VoxEU @ CEPR

@voxeu.org

VoxEU – CEPR’s policy portal - promotes "research-based policy analysis and commentary by leading economists". VoxEU columns cover all fields of economics broadly defined and are widely read.

In New Zealand, the Māori Advisory Committee advises on whether registration of a trademark incorporating Māori signs is likely to be offensive. This raises the cost of entry in New Zealand but increases the value of existing trademarks there. cepr.org/voxeu/column...

The figure plots event-study coefficients from a linear probability model where the outcome equals one if the trademark is successfully registered and zero if abandoned, cancelled, rejected, or withdrawn.

Across food, tourism, cosmetics, and luxury goods, indigenous language and heritage increasingly anchor premium market positions, but when it comes to trademarks, few jurisdictions have developed national rules to explicitly protect and preserve traditional knowledge. This column looks at the case of New Zealand, where the Māori Advisory Committee advises the Commissioner of Trademarks on whether registration of a trademark incorporating Māori signs is likely to be offensive to Māori. This embedded collective governance mechanism raises the cost of new entry in New Zealand but increases the value of existing trademarks held there.

Olivier Accominotti, Jason Cen, David Chambers, and Victor Degorce argue that covered interest parity deviations are not a recent phenomenon and have, in fact, been the norm over the past 60 years, with the notable exception of the years immediately preceding the GFC. cepr.org/voxeu/column...

This figure plots the 30-day moving average of the three-month US dollar basis for the British pound between 1968 and 2025, measured in basis points. The USD basis is defined as the difference between the three-month US dollar offshore Eurocurrency deposit rate and the synthetic USD interest rate obtained when swapping British pounds into US dollars. For visual clarity, the vertical axis is truncated at ±60 basis points.

The case of Bhutan shows that cultural norms matter for macroeconomic development. Four decades of data show that rapid income growth remains weakly associated with measured happiness, and that only by embedding the wellbeing measure in local cultural norms does it become effective.

Interest in the role of cultural norms, institutions, and development has resurfaced in recent years especially post the 2024 Nobel Prize in Economics. This column revisits Bhutan, which has pursued ‘gross national happiness’ rather than GDP since the 1970s. Four decades of data show that rapid income growth remains only weakly associated with measured happiness, even as multidimensional wellbeing has continued to rise alongside growth. The broader lesson is that a wellbeing measure becomes effective only if it is embedded in local cultural norms and given institutional teeth, showcasing the importance of cultural norms to economic development.

S Maria Bonin, S Darougheh, & A Kuchler find evidence that the adoption of AI is impacting the Danish labour market, even though the effect is not visible in aggregate statistics. They show that early adopters of AI fall behind their own pre-adoption employment trend. cepr.org/voxeu/column...

Log employment (full-time equivalents) of firms first using AI in 2023 (left panel) or 2024 (right panel), relative to firms reporting no AI use, each measured against a firm-specific linear trend; within three-digit industry × month. Left panel: trend estimated on 2020–2021, so 2022 is out of sample; indexed to January 2023. Right panel: first-time 2024 adopters compared with firms reporting no AI use through 2024; trend estimated on 2021–2022; indexed to January 2024. Shaded areas are 95% confidence bands. The vertical axis is in log points (−0.1 ≈ −10%). Survey-weighted.

Policymakers are asking whether generative AI is reshaping labour demand, and the literature has found mixed results so far. This column links Denmark’s official IT-use survey to monthly employer–employee records and compares firms that started using AI in 2023 with those that did not. Early adopters fall behind their own pre-adoption employment trend. The decline is most visible in AI-exposed jobs and among new hires, and only significant among small firms. The results indicate that AI is affecting the labour market, even though the effect is not visible in the aggregate statistics.

Worker-managed firms are a remedy for rising income inequality & protect social cohesion by increasing employees' bargaining power & providing an enhanced sense of self-worth. Data suggest well-designed co-ops compete with conventional firms in free markets with no evidence of a productivity gap.

