Research

An increasing number of young university graduates are not entering professional occupations, raising concerns about a decline in their skills compared to earlier cohorts. This perspective may overlook evolving occupational returns to skill and job amenities. I develop an economic model featuring heterogeneous skill supply, differentiated returns to skill across occupations, and changing preferences over non-wage job characteristics. Estimating the model using UK data from 2001-2019, I find a significant decline in average graduate skill levels – about 42% of a standard deviation – which explains the majority of reduced professional employment. Additionally, changes in amenity values push graduates into routine occupations, while lower returns to skill in service occupations create growing demand from low-skilled graduates.

This paper investigates the short-run effects of an aspiration-raising intervention delivered as part of a randomized controlled trial among postgraduate students at a UK university during the Covid-19 pandemic. We document suggestive evidence that a video-based role-modeling intervention led to an immediate increase in aspirations and a delayed increase in self-reported effort. However, both effects dissipated within a few weeks. Our findings suggest that timing and reinforcement are important considerations for the sustained effectiveness of aspiration-building strategies.

Understanding the extent and evolution of pandemic-induced mortality risk is critical given its wide-ranging impacts on population health and socioeconomic outcomes. We examine empirically the persistence and scale of influenza mortality risk following the main waves of influenza pandemics, an analysis which is missing from existing research despite its importance for understanding the true scale of pandemic-induced risk. We provide evidence from municipal public health records that multiple recurrent outbreaks followed the main waves of the 1918-19 pandemic in eight large cities in the UK, a pattern we confirm using data for the same period in the US and data for multiple influenza pandemics during the period 1838-2000 in England and Wales. To estimate the persistence and scale of latent post-pandemic influenza mortality risk, we model the stochastic process of mortality rates as a sequence of bounded Pareto distributions whose tail indexes evolves over time. Consistently across pandemics and locations, we find that influenza mortality risk remains elevated for around two decades after the main pandemic waves before more rapid convergence to background influenza mortality, amplifying the impact of pandemics. Despite the commonality in duration, there is heterogeneity in the persistence and scale of risk across the cities, suggesting effects of both immunity and socioeconomic conditions.

We develop a novel annual measure of current and expected financial constraints for publicly listed US firms over 1993 to 2024. Applying artificial intelligence to 10-K filings text enables more accurate and context-aware detection of financial constraints than traditional text classification techniques. Uniquely, we distinguish constraints affecting firms presently from those anticipated for the future. These constraint types are associated with distinct financial profiles and transition dynamics from which we distill three novel facts: (i) Expected constraints are seldom realized, instead, firms typically become unconstrained or postpone constraints further into the future. (ii) Firms frequently mitigate current constraints within a year, but persistence rises with severity. (iii) Firms prioritize resolving immediate over future constraints. Notably, timing-related heterogeneity impacts the practical application of the widely-used cash flow sensitivity of cash: while it identifies anticipated future financial constraints, it conflates distinct constraint types – unconstrained and currently constrained – and therefore fails to capture all financially constrained firms. A general implication of our work is that firms observable financial decisions remain informative for identifying financial constraints as liability-cashflow sensitivities distinguish unconstrained from currently binding constraints.

Group sorting segments housing markets into thin sub-markets, but its pricing consequences are theoretically ambiguous. Leveraging exogenous variation in neighborhood composition induced by British colonization of Northern Ireland during the early 1600s, we find a one standard deviation increase in neighborhood heterogeneity raises house prices by 9.6%. The premium is larger in inherently illiquid market segments. Consistent with a liquidity mechanism, properties in mixed neighborhoods attract more viewers and cross-group participation, sell faster, exhibit tighter listing-to-sale spreads and lower neighborhood-level price volatility. Property title deeds provide complementary evidence on the mechanism by showing that transactions in more heterogeneous neighborhoods are less likely to occur within groups, that counterparty identity does not systematically predict transaction prices, and that same-group transactions are associated with lower holding-period returns, consistent with trading in thinner and less liquid markets. Out-of-sample estimates from a different jurisdiction yield similar inferences. The mechanism we identify — an illiquidity discount driven by demographic segmentation of the buyer pool — represents a novel source of systematic pricing variation in the largest asset class most households hold.

Keywords: Market segmentation, House prices, Liquidity, Segregation.

The main waves of a pandemic and subsequent disease outbreaks in the following years influence the evolution of the distributions of health and wealth, leading to differences in the ability to mitigate future income shocks. We study consumption smoothing and precautionary behaviour associated with the main pandemic waves and recurrent outbreak risk in a model in which health and wealth are jointly determined under income and health risk that are related to disease outbreak risk. We calibrate the model to the UK and find that the impact shock of COVID-19 and recurrent outbreak risk amplify existing inequalities in wealth and health, implying persistent increases in wealth inequality that are characterised by increases in wealth for households in higher income groups and/or with higher initial wealth, and decreases for those in lower income groups and/or with lower wealth. These changes lead to inequality in exposure to post-pandemic income risk and, in particular,
an increase in the vulnerability of those already with very little wealth prior to
the pandemic. We assess public insurance policy to mitigate income losses for those with low wealth and find that, by disincentivising wealth accumulation and incentivising investment in health for those with low wealth and health, it reduces health inequality and, in the short run, the probability of low consumption, but increases wealth inequality and, in the medium run, the probability of low consumption.

Wealth and human capital are strongly intertwined. But what effect does the distribution of one have on the other? In this paper, I develop a general equilibrium heterogeneous agent incomplete market model with endogenous wealth and human capital to analyse the interactions between these two factors. I calibrate the model to the UK economy in the pre-Covid19 period and analyse the interaction of wealth and human capital in the stationary equilibrium. I find that there are important non-linearities in human capital investments, with workers with low levels of wealth investing considerably less in accumulating human capital than their counterparts with more wealth. I then analyse the economic dynamics of the distribution of human capital in the aftermath of an unexpected economic shock, showing that wealth poorer households are more exposed to these shocks, implying that the distribution of wealth matters for the recovery of the economy following recessions.
Finally, I assess the impact of the Covid19 pandemic and associated support measures in the UK. The model predicts that the UK economy will likely suffer a significant reduction in human capital in the aftermath of the Covid19 pandemic, but targeted policy action has helped to reduce the impact of the crisis particularly for low wealth households.

This paper proposes a model to study the effect of differentiated, cognitive skill biased, technological change on income and wealth inequality. It is one of the first attempts to combine elements of the “Task-Skill” literature with a heterogenous agent incomplete market model. The model includes a structural production environment that accounts for differentiated skill demand on the firm side and multidimensional skill supply on the workers side. Using measures of cognitive and non-cognitive skills from a comprehensive panel dataset for the UK (Understanding Society), I calibrate the model with appropriate micro-estimates. The calibrated model manages to capture many of the features of the income process observed in the data and provides additional features beyond other, more commonly used approximation techniques. I then use the model to assess the impact of cognitive task biased technological (CBTC) change due to increased Computer usage in the UK over the period 1980 – 2016. The model suggests that CBTC can account for the bulk of increases in labour income inequality observed over that period, and is generally consistent with stylized facts about changes to wealth inequality.