Health, human capital, and domestic violence

Nicholas W. Papageorge, Gwyn C. Pauley, Mardge Cohen, Tracey E. Wilson, Barton H. Hamilton, Robert Pollak 01 April 2017



Domestic violence is disturbingly common in the US, with almost one in four women reporting being physically assaulted at some point in their lives (Tjaden and Thoennes 2000). The direct costs of domestic violence (including medical costs and productivity losses) have been estimated to be approximately $5.8 billion per year. Unfortunately, policies intended to reduce the prevalence of violence, such as mandatory arrest laws, have been shown to be of limited effectiveness (Iyengar 2009).

Previous research has shown that violence is associated with poor labour market outcomes (Browne et al. 1999, Swanberg and Macke 2006, Aizer 2010). Yet the relationship between health and domestic violence is not well understood. Though exposure to violence can negatively affect health, it is not known how health affects a woman’s likelihood of suffering violence.

In this column, we are the first to show that health improvements can reduce domestic violence (see also Papageorge et al. 2016). We use the introduction of HAART, a new treatment for HIV, to show that positive health shocks reduced domestic violence among a group of HIV+ women. We also show that this same medical innovation decreased illicit drug use.

Our findings speak to the fact that health improvements can have broad impacts on social problems. The policy relevance of our findings is compounded by the fact that both domestic violence and illicit drug use have been frustratingly impervious to a variety of interventions (on cyclicality of violence, see e.g. Bowlus and Seitz 2006). The introduction of HAART provides a unique opportunity to test this relationship between health and these social outcomes because we can examine what happens when factors underlying violence or drug use shift exogenously.


Although there is no vaccine or cure for HIV, HAART is the current standard treatment. It was introduced in late 1996 after protease inhibitors (made widely available towards the end of 1995) were recognised to be an effective HIV treatment. At the same time, it was discovered that several anti-retroviral drugs taken simultaneously could indefinitely delay the onset of immune system decline among HIV+ patients.

The introduction of HAART is a natural setting to study the effect of health for two main reasons. First, the introduction of HAART was unanticipated, allowing us to causally estimate the effect of health on social outcomes. Second, HIV infection was, at the time HAART was introduced, an extremely severe disease. HAART essentially turned HIV infection from a death sentence to a manageable, chronic disease. In fact, within two years of its introduction, HAART decreased mortality rates by 80% (Bhaskaran et al. 2008). This allows us to observe the effect sizes large enough to detect the nuanced relationship between medical innovation, health, and other social outcomes.


We use the Women’s Interagency Health Study (WIHS), a unique source of data, to study how health improvements affected women. The WIHS began in the early 1990s to study how women were affected by the HIV crisis. It is a longitudinal study and continues to interview women about their labour force participation, illicit drug use, domestic violence, and health every six months. Women were recruited from a variety of places including health clinics and drug rehabilitation programs (Barkan et al. 1998). Importantly for this study, data collection began before the introduction of HAART, allowing us to compare women before and after the innovation.


We study a group of women who were just starting to feel ill and were therefore most affected by the introduction of HAART. These women were beginning to experience HIV-induced health deterioration just prior to the introduction of HAART. We compare this group of women to a control group consisting of HIV+ women who were in relatively good health and had not started to experience immune system deterioration. We base our choice of control group on the idea that it consists of women affected by the introduction of HAART less than women who were sicker and thus would expect to rely on it sooner. Although the control and treatment groups both benefit from HAART in expectation, our treatment group immediately experienced improvements in health because of HAART.

Health, domestic violence, and risky behaviours

We find that the group of women most affected by HAART experienced substantial decreases in domestic violence as compared to HIV+ women who were in better health. The decrease in violence of about 5% is even larger, over 10%, when we focus on a subsample of black women. Turning to illicit drug use, we find that this same group of women reduced their use of heroin by over 10%. Again, this change was larger for black women.

Investments in human capital

To help explain our findings, we turn to economic underpinnings of human capital. Namely, we think of health as a form of human capital and argue that the introduction of HAART caused women whose life expectancy was dramatically increased to have a stronger incentive to invest in themselves (Grossman 1972, Becker 2007). One way that they do so is by avoiding risky behaviours, such as illicit drug use.  

Additionally, we also show evidence that the women who were most affected by the introduction of HAART were more likely to be employed after the medical innovation. Although difficult to pin down the mechanism driving these findings, we argue that these findings strengthen our claim that health is human capital and that HAART, by improving health trajectories, boosted women’s human capital.


These findings speak to the interaction between health and social issues.  We show that two persistent social ills, violence and drug use, can be reduced by improving health. By ignoring the link between medical innovation, health, and these outcomes, under investment in research is likely to occur. As we debate the future of health care in the US, this important link should not be forgotten.

Although our study is focused on low income, HIV+ women, this does not mean that our findings are not generalisable. Violence and drug use are not limited to women with HIV. Further, HIV is a chronic disease that when treated properly is manageable, similar to diabetes. Our focus on low-income, HIV+ women helps us to understand how chronically ill women with limited resources respond to and benefit from a medical breakthrough.


Aizer, A (2010) “The gender wage gap and domestic violence”, American Economic Review, 100(4): 1847-1859.

Barkan, S E, S L Melnick, S Preston-Martin, K Weber, L A Kalish, P Miotti, M Young, R Greenblatt, H Sacks, and J Feldman (1998) “The women’s interagency HIV study”, Epidemiology, 9(2): 117–125.

Becker, G S (2007) “Health as human capital: Synthesis and extensions”, Oxford Economic Papers, 59(3): 379–410.

Bhaskaran, K, O Hamouda, M Sannes et al (2008) “Changes in the risk of death after HIV seroconversion compared with mortality in the general population”, Journal of the American Medical Association, 300(1): 51–59.

Bowlus, A J and S Seitz (2006) “Domestic violence, employment, and divorce”, International Economic Review, 47(4): 1113–1149.

Browne, A, A Salomon and S S Bassuk (1999) “The impact of recent partner violence on poor women’s capacity to maintain work”, Violence Against Women, 5(4): 393–426.

Iyengar, R (2008) “Does the certainty of arrest reduce domestic violence? Evidence from mandatory and recommended arrest laws”, Journal of Public Economics, 93(1): 85-98.

Grossman, M (1972) “On the concept of health capital and the demand for health”, Journal of Political Economy, 80(2): 223–255.

Papageorge N W, G C Pauley, M Cohen, T E Wilson, B H Hamilton and R A Pollak (2016), “Health, Human Capital and Domestic Violence”, NBER Working Paper No. 22887.

Swanberg, J E and C Macke (2006) “Intimate partner violence and the work-place consequences and disclosure”, Journal of Women and Social Work, 21(4): 391–406.

Tjaden, P and N Thoennes (2000) “Full report of the prevalence, incidence, and consequences of violence against women”, Technical Report, US Department of Justice. 



Topics:  Health economics Labour markets

Tags:  health, HIV/AIDS, domestic violence, human capital, drug use

Assistant Professor of Economics, Johns Hopkins University

Postdoctoral Fellow, Schaeffer Center for Health Policy & Economics, University of Southern California

Professor of Medicine, Department of Medicine, Rush University in Chicago

Professor, Department of Community Health Services, SUNY Downstate Medical Center

Robert Brookings Smith Distinguished Professor of Entrepreneurship, Olin Business School, Washington University in St. Louis

Research Fellow, Institute for the Study of Labor (IZA) and CESifo; Research Associate, NBER