Rising income inequality in Western high-income countries threatens democratic governance and social cohesion, and AI risks making it worse. Drawing on a forthcoming book, this column argues for worker-managed firms as a durable remedy. With no external owners taking profits, worker cooperatives protect against the rising bargaining power of employers and provide an enhanced sense of self-worth and self-determination for employees. Furthermore, the data suggest that well-designed worker cooperatives compete with conventional firms in free markets with no evidence of a productivity gap. Because they are underprovided by the market, it calls for government action to support worker cooperatives.

S Reyes Ortega, K Freeman, P Hernando-Kaminsky, J Huber, & @paolomauroecon.bsky.social use data from 150 economies to study how venture capital financing differs across countries. Most of the cross-country gap is associated with how many startups enter the pipeline rather than their funding.

This figure compares the average sectoral distribution of VC volume within countries across high-income and low- and middle-income countries over 2012–2023. For each income group, the figure reports the average share of total VC volume accounted for by each selected sector, weighted by country GDP. The first three sectors—IT, Services–R&D, and Manufacturing–Chemicals and Pharmaceuticals—are sectors in which the average VC share is higher in high-income countries, whereas the remaining three sectors—Wholesale and Retail, Financial Services, and Transportation and Logistics—are sectors in which the average VC share is higher in low- and middle-income countries.

Existing views on what venture capital finances and how its markets are structured are primarily shaped by the experience of the US and a few other high-income economies. This column presents evidence covering nearly 270,000 venture capital-backed startups in over 150 economies to provide a more complete picture. Venture capital markets differ across countries in size and also in kind. At lower levels of economic development, the composition tilts away from knowledge intangibles towards organisational intangibles. Yet differences in what venture capital finances tell only part of the story. Most of the cross-country gap in venture capital activity is associated with how many startups enter the funding pipeline rather than how much each raises.

During banking crises, nonbank financial institutions contract syndicated lending relative to banks but expand corporate bond underwriting. For firms with access to bond markets, relationships with nonbank underwriters can help preserve market-based financing. cepr.org/voxeu/column...

Left panel: corporate bond issuance as a share of corporate bonds and syndicated loans one year before and in the first year of country-specific banking crises, the GFC, and the COVID-19 shock. Right panel: volume-weighted NBFI share of corporate bond underwriting in borrower-country-years without and with a systemic banking crisis.

Nonbank financial institutions are often seen as procyclical lenders, cutting credit more sharply than banks during stress, while bond markets can substitute for bank lending when credit supply contracts. This column brings these perspectives together and argues that this cyclicality depends on the function intermediaries perform. During banking crises, nonbank financial institutions contract syndicated lending relative to banks but expand corporate bond underwriting. For firms with potential access to bond markets, we find that relationships with nonbank underwriters can therefore help preserve market-based financing when bank credit is impaired.

Ruiqi Sun & Daniel Trefler show that AI makes digital services dramatically more attractive, increasing their user base at home and abroad. But when governments restrict the movement of data across borders, the gains from AI are halved. cepr.org/voxeu/column...

The curves show the cumulative distribution of app-level foreign user shares in 2020 through the 60th percentile. Small apps are those at or below the median number of total users.

AI is transforming digital services, and mobile apps, which routinely reach users far beyond their home markets, show just how global that transformation has become. This column shows that AI makes these products dramatically more attractive, increasing their user base at home and abroad by more than threefold. But when governments restrict the movement of data across borders, the gains from AI are halved. This creates a genuine policy tension. Rules designed to protect privacy, national security, and democratic institutions can also make it harder to build better products and spread the gains from new technologies.

When work tasks are classified by what they require of a human, Benjamin Verschuere & Angus Cameron show that AI both substitutes and complements work. Highly substitutable work has grown 3.3 pp/year more slowly since 2021 and complemented work 3.3pp faster. cepr.org/voxeu/column...

The figure shows every US occupation mapped by its substitution and complement shares. Applied to the full US occupational database, employment-weighted work comes out 32% substitution, 15% complement, and 53% inert. 

Estimates of AI's employment effect disagree in size and even in sign. This column argues the disagreement is built into the measurement: AI substitutes for humans on some tasks and complements humans on others, and any single exposure index nets these opposing forces into one number. Classifying US work tasks by what they require of a human and letting the two forces enter separately, both effects appear at once: since 2021, highly substitutable work has grown 3.3 percentage points a year more slowly, and complemented work 3.3 points faster. The adjustment runs primarily through hiring rather than firing, and it lands on the young.

