Published Papers

The Contribution of Infrastructure Investment to Britain’s Urban Mortality Decline 1861-1900.

Economic History Review, 72(1) pp 233-259, February 2019.

Awarded 2020 T.S. Ashton Prize for best paper in the Economic History Review by a junior scholar.

Download Paper

Idea: Sanitation infrastructure significantly reduced urban mortality from waterborne, and maybe airborne, disease.


It is well-recognized that both improved nutrition and sanitation infrastructure are important contributors to mortality decline. However the relative importance of the two factors is difficult to quantify, since most studies are limited to testing the effects of specific sanitary improvements. This paper uses new historical data regarding total investment in urban infrastructure, measured using the outstanding loan stock, to estimate the extent to which the mortality decline in England and Wales between 1861 and 1900 can be attributed to government investment. Fixed effects regressions indicate that infrastructure investment explains approximately 22-25% of the decline in mortality between 1861 and 1900, once time trends are accounted for. Since these specifications may not fully account for the endogoneity between investment and mortality, I perform additional specifications using lagged investment as an instrument for current investment. These estimates suggest that government investment was the major contributor to mortality decline, explaining up to 60% of the reduction in total urban mortality between 1861 and 1900 and 88% between 1861 and 1890. Additional results indicate that investment in urban infrastructure led to declines in mortality from both waterborne and airborne diseases.

Democratic Reform and Opposition to Government Expenditure: Evidence from Nineteenth-Century Britain.

Quarterly Journal of Political Science, Vol. 13: No. 4, pp 363-404 (2018).

Download Paper

Idea: Spending on public goods higher where elites are middle class. Test using 1894 legislation as a natural experiment.


Several theories have argued that democratic reform will lead to higher government spending. However, these theories have generally focused on expenditure on redistribution rather than expenditure on public goods. This paper presents a model predicting that democratization leads to lower government expenditure on infrastructure if the median pre-reform voter is middle class. This prediction is tested using a new panel dataset of town council infrastructure spending and revenue in nineteenth-century Britain. An 1894 national reform implementing a system of “one-household-one-vote” and the secret ballot is used as the treatment event in a difference-in-difference analysis. The results show that democratic reform slowed the growth of town council investment in public goods, including water supply and other public infrastructure. In line with the theoretical prediction, this negative effect was strongest when democratic reform transferred power from the middle class to the poor.

Saving Face Through Preference Signaling and Obligation Avoidance.

(with Matthew Chao)

Forthcoming at Journal of Economic Behavior & Organization

Idea: Some individuals pay to publically signal reciprocity, even at the cost of the person they are reciprocating to.


Many individuals act more selfishly in games when actions are hidden and their image is not at risk. However, some individuals may still desire to publicly signal reciprocity or other socially desired behavior in these contexts. These individuals may view hidden actions not as an opportunity to act selfishly, but rather as an obstacle to signaling preferences or type. Study 1 tests this by implementing a trust game where nature stochastically intervenes and allocates nothing in place of the second-mover’s choice. When nature intervenes, many second-movers choose to sacrifice pay in order to truthfully signal that they attempted to allocate more, and that they therefore tried to reciprocate. Since signaling can be costly, Study 2 tests whether some individuals strategically reject interactions that could necessitate this type of signaling response. Players play two rounds of dictator games of increasing size, swapping roles in between. In treatments that allow it, many players reject allocations from their partner in the first round; they then act more selfishly as the dictator in the subsequent, higher-stakes round. Together, these results emphasize that the need to signal reciprocity or other socially desired behavior can influence how people engage with and respond to others in strategic contexts.

Working Papers

The Extent of the Franchise and Government Spending: A Non-Monotonic Relationship in Nineteenth-Century England.

(Previously: Extension of the Franchise and Government Expenditure on Public Goods: Evidence from Nineteenth-Century England)

Idea: After a certain point extending voting rights to poorer citizens leads to lower spending on public goods.


Working Paper


This paper investigates how extending the franchise affected government spending in British towns between 1867 and 1900. The effect of franchise extension is identified by exploiting extensive regional and temporal variation in the right to vote. The results show robust evidence of an inverted U-shaped relationship: extending the right to vote from the rich to the middle class led to increased spending, but further franchise extensions—beyond around 50% of the adult male population—led to lower expenditure. Further, government spending was lower in towns where the poor were enfranchised by national reforms, in contrast to the predictions of many theoretical models of democratization. Analysis of local newspaper reports also demonstrates that working class citizens resisted higher expenditure. A simple model, supported by historical evidence, explains these results by emphasizing the trade-offs between public spending and private consumption faced by poor voters. Many theories of democratization suggest that extending the right to vote will lead to increased government expenditure (e.g. Meltzer and Richard, 1981; Lizzeri and Persico, 2004; Acemoglu and Robinson, 2000), these results suggest that may only be true when government has the ability to engage in widespread redistribution.

Democracy, Redistribution, and Inequality: Evidence from the English Poor Law.

Under Review

Working Paper

Idea: Democratization led to higher welfare spending in high inequality areas, consistent with theoretical models.


This paper tests the relationship between inequality, democratization, and expenditure on poor relief in England and Wales between 1885 and 1905. Poor relief served as the main form of social insurance at that time and, in contrast to modern-day welfare programs, was provided by elected local governments. As a result, policy varied substantially across the country in terms of both the magnitude and nature of the support provided. Prior to 1894, a number of institutional features —a graduated franchise, the absence of a secret ballot, and the participation of unelected magistrates—helped landowners to control these local councils, and hence poor law policy. These advantages were removed by a national reform in 1894, an event which serves as the treatment event in a difference-in-difference analysis. The analysis tests whether the effect of this reform varied according to the level of inequality in each district, as predicted by theoretical models of democratization. The results show that, consistent with these theories, unequal districts experienced greater increases in poor law expenditure post democratic reform.

