April 17, 2015
Harry Krashinsky, University of Toronto, and Philip DeCicca, McMaster University and NBER
While much descriptive work implies less-educated women are more likely to give birth as teenagers, there is much less evidence the relationship is causal. DeCicca and Krashinsky investigate this possibility using variation in compulsory schooling laws (CSLs) in an instrumental variables framework to identify the impact of formal education on teen fertility for a large sample of women drawn from multiple waves of the Canadian Census. The researchers find that greater CSL-induced schooling reduces the probability of giving birth as a teenager by roughly four to eight percentage points. They also explore possible mechanisms underlying this relationship by examining the timing of their estimates. The authors find evidence that education affects the timing of births in a way that strongly implies an "incarceration" effect of education. In particular, they find large negative impacts of education on births to young women aged seventeen and eighteen, but little evidence of an effect after these ages, consistent with the idea that being enrolled in school deters fertility in a contemporaneous manner, but not in the longer-run. The researchers' findings are robust to the inclusion of several province-level characteristics including multiple dimensions of school quality as well as expenditures on public programs.
Thomas Goldring, Carnegie Mellon University; Fabian Lange, McGill University; and Seth Richards-Shubik, Carnegie Mellon University and NBER
Goldring, Lange, and Richards-Shubik develop a flexible test for changes in the SES-mortality gradient over time that directly accounts for changes in the distribution of education, the most commonly used marker of SES. The researchers implement the test for the period between 1984 and 2006 using microdata from the Census, CPS, and NHIS linked to death records. Using their flexible test, the authors find that the evidence for a change in the education-mortality gradient is not as strong and universal as previous research has suggested. Their results indicate that the gradient increased for females during this time period, but they cannot rule out that the gradient among males has not changed. Informally, the results suggest that the changes for females are mainly driven by the bottom of the education distribution.
Elaine M. Liu, University of Houston and NBER; Jin-Tan Liu, National Taiwan University and NBER; and Hazel Tseng, University of Houston
Liu, Liu, and Tseng examine the impact of a magnitude-7.3 earthquake on fetal losses and birth outcomes in Taiwan. The researchers' main identification strategy is a difference-in-differences method. They compare the pregnancy outcomes and birth cohort size of those who resided in areas with high earthquake intensity to those who resided in areas with low earthquake intensity, before and after the earthquake. The authors' analysis suggests that the incidence of fetal losses increased by 4.4 percentage points in the most affected regions relative to the least affected regions for those who had exposure during the first trimester. Almost all of the fetal losses are due to the loss of male fetuses. Exposure during the second or third trimester results in lower birth weights, but no significant impact on fetal losses. The researchers do not find women from lower socioeconomic status to be more affected by earthquakes. Lastly, the authors suggest that the results are mainly driven by maternal stress.
Mark L. Egan, University of Chicago, and Tomas Philipson, University of Chicago and NBER
Non-adherence in health care results when a patient does not initiate or continue care that a provider has recommended. Previous researchers have identified non-adherence as a major source of waste in U.S. health care, totaling approximately 2.3% of GDP, and have proposed a plethora of interventions to raise adherence. However, health economics has provided little explicit analyses of the important and dynamic demand behavior that drives non-adherence, and it is often attributed to uninformed patients. Egan and Philipson argue that whereas providers may be more informed about the population-wide effects of treatments, patients are more informed about the individual value of a treatment. The researchers interpret a patient's decision to adhere to a treatment regime as an optimal stopping problem in which patients learn the value of a treatment through experiencing it. The authors derive strong positive and normative implications resulting from interpreting non-adherence as an optimal stopping problem. Their positive analysis derives an "adherence survival function," depicting the share of patients still on treatment as a function of time, and predicts how various observable factors alter it. The researchers' normative analysis derives the efficiency effects of non-adherence and the conditions under which adherence is too high or low. They consider the efficiency implications of this analysis for common adherence interventions. The authors' argue that personalized medicine, by replacing learning through experience with a companion diagnostic, speeds up our leaning process and raises efficiency through cutting over-adherence. The researchers assess the quantitative importance of these implications by calibrating the degree of over- and under-adherence for one of the largest U.S. drug categories, cholesterol-reducing drugs. Contrary to frequent normative claims of under-adherence, their estimates suggest the ex-post efficiency loss from over-adherence is over 80% larger than from under-adherence, even though only 43% of patients fully adhere.
D. Mark Anderson, Montana State University; Benjamin Crost, University of Illinois, Urbana-Champaign; and Daniel I. Rees, University of Colorado, Denver
Drawing on county-level data from Kansas for the period 1977-2011, Anderson, Crost, and Rees examine whether plausibly exogenous increases in the number of establishments licensed to sell alcohol by the drink are related to crime. During this period, 86 out of 105 counties in Kansas voted to legalize the sale of alcohol to the general public for on-premises consumption. These counties experienced substantial increases in the total number of drinking establishments (e.g., bars and restaurants). Using legalization as an instrument, this study is the first to provide estimates of the elasticity of crime with respect to drinking establishments. Specifically, the researchers show that a 10 percent increase in drinking establishments is associated with a 3 to 4 percent increase in crime. Reduced-form estimates suggest that legalizing the sale of alcohol to the general public for on-premises consumption leads to a 10 to 12 percent increase in crime.
Chad D. Cotti, University of WisconsinOshkosh; Erik Nesson, Ball State University; and Nathan Tefft, Bates College
Cotti, Nesson, and Tefft use the Nielsen Homescan Consumer Panel (NHCP) to estimate the effects of tobacco control policies on purchases of tobacco-related products using within-household variation. The researchers also match 91% of cigarette product purchases in the NHCP to tar, nicotine, and carbon monoxide contents from the National Health and Nutritional Examination Surveys. Higher cigarette taxes reduce the number of cigarettes households purchase, but households also purchase cigarettes with higher tar, nicotine, and carbon monoxide contents. Contrary to previous findings, this effect is overwhelmed by the reduction in cigarettes purchased. Neither smokefree air laws nor smokeless tobacco taxes affect tobacco product purchases.