Abstract This study estimates the effects of an employment programme for disadvantaged unemployed individuals. The programme emphasized on‐the‐job training and contracting the unemployed for a few paid weekly work hours as a stepping stone into the labour market. Evaluated through a randomized controlled trial with 3,727 participants, the programme was found to accelerate transitions into part‐time work. Contrary to its intention, it permanently increased the share of participants receiving disability pensions among most disadvantaged groups. To explain this finding, we suggest that training, while enhancing productivity for some, simultaneously provided information of employability used in the assessment of disability pension eligibility.
Abstract We evaluate the impact of the disclosure of artificial intelligence (AI) involvement in a charitable fundraising experiment, where participants received either AI‐generated or human‐written fundraising letters, with or without the disclosure of their origin. We find that both AI‐generated letters (whether edited by humans or not) and human‐written letters (whether edited by AI or not) significantly increase donations, producing similar positive effects. Importantly, these effects disappear when AI involvement is disclosed, regardless of whether the letters are generated or edited by AI. These findings suggest that, while AI improves the effectiveness of charitable fundraising, disclosing its use offsets this benefit.
In this paper, we investigate changes in time allocated to routine housework around four life events - cohabitation, marriage, first birth, and divorce - to explore the influence of social norms on gendered housework patterns. Using data from the German Socio-Economic Panel (1991-2019), a timing-based event study design reveals that gender differences intensify with cohabitation and parenthood. Women's housework hours increase significantly with the first child, while men's decrease. Marriage has no additional impact beyond cohabitation, and divorce leads to a rise in men's and a reduction in women's housework. Couples with higher-earning women and those in East Germany show weaker adherence to these social roles.
Abstract The Royal Swedish Academy of Sciences awarded the 2025 Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel to Joel Mokyr, Philippe Aghion, and Peter Howitt “for having explained innovation‐driven economic growth”. Mokyr's work explains why sustained growth was historically rare: prosperity required societies capable of generating, diffusing, and applying useful knowledge. The theory of creative destruction established by Aghion and Howitt explains why economic progress is inherently disruptive, as new technologies continuously replace older technologies, firms, and rents. In this paper, I argue that these ideas moved innovation from the periphery to the center of growth economics and recast growth theory as a dynamic process driven by experimentation, rivalry, entry, and reallocation. Recent research has extended these insights in numerous directions, including heterogeneous firms, business dynamism, talent allocation, artificial intelligence, and green innovation. I conclude by arguing that the central challenge of modern growth policy is not simply to accelerate innovation, but to sustain institutions that keep economies open to creative destruction while preserving competition, broad opportunity, and political support for technological change.
This paper examines the impact of importer dispersion on exchange rate pass-through. Theoretically, we demonstrate that greater importer dispersion - characterized by larger variations in import size across importers - leads to higher demand elasticity and lower exporter markups, thereby resulting in greater exchange rate pass-through. Using transaction-level customs data from Colombia, we provide robust empirical evidence supporting this prediction. The quantitative effect of importer dispersion on exchange rate pass-through is significant: the importer dispersion channel is at least as important as the traditional exporter heterogeneity channel. Our results are robust to various empirical specifications and become even stronger in the context of the dominant currency paradigm.
Public tenders typically involve uncertainty and unforeseen costs, which might not be fully known to the bidders themselves. This uncertainty may also concern adaptation costs to the procurer after the delivery, and information about such adaptation costs may influence the bidding strategies, even though the cost is paid by the procurer. Consequently, the availability of information matters, and if it is obtained by the procurer, withholding information from bidders may be better than sharing all of it. To investigate the methods of gathering information, we consider pre-qualification stage questionnaires designed to obtain information from the bidders. If suitably designed, they can be used to disclose information of which the bidders are themselves unaware and can thereby serve as a tool for revealing the inherent tendency of the cost for the bidders, so that the unforeseen future cost can be estimated.
In many environmental problems, emissions are hard to monitor and abatement costs are privately known. We study how efficient abatement nevertheless can be induced through market-like arrangements that condition transfers on verifiable abatement rather than on measured emissions. We characterize the transfer schemes that implement efficient abatement in unique dominant strategies and show that a single scheme, the beneficiaries-compensate transfer scheme, emerges under two distinct characterizations: one based on equilibrium properties and one grounded in classical fairness axioms. Under the beneficiaries-compensate transfer scheme, each agent compensates others according to the benefits received from their abatement. We illustrate its usefulness in two applications.