Financing Sanitation Infrastructure in Nineteenth-Century England and Wales.

Under Review

Working Paper

Idea: High costs of borrowing deterred investment in sanitation infrastructure such as sewer systems and clean water supply.


This paper investigates the role of high borrowing costs in deterring sanitation investment in late nineteenth-century Britain. Many councils were not providing public goods such as water supply and sewer systems even at the end of the century, despite having had access to government loans since the 1860s. Using an annual dataset of the financial accounts of almost seven hundred town councils, the paper identifies significant variation in the interest rates that towns had to pay when borrowing to fund investment. Panel regressions show that higher interest rates were associated with significantly lower levels of sanitation investment, with the relationship robust to controlling for town tax base, non-tax revenue sources, demographic characteristics, and town fixed effects. The regression estimates imply that providing loans at the government’s cost of borrowing would have increased the stock of sanitation infrastructure in 1903 by around 25%, and so potentially hastened Britain’s mortality decline.

Willingness-To-Pay and Willingness-To-Accept are Probably Less Correlated than You Think.

with Mark Dean, Pietro Ortoleva, Erik Snowberg, and Colin Camerer.

Working Paper (January 2019 - Revisions requested at Econometrica.)

NBER Paper (May 2017)

Idea: Willingness to Pay and Willingness to Accept are very poorly, if at all, correlated in representative samples. Implications for theories of reference dependence.


An enormous literature documents that willingness to pay (WTP) is less than willingness to accept (WTA) a monetary amount for an object, a phenomenon called the endowment effect. Using data from an incentivized survey of a representative sample of 3,000 U.S. adults, we add one (probably) surprising additional finding: WTA and WTP for a lottery are, at best, slightly correlated. Across all respondents, the correlation is slightly negative. A meta-study of published experiments with university students shows a correlation of around 0.15-0.2, consistent with the correlation in our data for high-IQ respondents. While poorly related to each other, WTA and WTP are closely related to different measures of risk aversion, and relatively stable across time. We show that the endowment effect is not related to individual-level measures of loss aversion, counter to Prospect Theory or Stochastic Reference Dependence.


with Mark Dean, Pietro Ortoleva, Erik Snowberg, and Colin Camerer.

Working Paper (January 2019 - Revisions requested at Journal of Political Economy.)

NBER Paper (August 2018)

Idea: Finds clear patterns of correlations betweeen behavioral economic measures in a large-scale representative-sample incentivized survey.


We study the pattern of correlations across a large number of behavioral regularities, with the goal of creating an empirical basis for more comprehensive theories of decision-making. We elicit 21 behaviors using an incentivized survey on a representative sample (n=1,000) of the U.S. population. Our data show a clear and relatively simple structure underlying the correlations between these measures. Using principal components analysis, we reduce the 21 variables to six components corresponding to clear clusters of high correlations. We examine the relationship between these components, cognitive ability, and demographics, and discuss the theoretical implications of the structure we uncover.

Loss Attitudes in the U.S. Population: Evidence from Dynamically Optimized Sequential Experimentation (DOSE).

with Erik Snowberg, Colin Camerer and Stephanie Wang (Previously “Dynamically Optimized Sequential Experimentation (DOSE) for Estimating Economic Preference Parameters”.)

Working Paper (March 2019)

NBER Paper (November 2018)

Idea: Only around half the U.S. population are loss tolerant. Introduces new technique (DOSE) to estimate economic preference parameters.


We introduce DOSE—Dynamically Optimized Sequential Experimentation — and use it to estimate individual-level loss aversion in a representative sample of the U.S. population (N=2,000). DOSE elicitations are more accurate, more stable across time, and faster to administer than standard methods. We find that around 50% of the U.S. population is loss tolerant. This is counter to earlier findings, which mostly come from lab/student samples, that a strong majority of participants are loss averse. Loss attitudes are correlated with cognitive ability: loss aversion is more prevalent in people with high cognitive ability, and loss tolerance is more common in those with low cognitive ability. We also use DOSE to document facts about risk and time preferences, indicating a high potential for DOSE in future research.

Work in Progress

The Historical Development of Fiscal and Legal Capacity in England and Wales, with Tim Besley, Dan Bogart and Nuno Palma.

The Impact of Inequality on Support for Franchise Extension: Evidence from Britain’s Age of Reform.

Inequality and Welfare Expenditure: Evidence from the New Poor Law.

The Effect of the 1984/85 British Miners’ Strikes on Social Cooperation: There’s No Such Thing as Society Anymore?, with Samantha Myers.

Econographics and Political Behavior, with Colin Camerer, Pietro Ortoleva and Erik Snowberg.

Time Stability of Econographics, with Colin Camerer, Pietro Ortoleva and Erik Snowberg.

Other Publications

Book Review of The River Pollution Dilemma in Victorian England: Nuisance Law Versus Economic E.fficiency. By Leslie Rosenthal. Farnham: Ashgate, 2014. Australian Economic History Review, November 2015.

Book Review of The Rise of a Victorian Ironopolis: Middlesbrough and Regional Industrialization. By Minoru Yasumoto. Woodbridge: Boydell, 2011. Published in the Journal of Economic History, September 2013.

An economic impact assessment of the CCPMO with Gavan Conlon and Patrice Muller (London: London Economics, 2008).