US-Israeli military intervention in Iran sharply reoriented mobilisation away from anti-government protest and toward pro-government activity. Overt foreign intervention can make sovereignty politically valuable to incumbent governments. cepr.org/voxeu/column...

The figure plots the average number of weekly protests per county in Iran. 

Military intervention may be undertaken to weaken a hostile regime and strengthen the position of its opponents. This column argues that evidence from the 2026 US-Israeli strikes on Iran points to a different outcome: the strikes left total protest levels broadly unchanged but sharply reoriented mobilisation away from anti-government protest and toward pro-government activity. The findings suggest that overt foreign intervention can make sovereignty politically valuable to incumbent governments by making domestic opposition easier to portray as aligned with an external adversary.

M Crosignani, L Han, & M Macchiavelli show that domestic US stocks expose investors to substantial geoeconomic risk through firms' global supply-chain relations. Stocks of US suppliers decline sharply after the announcement that export controls target one or more of their Chinese customers.

The chart shows cumulative abnormal returns for stocks issued by US firms that supply directly targeted Chinese customers around the announcement that those customers are added to a US export-control list. The chart is based on the Fama-French five-factor model.

Geoeconomic risk has become a first-order concern for investors. This column shows that domestic US stocks expose investors to substantial geoeconomic risk through firms’ global supply-chain relations. Stocks of US suppliers decline sharply after the announcement that export controls target one or more of their Chinese customers. Mutual funds holding stocks of affected US suppliers experience higher volatility and lower returns as well. Active managers reduce holdings of affected US suppliers and other China-linked firms. Overall, understanding global linkages of portfolio firms becomes increasingly important for understanding portfolio risk.

When undergraduates in Italy received either causal-reasoning training, ChatGPT, both, or neither, students with ChatGPT access scored better on the assignment but only those who received causal-reasoning training produced unique ideas. Chiara Fumagalli, Alfonso Gambardella cepr.org/voxeu/column...

Whether teaching a cognitive skill remains valuable in this era of AI is a crucial question in education and for firms. This column reports on an experiment in which around 1,000 undergraduates in Italy received either (1) causal-reasoning training, (2) access to ChatGPT, (3) both, or (4) neither. While students with ChatGPT scored better in an assignment to write a 180-word recommendation on a real-world problem, only those who had received causal-reasoning training produced unique ideas. The findings highlight the benefits of combining training in first-principles thinking with large language models.

The authors also summarised the report for VoxEU where they briefly touch on the main points, as well as outline two different paths to the next decade dependent on the economic policy actions of the US, Europe, and China. cepr.org/voxeu/column...

Changing structure and challenges of the international financial system: Imbalances, currencies, and financial stability

Concerns are growing about the state of the global economic and financial architecture. The 29th Geneva Report on the World Economy focuses on four key pillars of the global economic order – global im...

cepr.org

Centre for Economic Policy Research@cepr.org · last wk.

New CEPR #ebook out now! Geneva Report 29: Changing Structure and Challenges of the International Financial System By: Valentina Bruno, Steven B. Kamin, Cédric Tille, and Ángel Ubide Free download: cepr.org/publications...

Daniele Colombo and Francesco Toni show that a deterioration in expected future gas availability has a larger and more persistent effect on consumer prices than an interruption to current flows, even without an immediate physical shortage. cepr.org/voxeu/column...

Figure shows euro area supply instruments, allowing the authors to trace the effects of the different shocks separately over time and to estimate the short-run structural elasticities of gas demand and supply. Flow-supply surprises (blue) are gas-price revisions around realised flow disruptions; supply-security-news surprises (orange) are longer-dated futures-price revisions around news about future availability. Circled observations identify illustrative events. Values are percentage changes.