Public pension and public health care policies - collectively referred to here as social security policies - are costly and increasingly challenged by demographic shifts. Using a two-period overlapping generations model in which health investment affects both mortality and morbidity, we analytically study the optimal design of social security. Our model incorporates three market imperfections: an intergenerational externality, a constraint on saving for future consumption, and incomplete insurance markets. It also considers three policy instruments: subsidies for health investment, subsidies for health expenditures, and a pay-as-you-go pension replacement rate. We analyze how the imperfections, along with evolving trends in mortality and morbidity, shape the optimal set-up of social security policies. Our findings show that there is no universal design for optimal social security: the effectiveness of each instrument depends on the others and on market conditions, with stronger prevention policies emerging as key in aging societies and in systems with less generous pensions.
In 2018, two compensating-income variation studies, based on Household, Income and Labour Dynamics in Australia, reported sharply different monetary health values: A$42,000-A$67,000 for a full quality-adjusted life year (QALY) gain versus A$162,000 for a 0.10 QALY loss (implying A$1.6 million per QALY under linear scaling). We embark on econometric detective work to identify the sources of these discrepancies, carefully aligning samples, cross-examining methodologies, and uncovering methodological differences that drive the divergent results. Gain-loss framing is the largest margin (consistent with willingness-to-accept/willingness-to-pay evidence), while estimator choice (instrumental variables versus reduced form) and income equivalization further magnify differences, jointly resulting in the order-of-magnitude gaps with implications beyond health.
Does financial intermediation affect structural change? We address this question both theoretically and empirically, focusing on whether financial development reinforces structural change during the post-industrialization phase, where employment, value-added, and expenditure shares change towards services and away from manufacturing. We build a dynamic general equilibrium model in which structural change may be driven by mutually independent engines - sectoral productivity gaps, asymmetric factor elasticities - as well as by learning-by-doing. In all its variants, the model robustly predicts that exogenous reductions in intermediation costs (e.g., deregulation shocks) reinforce structural change. We take this prediction to the data by examining the effects of bank branching deregulation in the United States in the period from the 1960s to the 1990s. Within a staggered difference-in-differences framework, we show that bank branching deregulation reinforces the pattern of structural change already underway, leading to an increase in the services share of output and employment in deregulated states.
We analyse the general equilibrium implications of introducing a universal basic income (UBI) for employment and income in an economy characterized by involuntary unemployment. Our framework allows tax revenues and workers' reservation wages to adjust endogenously with employment to satisfy the government's budget constraint. We show that multiple equilibria may emerge and that a UBI can support higher employment when the initial equilibrium is characterized by high taxes. However, in this setting, replacing unemployment benefit schemes with UBI proves less effective at fostering employment, ultimately requiring higher taxes and yielding more limited welfare gains.
We study how poor hand-to-mouth and wealthy hand-to-mouth households in the United States form their expectations as compared to unconstrained households. To do so, we use monthly household data for the period 2005:2 to 2013:6 with information on the exact survey day for each household within a month. Utilizing a timeline of financial crisis events along with changes in stock market values and uncertainty in the days around those events, we assess the response of these households' expectations regarding inflation, unemployment, and the interest rate. Our estimates imply differences in the formation of expectations for liquidity-constrained households relative to unconstrained households. Wealthy hand-to-mouth households tend to revise their inflation expectations downwards substantially so that they make lower forecast errors following adverse financial crisis events that lower the actual future inflation rate, while other households appear inattentive to these shocks. This suggests that the wealthy hand-to-mouth households decipher these financial events' noisy signal regarding lower future inflation more accurately than other households, in line with having a greater incentive to do so.
