sample instrumental variables approach, and findings in the mobility literature suggest that these approaches could introduce bias in the IGE estimates. The current study rectifies this important omission for rural India, and develops strategies that will be useful for similar research in other developing countries. Using a panel dataset spanning 1994–2012, I find that the intergenerational income persistence for rural India is lower than analogous evidence from other developing countries (IGE: 0.30 [se:0.03]). However, the between-groups coefficient of 0.74, which is corroborated by repeated cross-sections from the National Sample Survey, indicates that India is progressing towards cross-caste equality, but at a disappointingly slow rate.

Liquid Assets? the Short-Run Liabilities of Binge Drinking

AbstractWe estimate the effect of binge drinking on road accidents, accident and emergency (A&E) attendances and arrests using a variety of unique English data and a two-sample instrumental variables estimation procedure. Drinking $10+$ units of alcohol in a single session increases road accidents by 18.6%, injury-related A&E attendances by 6.6% and arrests by 71%. The marginal increase from eight to $10+$ alcoholic units implies nearly 6,100 extra road accidents every year, 63,000 additional A&E attendances and 100,000 additional arrests. The externality per mile driven by a binge drinker is about five pence and the punishment that internalises this externality is equivalent to a fine of £22,800 per drunk driving arrest.

Intergenerational Earnings Persistence and Economic Inequality in the Long-Run: Evidence from French Cohorts, 1931-1975

This paper analyzes long-term trends in intergenerational earnings persistence in France for male cohorts born between 1931 and 1975. This time period has witnessed important changes in the French labor market and educational system, in particular an important compression of earnings differentials as well as a large expansion in access to secondary and higher education. Using a two-sample instrumental variables approach, I estimate two measures of intergenerational economic persistence: the intergenerational earnings elasticity (IGE) and the intergenerational correlation (IGC). Over the period, the IGE exhibits a V-shaped pattern. It falls from a high of value of .6 for cohorts born in the 1930s to around .4 for those born in the 1950s, but subsequently rises to a level close to the beginning of the period. In contrast, the IGC remains relatively stable over the period. This suggests that changes in the IGE are partly driven by transitory responses to changes in cross-sectional inequality rather than long-term changes in the degree of intergenerational persistence.

Efficient Estimation of Data Combination Models by the Method of Auxiliary-to-Study Tilting (AST)

We propose a locally efficient estimator for a class of semiparametric data combination problems. A leading estimand in this class is the average treatment effect on the treated (ATT). Data combination problems are related to, but distinct from, the class of missing data problems with data missing at random (of which the average treatment effect (ATE) estimand is a special case). Our estimator also possesses a double robustness property. Our procedure may be used to efficiently estimate, among other objects, the ATT, the two-sample instrumental variables model (TSIV), counterfactual distributions, poverty maps, and semiparametric difference-in-differences. In an empirical application, we use our procedure to characterize residual Black–White wage inequality after flexibly controlling for “premarket” differences in measured cognitive achievement. Supplementary materials for this article are available online.

The impact of mothers’ earnings on health inputs and infant health

This paper investigates the impact of mothers’ earnings on birth weight and gestational age of infants in the U.S. It also analyzes the impact of earnings on mothers’ consumption of prenatal medical care, and their propensity to smoke and drink during pregnancy. The paper uses census division-year-specific skill-biased technology shocks as an instrument for mothers’ earnings and employs a two-sample instrumental variables strategy. About 14 million records of births between 1989 and 2004 are used from the Natality Detail files along with the CPS Annual Demographic Files from the same period. The results reveal that an increase in weekly earnings prompts an increase in prenatal care of low-skill mothers (those who have at most a high school degree) who are not likely to be on Medicaid, and that earnings have a small positive impact on birth weight and gestational age of the newborns of these mothers. Specifically, if a mother's earnings double, this produces a weight gain of the newborn by about 100g and an increase in gestational age by 0.7 weeks. An increase in earnings does not influence the health of newborns of high-skill mothers (those with at least some college education). Variations in earnings have no impact on birth weight for mothers who are likely to be on Medicaid.