Europe’s output losses proved limited following the 2022 cutoff of Russian gas flows, yet inflation rose sharply. This column uses weather variation and gas-market events to distinguish gas-demand shocks from interruptions to current flows and news about future availability. A deterioration in expected future gas availability has a larger and more persistent effect on consumer prices than an interruption to current flows, even without an immediate physical shortage. Furthermore, lower gas use can coincide with weaker spending and production. Energy-security policy must make future availability credible by ensuring that alternative supplies and stored gas can be accessed under stress.

Though stock returns and consumption growth might appear disconnected, Carlo A. Favero, Alessandro Melone, Sean Myers, & Andrea Tamoni show that consumption helps identify the slow-moving component around which prices fluctuate. cepr.org/voxeu/column...

Debate over whether buoyant equity markets reflect durable economic gains or a widening gap between Wall Street and Main Street has intensified with the AI boom. This column argues that aggregate consumption provides a useful real-economy anchor for stock prices. Temporary departures from that anchor predict stock market returns from one quarter to two years ahead, including out of sample, but do not predict consumption growth. The evidence offers policymakers a real-time indicator of macro-financial dislocation, while stopping well short of a mechanical bubble test.

Geopolitical tensions are reshaping bank lending. Banks are shown to systematically curtail lending to firms exposed to geopolitical risk, with the strongest effects linked to sanctions and lending across geopolitical blocs. By: Dennis Reinhardt, Rhiannon Sowerbutts, Julian Reynolds


Geopolitical tensions are increasingly reshaping the flow of trade, investment and capital flows. This column documents that the same process is also reshaping bank lending. Using confidential supervisory data on thousands of international bank lending relationships, the authors show that banks systematically curtail lending to firms exposed to geopolitical risk. The effects are strongest when risks are linked to sanctions and when lending crosses geopolitical blocs, suggesting that geopolitics is a key determinant not only of how much banks lend but also of where they lend – even to their largest clients.

Randomised access to an AI patent-writing product raised the quality of drafts and compressed performance differentials among junior lawyers. However, on a subsequent task without AI tools, expertise gained from previous AI use appeared only among senior lawyers with seasoned judgement.

Figure: Estimated impact of AI access on patent drafting and redlining quality. Results are reported for all lawyers (black squares), junior lawyers under 7 years of experience (blue triangles), and senior lawyers with 7+ years of experience (orange diamonds). All estimates account for differences across law firms and adjust for slight variations in how many ratings each lawyer received.

Professional judgement comes from learning by doing, but AI has the potential to disrupt this learning process. This column explores how randomised access to an AI patent-writing product affects the performance of lawyers on certain writing tasks. Using the AI tool raised the quality of the drafts and compressed performance differentials, especially among junior lawyers. However, on a subsequent task without AI tools, expertise gains from the previous AI use appeared only among senior lawyers with seasoned judgement.

As the EU's role as a borrower evolved, so did the availability of relevant statistics. Martin Larch & Alice Lucius show that the EU's debt issuance on behalf of its member states increased 5x from 2020-2025 as well as highlight the implications associated with this trend. cepr.org/voxeu/column...

The EU’s far-reaching response to a series of major shocks has turned it into a major borrower on international financial markets. Debt issued by the EU on behalf of its member states now stands at close to 5½% of GDP, almost twice as much as six years ago. This column highlights two crucial issues associated with this trend. It reviews implications for conventional debt statistics and looks at debt servicing obligations.

In emerging markets, capital inflows raise output by strengthening domestic demand and easing financing conditions. Equity-type inflows have larger and more persistent effects, while inflow reversals impose disproportionally large costs.

This figure shows the cumulative responses of real GDP to a one-percentage-point increase in capital inflows relative to annualised trend GDP over an eight-quarter horizon. “Base” denotes our baseline estimates based on capital inflow shocks purged of expectation-driven components.  “Net Capital Flows” uses net rather than gross capital inflows; “Excluding Post-Pandemic Period” excludes observations after 2020Q1; “IV” shows estimates using measures as external instruments; and “Actual inflows (OLS)” reports estimates from a regression using actual capital inflows. Shaded areas indicate 68 and 90 percent confidence intervals for the baseline specification in the left panel and the 90 percent confidence interval in the right panel.

