Working Papers
Assortative Matching and Human Capital Investment
I develop a dynamic model of sorting between workers and firms in which it is possible to endogenously invest in the worker's general human capital. The capability of a worker-firm pair to produce both tradeable output and further human capital may depend non-parametrically on both the worker's current human capital and firm type. Supermodularity of the technology with respect to the types does not suffice for strict positive assortative matching (PAM) in the competitive equilibrium; much stronger assumptions must be imposed. If high-productivity firms are better at training and production is concave in human capital, then PAM is not guaranteed even if production is supermodular in the types. In particular, with enough concavity, the importance of getting low-skilled workers paired with the best firms may outweigh the effects of supermodularity. With simple examples, it is shown that randomisation in the matching process may be an endogenous outcome, even in the absence of search or informational frictions. I prove that under weaker conditions, it is sometimes possible to determine whether the correlation between worker and firm type will be positive or negative, without any knowledge of the distribution of types. Furthermore, I prove that under some conditions, workers sort in a manner such that the highest skilled workers see their wages increase at the fastest rate, giving firms a highly active role in the dispersion of wages and inequality over the life cycle.
Employer-Provided Training in the UK: A Search-Theoretic Model
I document employer-provided training for full-time employed workers in the UK using an ``effective training" measure that weighs off different types of self-reported work-related training. This form of training tends to be higher in already highly educated workers and is provided in greater amounts at larger firms. Moreover, occupations and industries that tend to pay higher wages also tend to provide more training; this is consistent with the idea that training enhances the productivity of the worker and that workers with high earning ability may sort into high training environments. In conjunction with these findings, I develop a search model of the labour market that includes heterogeneity in both workers and firms and endogenous human capital distribution. Workers vary in their level of human capital and firms vary in productivity. Worker-firm pairs can increase the worker's human capital at the cost of losing output. I show that this framework can replicate key facts from the data; namely, higher educated workers receive more training throughout their lifetime and earn more, and that the firms that pay higher wages also provide more training. Finally, the model features inefficiently low human capital investment due to the social returns not being fully internalised under random search; a policy of subsidising low-skilled young workers covered by income taxation is shown to improve aggregate welfare and social mobility in the model.
Implicit Contracts and Asymmetric Pass-Through of Productivity Shocks
with Andy Snell (UoE), Heiko Stüber (UALS), Jonathan P. Thomas (UoE)
We document distinctive empirical features of wage pass-through in Germany that are consistent with Thomas-Worrall wage contracting in the presence of both idiosyncratic and nonstationary aggregate productivity components. These empirical features are hard to reconcile with the predictions of search models based on period-by-period Nash bargaining over match surplus and with the predictions of financial models where risk neutral firms may costlessly shield risk averse workers from idiosyncratic shocks (Guiso, Pistaferri et al. 2005).