Parenthood and legislators' temporal focus: evidence from the UK House of Commons, 1946 - 2022

Many important societal problems concern the future. It therefore matters whether the temporal focus of our elected representatives centres on the past, the present, or the future. Temporal focus can change as a result of contextual factors or as a result of changes in representatives' interests. One way in which representatives' interests, both in general, and regarding the future, can change is through parenthood. Individuals who become parents have an interest in their child's wellbeing which extends beyond their own lifespan. In our paper, we test whether experiencing parenthood for the first time causes a change in temporal focus. We study temporal focus through spoken language in the UK House of Commons which, because of its large size and long history, allows us to identify the effects of parenthood quite precisely. We base our measure of temporal focus on a large language model fine-tuned with approximately 3,600 hand-coded sentences. This gives our the proportion of legislative speech focused on the future for each legislator/month from 1945 to the present. We then combine this with information on MPs' parenthood. We focus specifically on two groups of MPs: MPs who became parents for the first time after entering the House of Commons, and MPs who never experienced parenthood during their time in the Commons. By focusing on these "treatment" and "control" groups, we can identify the effect of parenthood comparing levels before and after first experience of parenthood, comparing these before and after trends to trends in individuals who did not experience parenthood. Our design allows us to identify a main effect of first parenthood in general, rather than an effect of becoming a mother or becoming a father; this average effect may therefore be thought of as a weighted combination of these two distinct effects. Our research contributes in three different ways. First, we present new data on the experience of first parenthood amongst parliamentarians in a large national legislature. Second, we show how large language models can be fine-tuned to provide insight into substantively interesting patterns of legislative speech beyond the usual applications (sentiment, position). Third, we tackle the link between parenthood and representation for a group (future generations) which is important (and will be more important), but which is generally difficult to study.