Research
Taxation and technology under autocracy, measured on the state's own farm-level records; technology and contracts in a market.
Research statement (PDF) · Curriculum vitae (PDF)
Job market paper
A state that cannot verify the harvest it taxes chooses between a crop share, which invites a reporting game between the farm and the assessor, and a fixed rent per hectare, which cannot be moved by a report of the harvest and cannot follow it: a fixed rent taxes uniformly farms whose harvests differ. In January 1933 the Soviet state replaced its claim on the assessed grain harvest of collective farms with a fixed rent per sown hectare, one rate for every farm of a district. On the accounts of 272 collective farms of the Urals for 1933 and 1934, I measure what uniform taxation cost in people and divide the cost between information and incentives. The sequence of the agricultural year fixes what each party observed when it acted and hence the form of each claim; the rules estimated from the accounts, the district office's schedule, the farm's division of grain between seed and food, the seed relief and the response of population to food, convert any schedule of deliveries into members a year later. Relative to the same statutory total allocated on harvests with subsistence first, the policy as run cost 4,392 people, one member in twenty-eight: 707 from taxing uniformly farms whose harvests differed, 2,755 from the level of the wedge, the offices' collection above the norm, and 930 from its allocation across farms the office could not tell apart. Relief that came as seed in the spring rather than food in the autumn cost a further 964.
Working papers
A state that cannot verify its producers' harvest taxes what it can see. Soviet grain procurement of 1932 was written on the sown area, on past deliveries and on a yield assessed from the standing crop, and it was collected by local officials who answered for a district's quota. This paper asks whether that tax, as it was actually collected, was enough to produce the famine of 1933. In the state's own post-harvest surveys of collective farms, recorded by region, the harvest of 1932 was in most regions enough to feed the farms' members; after the state's claims most of their farms fell below subsistence, and the depth of the shortfall predicts the rise in mortality between regions under every allocation of food within a region that the printed distributions admit. Ukraine, where the famine struck hardest, lost no more to the extraction than the other grain-producing regions, and the schedules make no reference to nationality. Within regions, districts with more Ukrainians had higher mortality, a difference the regional data cannot test; the archives point to the local campaigns of enforcement, the conditioning of relief and the restriction of movement, which no survey measures.
We examine how Big Push policies perform under the constraints of extractive institutions. Using novel micro-data from an All-Union audit of collective farms, we exploit the rollout of Machine Tractor Stations, the central pillar of Soviet agricultural policy, to estimate the impact of mechanization on collective farm performance, state procurement and peasant welfare in the early Soviet command economy. The Soviet system rapidly expanded technological adoption but failed to realize the full gains of increasing-returns technologies; adoption also did not raise living standards, leaving peasants on mechanized farms at or below subsistence levels.
This paper examines how technology adoption and contractual coordination jointly modernize an agricultural supply chain, using the Wisconsin fresh potato industry from 2010 to 2019, where a wholesaler invested in optical sorting systems and then entered long-term marketing contracts with a retail chain. A random-coefficients demand model and a difference-in-differences design show that the sorting technology lowered marginal cost by about 4.6 cents per pound, roughly 10 per cent of the price, and that the contracts added about 1.5 cents to price-cost margins, with the gains reinforced by improved inventory management and reduced spoilage.
Ongoing research
Repression and the collective farm in the Urals, 1930–1934 (data in progress)
My ongoing research turns to the other instrument of the autocratic state, repression. In the early 1930s the centre set for each region the number of households to be removed as kulaks and left the choice of whom to the district: a similar delegation problem. I have assembled a registry of the persons dispossessed or disenfranchised in the Ural region from a database built on the district archives, which covers rural repression more completely than the widely used Memorial database, and matched it to the 1928 settlement census; where the database leaves a part of the region uncovered, I am reading the archives myself. Joined to the farm accounts, the registry will show who was removed and what their removal did to the farms; my aim is to bring the state's capacity to repress into the same account as its capacity to extract. Alongside it I am extending the coverage of the district panel behind these papers, which follows the state's agricultural policy across the Urals year by year.
Presentations
2026 — Economic History Association Annual Meeting, St. Louis (poster)
2026 — Summer Workshop in the Economic History and Historical Political Economy of Russia, University of Chicago Center in Paris
2025 — Central Asian Economics Conference, Almaty
2024 — Agricultural and Applied Economics Association Annual Meeting, New Orleans
2023 — Economic History Association Annual Meeting, Philadelphia