Home | Business News | Browse by Publication | E | Economic Inquiry

The health effects of military service: evidence from the Vietnam draft.

Article Excerpt
I. INTRODUCTION

The impact of military service on long-term health is reemerging as a vital policy question as the United States conducts major military operations in Iraq and Afghanistan. Health professionals have raised concerns that significant numbers of the veterans of these conflicts are experiencing psychological and physical health problems. The Department of Veterans Affairs estimates that anywhere from 12 to 20% of servicemen returning from Iraq suffer from posttraumatic stress disorder, which the medical literature suggests can have a long-term adverse effect on an individual's health. In the 2005 fiscal year, the Department of Veterans Affairs spent about 32 billion dollars on medical care for veterans: a substantial liability for the American taxpayer. Thus, understanding the long-term health impacts of military service is important both for allocating treatment resources and for correctly estimating the cost of waging these two wars.

Comparing the health of veterans with that of nonveterans may not give us credible estimates of the effect of military service. People voluntarily enlisting in the military differ in their observable characteristics from those who do not. These observable differences between veterans and nonveterans leave us concerned that cross-sectional estimates of the long-term health consequences of military service will be biased by systematic differences in the unobservable characteristics of the two groups. For this reason, we use the draft lottery implemented during the Vietnam War as an instrument for military service. The draft lottery required randomly selected men born between 1944 and 1952 to report for possible induction into the military.

We use confidential versions of the 19742004 National Health Interview Surveys (NHIS) (1) to examine the impact of military service on veterans' health. We focus on the 67,608 men who were born between 1950 and 1952 because, as we document, this is the cohort most affected by the Vietnam draft lottery. We begin with a cross-sectional comparison of veterans and nonveterans, which reveals that veterans are in systematically worse health than nonveterans and that these health disparities increase with age. These estimates probably overstate the causal effect of military service because veterans differ systematically from nonveterans on dimensions such as education that are strongly associated with adverse health. When we correct for this problem using draft status to instrument for military service, we find that despite a very strong first-stage relationship between draft eligibility and military service, the 2-stage least squares (2SLS) estimates of the impact of military service on health are fairly imprecise. For most of the outcomes, we examine the confidence intervals from the 2SLS regressions containing both zero and the point estimate from the cross-sectional regression.

This paper makes two contributions to the literature. First, unlike the cross-sectional literature on the impact of military service on health, the instrumental variables approach used in this paper provides unbiased estimates. Second, this paper documents that though the Vietnam lottery very significantly increased military service, it led to only a modest increase in combat exposure. (2) In the following section, we discuss the existing literature on veterans' health and on the impact of the Vietnam Era Draft. In Section III, we describe our primary data source, the NHIS. In Section IV, we discuss the research methodology and some of the issues that complicate the use of draft eligibility as an instrument. Section V discusses our results, and Section VI concludes.

II. PREVIOUS LITERATURE

There is a significant literature documenting the disproportionate prevalence of health problems among military veterans. Hoge et al. (2004) find that mental health problems are common in veterans who served in Iraq and Afghanistan. Kang and Bullman (2001) document an increase in mortality caused by motor vehicle accidents in a 7-yr follow-up study of Gulf War veterans. McKinney et al. (1997) show that veterans are more likely to smoke than nonveterans. In one of the early papers in this literature, Card (1987) finds that Vietnam veterans are much more likely to report problems associated with posttraumatic stress disorder including "nightmares, loss of control of behavior, emotional numbing, withdrawal from the external environment, hyperalertness, anxiety, and depression." Jordan et al. (1991) document that exposure to combat in Vietnam is associated with higher prevalence of specific psychiatric disorders, including posttraumatic stress disorder. McFall, Mackay, and Donovan. (1992) and Price et al. (2004) present evidence of a strong link between posttraumatic stress disorder and substance abuse.

The papers described above share a common limitation: given the very significant differences in the characteristics of veterans and nonveterans, it is likely that some of the differences they document are due to omitted variables bias rather than the adverse effects of military service. (3) The first paper to deal convincingly with the omitted variables problem was by Hearst, Newman, and Hulley (1986). They used draft eligibility as an exogenous source of variation in the probability of military service. The authors found that men born on days that made them eligible for the draft hada 4% higher mortality rate and were 13% more likely to commit suicide and 8% more likely to the in a motor vehicle accident. A second paper that tackles the omitted variables problem convincingly is Bedard and Deschenes (2006). The authors use an instrumental variables strategy to credibly document that World War II and Korean War veterans suffer from substantially increased rates of premature mortality. One major contribution of their paper is that it documents that military-induced smoking is a major causal pathway through which military service causes premature mortality.

This paper contributes to this literature using the Vietnam draft as an instrument to get unbiased estimates of the long-term impact of military service on health behaviors and morbidity. An additional advantage of this paper is that it examines the broad set of outcomes that the literature suggests are...

View this article FREE - Now for a Limited Time, try Goliath Business News
Free for 3 Days!