This paper develops a theory of when annuitization improves or reduces social welfare. The analysis is based on a small open economy with exogenous prices, populated by overlapping generations of non-altruistic agents. Annuities provide longevity risk insurance and above-market returns, but also reduce accidental bequests that transfer resources from the old to the young. I show that the welfare trade-off between these channels is governed by the interest rate: above a threshold, the cost of reducing bequests dominates and no annuitization is socially optimal; below it, partial annuitization improves welfare. Socially optimal allocations can be implemented by mandating a savings portfolio with a predetermined share invested in annuities. Governments that instead mandate annuities through fully funded pensions cannot replicate these allocations, even when annuity markets are missing and agents cannot borrow against future benefits. Although derived for small open economies, the results provide a benchmark for identifying when annuitization raises or lowers welfare and presumably extend to closed economies, where the boundary may still depend on the rate of return but now through threshold values of the parameters that determine it.
We examine the health effects of a labor market activation policy in Sweden, the Youth Job Guarantee. To estimate causal effects, we implement a regression discontinuity design using the policy's age-eligibility threshold, together with detailed administrative data on health-related outcomes, including indicators of mental health. Health effects could arise indirectly, via effects on employment, or directly, for example via improved daily routines. In contrast to most existing studies on the health effects of active labor market programs, our results indicate that the activation policy did not have clear positive effects on health one year after the start of unemployment, measured by prescribed medication or medical contacts.
An important transmission channel of monetary policy is the expectations channel. However, although this channel is important in theory, it has received limited attention in the empirical literature. One reason for this lack of focus is the scarcity of data on firms' expectations. This paper uses a novel data set on firms' expectations and estimates how they respond to monetary policy announcements. I find that firms respond to unexpected changes in the current policy rate by revising downward expectations about price pressure, output and employment growth, and labor market tightness. However, following an unexpected tightening in the future path of monetary policy, firms revise upward expectations about output growth. While the first result is in line with the predictions of the New Keynesian model, the second provides support for the information effect of monetary policy.
We study how acquisition-related foreign direct investment during economic crises affects R&D investments and the direction of innovation of target firms, compared with acquisitions made during periods of strong economic growth. Using a panel of Spanish firms, we find that foreign multinationals cherry-pick the best domestic firms, irrespective of the timing of acquisition. Using matching and difference-in-differences regressions, we find that firms acquired during economic crises experience smaller declines in R&D than those acquired during boom periods. Our results are consistent with the opportunity cost theory of R&D over the business cycle: during recessions, the relative cost of R&D falls, incentivizing innovation, particularly in new product development. However, only firms with sufficient access to finance can capitalize on these conditions. Our results suggest that acquisitions can ease financial constraints at a critical time, enabling target firms to sustain or redirect innovation efforts when incentives are most favorable.
How do regional comparative advantages influence the long-term effects of government interventions? This paper explores this question using disaggregated data on factor endowments and industrial construction in China under the centrally planned economic system during 1953-1978. These investment decisions, primarily driven by political considerations, led to significant variations in the congruence between the factor intensity of policy-targeted industries and local endowment structure. Our findings show that while government interventions can have positive and lasting impacts on the size and efficiency of local industries, misalignments between the production technology of targeted industries and the factor endowment structure significantly weaken these effects.
If old firms are, on average, much larger than young firms, does this mean that firms get better with age? Using Danish administrative data on firms aged 0-65, we study the relationship between size and age. In the cross-section, average size is increasing with age. However, by exploiting the panel structure through the estimation of fixed effects or by using a partial identification approach to the so-called age-period-cohort problem, we find evidence that firm size increases with age only for the first 10-15 years, and falls after that. Moreover, sample composition effects seem to be important to understand the patterns in the cross-section: we find significant differences in exit rates by firm size and strong cohort effects for firms entering in the late 1950s. We also find that the exit rate is not monotonically decreasing with age; for smaller firms, it starts increasing again in their 20s, spiking in their late 30s.
Is long-term economic stress from occupational decline linked to poor health or death? Using Swedish administrative data matched with US occupational trends, I examine this in reduced form and using instrumental variables. Workers who in 1985 worked in occupations that subsequently declined unexpectedly were more likely to die early than similar workers in non-declining occupations, with effect sizes of 6-19 percent of mean mortality. Cardiovascular deaths rose among men, while women faced higher mortality from alcohol, drugs, and suicide. Hospitalization days rose, as did prescription drug use for mental health problems. Effects were strongest for the lowest-paid.
