Cats, Dogs, and Babies: Quasi-Experimental Evidence on Substitutes or Complements from Linked Administrative Records

Falling fertility is one of the most pressing demographic and economic challenges of the twenty-first century, particularly in East Asia. Among the many popular explanations, the rise of companion animals has captured public imagination, often portrayed as substitutes for children or as complements in family formation. Yet, despite this widespread narrative, systematic evidence remains scarce. This paper provides the first large-scale empirical analysis combining individual-level administrative data and quasi-experimental methods to examine the relationship between pet ownership and childbearing. We link Taiwan’s universe of pet registry records with administrative tax files, constructing a ten-year panel covering 23 million individuals and millions of registered cats and dogs.

Our analysis proceeds in three steps. First, we document descriptive patterns: pet ownership peaks in the twenties while childbearing peaks later, and cross-sectional correlations differ by species: dogs are positively associated with children, cats negatively. Second, we estimate substitution elasticities. Exploiting the staggered rollout of child bonuses, we find that dog ownership rises by 21% following a birth, yielding a Marshallian cross-price elasticity of −1.32. Using lottery winnings to isolate income effects, we estimate a Hicksian elasticity of −1.29, confirming that children and pets are net complements. Third, event studies show that adopting a dog raises the probability of subsequent childbearing by 33%, consistent with a “practice child” mechanism, while childbirth reduces later pet adoption.

These findings expand the economics of fertility by introducing pets as a previously overlooked determinant of household decisions. They also reveal that pro-natalist policies affect not only fertility itself but related choices such as pet acquisition, broadening our understanding of modern family economics.