The Causal Effect of Pet Ownership on Health and Well-being

Despite a large strand of literature investigating how pet ownership is associated with different health and well-being outcomes, almost all studies are cross-sectional and use non-random samples. Using Australian longitudinal data and leveraging a policy change that allowed renters in the state of Victoria to get pets but did not apply to renters in other parts of the country, we estimate the causal effect of obtaining a pet on life satisfaction, loneliness, mental health, and general health. Using the stability-controlled quasi-experiment (SCQE) estimator that provides partial identification within a quasi-experimental study, under plausible assumptions we find no significant effects of pet ownership on any of the outcomes considered. To obtain a significant effect, one would require underlying trends in the outcomes to be order of magnitudes greater than those observed in benchmark groups: renters in unaffected states and non-renters in Victoria. The null effects from our causal identification contradict existing work, based primarily on selective cross-section data, as well as the general belief that pets improve individual well-being.