The Gender Wage Gap and Domestic Violence

Anna Aizer asks whether the wages women could earn outside a household affect the risk of violence within it. She uses California hospital records for women admitted after assaults from 1990 to 2003. To estimate relative wage opportunities, she combines historical patterns of employment across industries in each county with changes in industry wages. The approach seeks variation in local labour demand and does not rely on a direct comparison of partners’ observed earnings.

Outside options and observed harm

The estimates associate a narrowing wage gap with fewer female hospital admissions for assault. This is consistent with an account in which the ability to support oneself changes bargaining power and makes leaving more feasible. Aizer checks the relationship with different statistical specifications. The inference still relies on assumptions about how the constructed measure of wage opportunities relates to other local changes.

Aizer attributes about nine per cent of the observed decline in her violence measure to the narrowing wage gap. The denominator matters: it is the decline that took place during the study, rather than the number of women experiencing violence or a forecast of how an individual’s risk will change. The estimate leaves room for other mechanisms and for policies that affect access to housing, shelter and support directly.

A hospital record captures some severe physical harm. It leaves out much coercion, threats and violence for which someone does not seek hospital care. The data do not identify every attacker as a partner, and changes in hospital use matter to the count. The estimates have a specific place, period and measure.

A measure of offers available outside the household

In the design section, notice why the study does not take partners’ actual pay as the starting point. Violence can cause someone to miss work or lose a job, so lower earnings may also be a consequence of violence. The county measure combines earlier patterns of employment across industries with later shifts in industry wages. Imagine two counties with different industrial histories. A rise in pay in one industry can change their residents’ opportunities by different amounts. Other local changes that affect both jobs and hospital admissions are a challenge for the design.

The bargaining model explains why being able to leave may matter even if a woman does not immediately leave. The wage opportunity is one resource in that decision. Its value can depend on whether a person can find and keep a job safely, pay for housing or arrange childcare. Hospital records capture an outcome near the severe end of the spectrum. Reading them alongside the proposed mechanism makes it possible to take the effect of economic opportunities seriously without inferring an individual’s choices from a county total.

Spain’s Organic Law 1/2004 supplies a policy comparison through its employment provisions. The study and the law address different countries and questions. Reading them together shows why safety can depend on the economic means to seek help or leave.

Original text