Intergenerational earnings mobility in Japan among sons and daughters: levels and trends

This paper estimates the extent of intergenerational income mobility in Japan among sons and daughters born between 1935 and 1975. Our estimates rely on a two-sample instrumental variables approach using representative data from the Japanese Social Stratification and Mobility surveys, collected between 1965 and 2005. Father’s income is predicted on the basis of a rich set of variables, and we discuss changes in the Japanese earnings structure for cohorts born between the early 1900s and the 1960s. Our main results indicate that the intergenerational income elasticity (IGE) for both sons and daughters in Japan lies around 0.35, which is an intermediate value, by international standards. We discuss the sensitivity of the IGE to using either personal or family income as the income variable for both fathers and children. We also examine changes across cohorts in the IGE. Results indicate that intergenerational mobility has been roughly stable over the last decades.

The business cycle and health behaviors

While it is well documented that economic expansions provide widespread and immediate financial benefits, the evidence on how an economic downturn affects individual's health behaviors is surprisingly mixed. In this paper, we take a structural approach to investigate the effects of wages and working hours on health behaviors of low-educated persons using variations in wages and hours caused by changes in local economic activity. In the empirical analysis, we adopt a two-sample instrumental variables approach to combine the data on individual health behaviors from the Behavioral Risk Factor Surveillance System (BRFSS) and the National Health Interview Survey (NHIS) with the data on individual employment from the Current Population Survey (CPS). The total sample size of the combined CPS-BRFSS data for the time period of 1984–2005 is 967,594, while that of the combined CPS-NHIS data for the time period of 1976–2001 is 364,078. We find that increases in wages caused by economic expansions are associated with greater consumption of cigarettes in the United States. We also find that increases in hours of work caused by economic expansions are associated with more cigarette consumption, but less physical activity and physician visits. More importantly, the evidence suggests that most of such effects associated with changes in hours of work can be attributed to the changes at the extensive margin of employment, meaning the changes in employment status, rather than the changes at the intensive margin, meaning changes in hours of work conditional on being employed. These findings imply that changes in employment may have heterogeneous impacts on time-intensive and less time-intensive health behaviors and also provide additional evidence on the importance of time in health production, particularly for time-intensive activities.

Intergenerational Earnings Mobility in Japan Among Sons and Daughters: Levels and Trends

This paper estimates the extent of intergenerational income mobility in Japan among sons and daughters born between 1935 and 1975. Our estimates rely on a two-sample instrumental variables approach using representative data from the Japanese Social Stratification and Mobility (SSM) surveys, collected between 1965 and 2005. Father's income is predicted on the basis of a rich set of variables including education, occupation and job characteristics and we discuss changes in the Japanese earnings structure for cohorts born between the early 1900s and the 1960s. Our main results indicate that the intergenerational income elasticity (IGE) for both sons and daughters, in Japan lies around .35, which is an intermediate value, by international standards. We discuss the sensitivity of the IGE to using either personal or family income as the income variable for both fathers and children and perform some robustness analysis with respect to the first-step specification and to the age selection rule for children. We also examine changes across cohorts in the IGE. Results indicate that intergenerational mobility has been roughly stable over the last decades. Lastly, we estimate the intergenerational correlation of earnings.

Smart Money: The Effect of Education on Financial Behavior

Household financial decisions are important for both households and the greater economy. Yet, our understanding of the process of financial decision-making is limited. Applying standard and two-sample instrumental variables strategies to census and credit bureau data, we provide the first precise, causal estimates of the effects of education on financial behavior. Education has large effects on financial market participation and smaller, but statistically and economically significant effects on financial management. We find that education improves credit scores, and dramatically reduces the probability of declaring bankruptcy or suffering foreclosure during the financial crisis. Examining mechanisms, we show that cognitive ability increases financial participation, and discuss how education may affect decision-making through: attitudes, borrowing behavior, discount rates, risk-aversion, and the influence of coworkers and neighbors.

