State-owned banks play a central role in stabilizing credit markets during downturns. Their public policy mandate, however, gives governments greater influence over lending decisions than in privately owned banks, raising concerns that lending decisions may reflect political objectives alongside economic ones. This paper studies how electoral incentives interact with the countercyclical role of public banks. We develop a political-economy framework showing that electoral incentives depend on the type of credit recipient. Household and firm lending generates electoral returns by relaxing credit constraints and improving economic conditions among marginal voters, whereas municipal lending operates through visible public investment and the mobilization of incumbent supporters. The model therefore predicts different geographic patterns across lending activities. Using bank–municipality–month data and a triple-difference design comparing public and private banks within the same local credit markets, we show that public banks expand retail and business lending during presidential election campaigns in politically competitive municipalities, mainly during aggregate credit contractions. No comparable response is observed during mayoral elections, when the federal government lacks direct electoral incentives. In contrast, using administrative data on municipal borrowing operations, we find that public-bank lending to municipalities is concentrated in areas with low political turnover and limited electoral competition, consistent with a mobilization motive. Additional evidence shows that electoral lending operates through both quantities and prices as public banks increase credit supply while reducing lending rates relative to private banks around presidential elections. These expansions have substantial balance-sheet consequences. The deterioration in loan quality implies losses between 1.3 to 5.7 cents per additional dollar lent, corresponding to losses of approximately $3.7 billion per presidential election cycle. The findings show that electoral incentives do not replace the stabilization role of public banks but shape where stabilization benefits are allocated. Political considerations therefore influence both the distribution and the financial sustainability of public credit.

Changes in relative energy prices are central to the green transition, but their effectiveness depends on firms’ ability to adjust their energy input mixes. This ability is captured by elasticities of substitution across energy inputs—parameters for which we lack empirical knowledge. Using firm-level microdata from several European countries, I estimate elasticities between clean and dirty energy, as well as among individual energy sources (e.g., electricity, natural gas, coal, etc.). They have declined over the past decades, vary significantly across sectors and countries, and are systematically lower among larger firms. I then embed these estimates in a standard model of technical change to simulate countries’ decarbonization paths. I document large cross-country differences in the reduction of dirty energy shares and the underlying drivers. Between 50 and 85 percent of the reduction is accounted for by within-firm substitution, with the remainder driven by between-firm reallocation, entry, and exit.

Carbon pricing is intended to reduce emissions by raising the cost of carbon-intensive energy. Its effectiveness depends on how firms adjust and if price signals reach the firms responsible for the largest share of energy use in the economy. We study the impact of carbon pricing on manufacturing firms' margins using microdata from France, Portugal, and Slovenia. Exploiting differences in firms' pre-existing fuel mixes and policy stringency shocks, we and that the most exposed firms to carbon pricing reduce their relative use of dirty to clean energy and substitute toward electricity, with no detectable effect on their productivity. However, such a response is considerably weaker among the largest energy users. These firms respond 47 to 85 percent less to carbon pricing than the rest of the firms. Our results suggest an incomplete pass-through of carbon costs to large rms. Since the largest firms account for most energy use, this heterogeneity substantially reduces the aggregate response to carbon pricing and reverses its effect on some adjustment margins. Carbon pricing therefore has its weakest effects among the firms with the greatest scope for abatement.

We develop a novel product-level measure of manufacturing energy use by combining firm-level energy records with product-level production data for France, Slovenia, and Portugal. This allows us to measure energy use across products and, critically, across firms producing the same product. Energy use is highly concentrated in a small set of energy-intensive products, particularly in basic metals, non-metallic minerals, paper, and chemicals. While energy use rises sharply with firm size, energy intensity per unit of turnover is relatively stable. Most importantly, we find substantial energy-use heterogeneity within narrowly defined product markets: the median within-product ratio of the 75th to 25th percentile of firm energy use is approximately 14 in France and 13 in Portugal. This suggests that a large component of energy-productivity dispersion occurs among firms producing the same products, rather than across sectors or product mixes. The results provide new evidence for product-level analysis of manufacturing decarbonization and energy-efficiency policies. 

How do energy-price shocks affect firms' demand for labour? Using firm-level data for manufacturing firms in France, Portugal, and Slovenia, we estimate the elasticity of substitution between energy and labour and identify firms' employment and wage responses to energy-price changes. We instrument firm-level energy prices using sector-level price shocks interacted with firms' initial price exposure. Labour and energy are gross complements, with an estimated elasticity of substitution below one in all three countries. A 10% increase in energy prices reduces employment by approximately 0.8–2.0% in France and 0.4–0.9% in Portugal, while wages fall by around 1–1.5%. These effects are substantially larger among firms with lower substitution elasticities: employment falls by approximately 5% in France and 8% in Portugal for this group. Standard firm characteristics explain little of the heterogeneity in substitution elasticities. The results show that the labour-market incidence of energy-price shocks depends critically on firms' ability to reorganise production.