Research
Working Papers
I estimate a structural model of sole-parent families to analyse the impact of TANF. Its benefits fail to reach the poorest mothers, who typically prefer not to meet the 30 hr/wk work requirement. Using the model to measure the intra-household allocation of resources, I find that poverty rates among children of sole mothers have risen by four percentage points since 1996. I compare TANF to alternative policies, such as free childcare, which promote labour supply by increasing the returns to work. Such policies are more than twice as effective at targeting household resources to children, per dollar spent.
This paper estimates how effective the major US cash welfare programmes are at targeting household resources to children. To do this I estimate a static intra-household model in which children benefit from private consumption, parental time and a public good. After accounting for the intra-household allocation, I find that the EITC, CTC and SNAP combined bring only 2 per cent of children out of poverty, a number far less than that suggested by official government statistics. I also find 24 percent of spending on these welfare programmes translates into additional household resources for children. In sole-parent households, cash welfare programmes that promote mother's labour supply are especially effective at targeting household resources to children.
Using detailed spatial information from the PSID-CDS, we estimate the relationship between spatial variation in parents’ investment in children’s human capital, government investment in children, neighborhood characteristics and geographical variation in inter-generational mobility across counties in the US (estimates for the latter taken from Chetty, Hendren, Kline, and Saez, 2014). We find that parents’ time spent in educational activities with children under 6 is especially important for explaining mobility in middle-income households: an increase in educational time of 1 hour per week is estimated to increase children’s income at age 30 by $1500 p.a. Our estimation method incorporates a kernel-weighted spatial averaging procedure, which is shown to yield much more informative estimates than does a simple averaging procedure. We supplement this analysis using the ATUS (2003-12), with which we demonstrate that parents spend much more time in educational activities with children in areas of high mobility.
To address McLanahan (2004)'s diverging destinies hypothesis, this paper constructs a monetary measure of children's household resources utilising estimates from a structural household model. Consistent with the literature on this topic, I find that trends in my measure cannot be fully accounted for by the increase in individual wage inequality, but in contrast my results suggest demographic shifts, such as the increased prevalence of sole-parent households, have little effect. Rather, changes to the behaviour of middle- to high-income households are important as they now care more about their children's cognitive development, and devote more household resources to them accordingly.
"TANF and Long-run Outcomes for Recipients' Children" (2020, research supported by The British Academy)
Welfare for sole mothers has changed dramatically over the last two decades, with the introduction in 1996 of TANF. The effects of these reforms on child outcomes are poorly understood. We address this question by examining how maternal labour supply and income change due to the reforms at critical stages of child development. We instrument for maternal labour supply and income using Lasso to select the relevant policy variables in the first stage. Results still under construction.
"The Optimal Design of TANF for Targeting Household Resources to Children" (draft available on request, research supported by The John Fell Fund)
This paper considers the optimal design of TANF policy for targeting household resources to children in sole parent households, building on the work of Blundell and Shepard (2011). Here I adopt the intra-household model estimated in Bruins (2016) where children benefit from private consumption, time with their mother and a home produced public good. TANF policy consists of work requirements, and a disutility from enrolment (parameters are allowed to vary across households with and without children under 6). I also allow income taxation to adjust flexibly to help finance TANF. Finally, optimal TANF policy is compared to the effectiveness of wage subsidies in improving child welfare. Intial results suggest that current work requirements are too high for sole parents, which might explain why enrolment in TANF is so low for these households below the povety line.
Published/Forthcoming
"Women's Economic Opportunities and the Intra-household Production of Child Human Capital" (2017, Labour Economics)
This paper exploits variation in the relative demand for male and female labor during the Great Recession to estimate the effect of woman’s relative economic opportunities on the resources parents allocate to children. Estimates from the ATUS using Bartik instruments suggest the increase in the median female-to-male wage ratio over 2003–12 increased parents’ time with children by 80 min/wk; child test scores are estimated to have consequently risen by 10 percentage points of a standard deviation. Further analysis of our results suggests that an increase in female bargaining power is necessary to explain our empirical findings.
"Generalised Indirect Inference in Discrete Choice Models" with James Duffy, Mike Keane, and Tony Smith (2018, Journal of Econometrics)
This paper develops and implements a practical simulation-based method for estimating dynamic discrete choice models. The method, which can accommodate lagged dependent variables, serially correlated errors, unobserved variables, and many alternatives, builds on the ideas of indirect inference. The main difficulty in implementing indirect inference in discrete choice models is that the objective surface is a step function, rendering gradient based optimization methods useless. To overcome this obstacle, this paper shows how to smooth the objective surface. The key idea is to use a smoothed function of the latent utilities as the dependent variable in the auxiliary model. As the smoothing parameter goes to zero, this function delivers the discrete choice implied by the latent utilities, thereby guaranteeing consistency. We establish conditions on the smoothing such that our estimator enjoys the same limiting distribution as the indirect inference estimator, while at the same time ensuring that the smoothing facilitates the convergence of gradient-based optimization methods. A set of Monte Carlo experiments shows that the method is fast, robust, and nearly as efficient as maximum likelihood when the auxiliary model is sufficiently rich.
Selected work in progress
"The Econometrics of Time-use Data" with James Duffy (Nov 2015)
Time diaries typically record an individuals activities over a 24-hour period. Thus measures of time use constructed by aggregating categories from time diaries are taken over a 24-hour window, when the length of time we are generally interested in is a typical week. Measures are unbiased, but more variable than those corresponding to a weekly diary, resulting in attenuation bias when used as an independent variable in regression analysis. Using an unusual Dutch survey, which includes time diaries for a full week, this paper quantifies the size of this bias for measures of parents’ time with children constructed from diaries over a single day. We find that the variance of these measures is 50% larger than the measure constructed for the full week and subsequently the estimated coefficient on time-use is half the size of the true coefficient. We develop procedures to correct for this bias if, rather than randomly sampling one day per week, the time diary is constructed by randomly sampling two days per week without replacement. With two observations rather than one we are able to estimate the measurement error in the daily diary relative to a weekly diary.
