Adrien Couturier

Why do financial crises cause such deep recessions, and what can policy do about them? We distinguish between financial intermediaries’ capacity to lend and their willingness to lend in a general-equilibrium model with heterogeneous intermediaries and occasionally binding leverage constraints. Two moments determine whether forward-looking intermediaries with room to lend offset constrained intermediaries’ retrenchment and dampen downturns, or cut lending and amplify them: how strongly lending responds to expected returns and how those expectations respond to aggregate shocks. We estimate both using 40 million bank forecasts of borrower default probabilities. First, comparing banks assessing the same borrower at the same time, we show that a 1 p.p. increase in expected portfolio returns raises lending by 12%. Second, expectations overreact to aggregate shocks, leaving banks excessively pessimistic following adverse news. These two moments allow the calibrated model to generate dynamics for aggregate output during crises similar to those of a model driven by balance-sheet constraints alone, but with sharply different policy implications. In a crisis, the marginal intermediary is unconstrained and cuts lending because it expects low returns. Recapitalizations provide limited stimulus, whereas loan guarantees raise expected returns directly and stimulate output four times as much at the same fiscal cost. Our estimates also explain how booms build fragility: optimism raises leverage in booms, so a larger expansion deepens the subsequent recession.

How does government spending stimulate the economy? We uncover a new transmission channel—--the sentiment channel—--by showing that a rise in government spending makes firms overoptimistic about their future demand, thereby stimulating investment and output. We assemble a novel dataset linking microdata on Italian firms’ sales, sales forecasts, and public procurement contracts. Using a natural experiment that shifted public spending across municipalities, we find that an increase in government demand makes firms systematically overoptimistic about their future sales. This overoptimism is pervasive, as firms also raise their expectations about export sales. To interpret these findings, we develop a theory of expectations in which shocks to total sales make a firm overoptimistic about both its public and private sales. We embed this model of expectations in a heterogeneous-firm New Keynesian model disciplined with our causal estimates. Government spending boosts firms’ optimism which prompts them to invest, crowding-in investment in general equilibrium, thereby producing a 75% increase in the government spending multiplier. The sentiment channel is state-dependent: tighter financing conditions limit firms’ ability to act on improved expectations and weaken its contribution to the fiscal multiplier.

How household greenhouse gas emissions scale with income is central to forecasting the environmental consequences of global growth. We construct a new cross-country database that harmonizes household consumption surveys for twenty-four countries with global data on direct and indirect emissions, enabling the first consistent estimation of environmental Engel curves worldwide. We show that the global elasticity of GHG emissions with respect to household expenditure is around 0.81 when evaluated using a common world-average vector of supply-chain carbon intensities per euro of expenditure. Decompositions highlight shifts across broad consumption categories and substitution toward less GHG-intensive goods within utilities as forces flattening the curve. Using retail scanner data and detailed product-level emission data, we show that quality differences and within-category heterogeneity generate only minor biases. Thus, global GHG Engel curves are largely shaped by broad consumption patterns. Overall, the contribution of demand-side non-homotheticities to long-run decarbonisation is comparable to the contribution of an entire “hard-to-abate” sector like cement, steel, or global aviation and shipping combined.

We study how regional heterogeneity shapes the aggregate transmission of monetary policy and its distributional implications across space. We build a multi-region Heterogeneous-Agent New Keynesian model with 3,140 U.S. counties and cross- county differences in (i) intertemporal Marginal Propensities to Consume (MPCs) and (ii) non-tradable employment shares. We analytically characterize the nationwide consumption response to monetary policy in terms of the joint distribution of (i) and (ii). Using U.S. and Italian micro-data, we construct novel empirical measures of regional MPCs to validate our theory. Quantitatively, geographic heterogeneity leads to large distributional consequences of monetary policy across space and can sizably amplify its aggregate effects.

We embed diagnostic expectations into an otherwise standard incomplete-markets model of consumption-saving with idiosyncratic income risk. In this framework, households form beliefs that overweight recent income shocks, a bias we summa- rize with a sufficient statistic—sentiment—that distorts perceptions of future income. We discipline our model empirically using the Italian Survey of Household Income and Wealth. Taking advantage of a continuous-time formulation, we derive a closed- form characterization of how sentiment dampens the saving motive under optimism and amplifies it under pessimism, causing households to overreact to income shocks in their consumption-saving choices. We then show that the interaction of sentiment with borrowing constraints generates a “diagnostic poverty trap”: positive shocks fuel over-consumption rather than asset accumulation, making it harder for constrained households to escape the hand-to-mouth state. This simple behavioral friction ratio- nalizes the persistence of hand-to-mouth households observed in the data and helps match their empirical prevalence without invoking illiquid assets or preference het- erogeneity.

We develop a multi-country Heterogeneous Agents New Keynesian (HANK) model of a monetary union with ex-ante heterogeneity in legacy public debt across mem- ber states. Following symmetric aggregate shocks, the systematic monetary policy reaction induces heterogeneous responses driven by national fiscal space. This gen- erates a trade-off between union-wide macroeconomic stabilization and cross-country synchronization of economic activity for the central bank. We characterize a possibil- ity frontier between union-wide inflation stability and cross-country synchronization, which is traced out by varying the degree of the central bank’s hawkishness towards inflation. We study the role of deficit caps, fiscal and political unions, and augmented Taylor rules as instruments to navigate the stabilization-synchronization trade-off.

Ce rapport évalue l'impact du plan France Relance sur le PIB, l'investissement, la consommation et l'emploi, ainsi que leur répartition sectorielle et distributive, à l'aide d'un modèle macroéconomique de type HANK. Les résultats montrent une augmentation moyenne de 0.48\% du PIB par rapport à un scénario de référence (soit 206 000 emplois trimestriels), tirée principalement par l'investissement. Les effets sur le niveau des prix sont faibles en 2023 (hausse de 0.8\% attribuable à France Relance). Les ménages les plus riches, bénéficiant davantage de l'augmentation des revenus du capital, ont tiré un gain disproportionné en termes de bien-être.