Research
Research
This paper estimates the causal effect of temperature and precipitation shocks on provincial GDP per capita growth in Spain between 1904 and 1934, exploiting within-country variation in aridity to identify the structural determinants of climate vulnerability in an organic agricultural economy. Following Nath et al. (2024), we extract weather shocks as innovations from province-specific autoregressions and find that a 1°C temperature shock reduced provincial GDP growth by 3.3–3.4 percentage points, more than twice estimates for modern economies, reflecting the absence of buffering inputs in Spain’s organic farming system. The aggregate null result for precipitation conceals a large effect concentrated in the most arid provinces, where a 100mm shock reduced growth by 13.9 percentage points. Aridity predicts climate vulnerability independently of income, agricultural dependence, land tenure structure, and crop composition. The most arid provinces were vulnerable not because they were poor, but because their ecological structure left dryland farming systems exposed to heat stress and structurally ill-equipped to manage unusually high rainfall.
The Effect of Temperature on Fertility in 19th Century France (Under Review)
France was the first country to undergo a sustained fertility decline, already well underway by the mid-nineteenth century. This paper asks whether climatic variation helped shape fertility levels during this period. Exploiting exogenous temperature variation across 80 French d\'{e}partements from 1851 to 1911, I find that a one degree increase in the five-year average temperature raises the Coale fertility index by 16.6 percent, equivalent to roughly 4 additional births per thousand. Higher temperatures reduce wheat prices, consistent with increased agricultural supply, and the fertility response is strongest in the most agriculture-dependent departments, consistent with the income effect operating through own-account production rather than market revenue. The temperature effect is robust to controlling for internal migration, infant mortality, and female education, consistent with agricultural income operating as a primary channel alongside complementary roles for these factors.
This paper estimates the causal effect of temperature shocks on wages across nine occupations in Spain between 1915 and 1931, tracing how an agricultural supply shock propagated through the occupational structure of an organic farming economy. Following Nath et al. (2024), we extract temperature shocks as innovations from province-specific autoregressions and estimate dynamic wage responses using local projections over a five- year horizon, retaining peak and off-peak wages separately to identify the transmission mechanism. Agricultural wages bear the largest and most persistent losses, falling by 0.153 log points at h = 1 with no reversion across the full horizon. Urban wages follow with a one- to-two-year lag, concentrated in construction trades and craft occupations most exposed to rural purchasing power, while stonemasons and shoemakers are largely unaffected. Nominal wage adjustment accounts for 65 to 94 percent of the real wage decline across affected occupation-horizon cells, with rising prices compounding rather than cushioning nominal losses. We rule out aggregate demand contraction as a mechanism: both rural and urban price indices rise contemporaneously with the shock, inconsistent with a demand-driven fall in wages and prices together. Formal tests of seasonal symmetry find no evidence that peak wages respond more strongly than off-peak wages, rendering seasonal labour market integration an unlikely primary channel. The timing, sequential occupational pattern, and price evidence are consistent with an agricultural supply shock operating through an income channel, in which harvest shortfalls reduce rural purchasing power and contract demand for the goods and services that urban craft workers produce.
This paper constructs and characterizes a provincial-level dataset of seasonal temperature and precipitation covering 48 Spanish provinces over 1500–2000, drawing on the reconstructions of Luterbacher et al. (2004), Xoplaki et al. (2005), and Pauling et al. (2006). Principal component analysis documents a fundamental asymmetry in the dataset’s spatial structure: temperature variation is overwhelmingly driven by a common national signal, while precipitation is substantially more spatially fragmented, particularly in summer. Three historically documented climate episodes are examined against these structural findings and are consistent with them, illustrating the reconstruction’s ability to recover known climate signals at provincial resolution, though as large, well-documented events they offer a necessary rather than sufficient test of reliability. The dataset is best suited to informing the climatic structure of the pre-industrial period directly, and to linking climate to provincial economic outcomes from the late nineteenth century onward, once consistent provincial economic data become available.
Rainfall Shocks and Bank Fragility: Evidence from the U.S. National Banking Era (with Luisa Bicalho-Ritzkat LSE)
Physical climate risk is increasingly viewed as a source of financial instability, yet the channels through which realized climate shocks affect banks remain difficult to identify. We exploit the U.S. National Banking Era (1867–1904), when banks lacked both geographic diversification through branching and access to a lender of last resort, to examine how precipitation shocks affect bank stability. Using bank-level weather variation and 110,831 bank-year observations covering up to 6,525 national banks, we find that adverse precipitation shocks increase the probability of bank failure by 0.19–0.23 percentage points, roughly half the unconditional annual failure rate. A 100mm rainfall shortfall reduces assets by approximately 1 percent, loans by 1.6–1.8 percent, and deposits by 1.2–1.5 percent, while banks subsequently increase their reliance on noncore funding. These responses are consistent with climate shocks generating liquidity pressures that banks must absorb through the market in the absence of a lender of last resort. We further show that favourable rainfall expands bank balance sheets faster than capital, leaving thinner capital buffers