Intergenerational Economic Mobility in the U.S., 1940 to 2000

We use a two-sample instrumental variables approach to estimate a time series of intergenerational economic mobility using the decennial U.S. Censuses. We find that the intergenerational income elasticity (IGE) followed a roughly U-shaped pattern from 1940 to 2000 that is similar to well known cross-sectional inequality trends. In particular, we find that intergenerational mobility (measured as 1 minus the IGE) increased from 1940 to 1980 but has declined sharply since 1980. The decline after 1980 is robust to alternative estimation approaches. This suggests that the rate of regression to the mean has been notably lower during the 1980s and 1990s compared to the three decades after WWII. Historical trends in the returns to education can only account for some of these changes in the IGE. The time pattern and the changes across birth cohorts may also help to reconcile previous findings in the literature that have used different surveys. Our estimates imply a somewhat different pattern for the intergenerational income correlation, a measure which is insensitive to changes in cross-sectional inequality and has implications for rank or positional mobility. We find that the post-1980 decline in intergenerational rank mobility marks a return to typical historical levels. At the end of the 20th century, the rate of movement of families across the income distribution across generations appears historically normal, but, as cross-sectional inequality has increased, earnings are regressing to the mean at a much slower rate, causing economic differences between families to persist longer than they had earlier in the century.

Two-Sample Instrumental Variables Estimators

Abstract Following an influential article by Angrist and Krueger (1992) on two-sample instrumental variables (TSIV) estimation, numerous empirical researchers have applied a computationally convenient two-sample two-stage least squares (TS2SLS) variant of Angrist and Krueger's estimator. In the two-sample context, unlike the single-sample situation, the IV and 2SLS estimators are numerically distinct. We derive and compare the asymptotic distributions of the two estimators and find that the commonly used TS2SLS estimator is more asymptotically efficient than the TSIV estimator. We also resolve some confusion in the literature about how to estimate standard errors for the TS2SLS estimator.

Are Public Housing Projects Good For Kids?

One of the goals of federal housing policy is to improve the prospects of children in poor families. But little research has been conducted into the effects of participation in housing programs on children, perhaps because it is difficult to find data sets with information about both participation and interesting outcome measures. This paper combines data from several sources in order to provide a first look at the effect of public housing projects on housing quality and the educational attainment of children. We first use administrative data from the Department of Housing and Urban Development to impute the probability that a Census household lives in a public housing project. We find that a higher probability of living in a project is associated with poorer outcomes. We then use two-sample instrumental variables (TSIV) techniques to combine information on the probability of living in a project obtained from the 1990 to 1995 Current Population Surveys, with information on outcomes obtained from 1990 Census. The instrument common to both samples is an indicator equal to one if the household is entitled to a larger housing project unit because of the sex composition of the children in the household. Families entitled to a larger unit because of sex composition are 24% more likely to live in projects. When we control for omitted variables bias using TSIV, we find that project households are less likely to suffer from overcrowding and less likely to live in high-density complexes. Project children are also 12 to 17 percentage points less likely to have been held back in school one or more grades, although this effect is confined to boys. Thus, most families do not face a tradeoff between housing quality and child outcomes -- the average project improves both.

Permanent Income, Current Income, and Consumption: Evidence From Two Panel Data Sets

In this article, I estimate Euler equations—that is, the first-order conditions of the consumers' maximization problem—using data from two data sets. Consumption data are taken from the Consumer Expenditure Survey. Income data are taken from the Panel Study of Income Dynamics. Because the data for the estimation come from two samples, I use a generalization of the instrumental variables estimator—the two-sample instrumental variables estimator. I find evidence that consumption is excessively sensitive to predictable income growth. The estimates of the coefficient of excess sensitivity for three consumption measures range from .2 to .5.