Emerging market economies that are expected to perform well may attract more foreign capital, making it difficult to establish the causal effects of capital inflows on output. Using professional forecasts to account for expectations about future economic conditions, this column constructs an expectations-based measure of capital inflow shocks. The results suggest that capital inflows raise output by strengthening domestic demand and easing financing conditions; equity-type inflows have larger and more persistent effects, while inflow reversals impose disproportionally large costs.

Frontier AI models turn the EU's reliance on foreign-controlled technology into a strategic dependence, widening asymmetries. Eric Törnqvist & Jerzy Kopiński argue that rather than more regulation, the EU must mobilise risk capital, retain talent, and help frontier technologies emerge and scale.

Figure shows the cycle of fragility. Technological debt and limited access to frontier AI models as means of defence would lead to higher risk and less operational resilience. Higher risk means higher funding costs, which in turn reduce margins for investments and reinvestments, slowing and reducing innovation, and widening the technological gap with better-protected jurisdictions. Over time, this would reduce EU financial sector competitiveness, causing business to gradually shift its flow and financial institutions to relocate.

The cyber threat landscape changed dramatically in April 2026 with the emergence of frontier AI models from OpenAI and Anthropic. Access to these models was initially limited, and the US imposed a de facto export ban, since lifted. This column argues that these models turn the EU’s reliance on foreign-controlled technology into a strategic dependence, widening asymmetries between jurisdictions, between attackers and defenders, and between less and more capable financial institutions. The result could be a self-reinforcing cycle of fragility. Rather than more regulation, the EU must mobilise risk capital, retain talent, and help frontier technologies emerge and scale.

Reaching net zero by 2050 would require a globally coordinated carbon tax of $1300 per tonne of CO2, 10x current prices anywhere. Costs are predicted by how exposed an industry is through its supply chain and not by how much it directly emits.

Figure shows the transition pathway of macroeconomic variables under laissez-faire and Paris Agreement scenarios (8.37% global emissions reduction per year). Variables: world emissions, carbon tax, temperature, and GDP.

Reaching net-zero CO2 emissions by 2050, as prescribed by the Paris Agreement, is hard to imagine without a carbon tax. This column estimates that reaching the goal by carbon taxation will require a tax path that peaks at around $1,300 per tonne of CO2, ten times current prices anywhere, and highlights how industries will bear very unequal transition costs. The intuitive smokestack principle, that the biggest direct emitters should pay the most, turns out to be false: costs are instead predicted by a simple statistic capturing how exposed an industry is through its supply chain.

Marijn Bolhuis, Jakree Koosakul, Neil Shenai, & Jie Yang analyse the role of financial repression in the reduction of debt burdens in advanced economies. It has been a persistent feature of modern policymaking, playing an important role post-WWII and rising again since 2008. cepr.org/voxeu/column...

Figure shows a time series of the use of fiscal repression. Measures are normalised relative to the minimum value for each country. Solid lines denote the mean across countries and dashed lines the 25th and 75th percentiles.

With public debt at high peacetime levels, how advanced economies can reduce debt burdens is again at the centre of policy debate. This column examines the role of financial repression, perhaps the least studied channel, by applying new quantity-based measures to more than a century of data for 17 advanced economies. The findings show that repression has been a persistent feature of modern policymaking, playing an important role in post-World War II debt reduction and rising again since the Global Financial Crisis. With several of the conditions historically associated with repression present today, its use may increase going forward.

Data from France show that firms respond to increases in the minimum wage not only by employing fewer workers, but also by increasing training and flattening management hierarchies. Nicholas Lawson, Claire Lelarge, Grigorios Spanos cepr.org/voxeu/column...

Figure: Estimated firm responses to a 3.7% rise in minimum labour costs. Point estimates at average initial exposure to the GMR2 wage-guarantee group. Left panel: percentage changes. Right panel: percentage-point changes (pp). Revenue TFP is revenue-based total factor productivity, not physical productivity. The training estimate concerns production workers in relatively simple firms. Confidence intervals are not shown

How far can minimum wages rise without damaging jobs and output? This column uses evidence from France to show that firms respond not only by employing fewer workers, but also by increasing training and flattening management hierarchies. A quantitative model suggests that this reorganisation cushions the impact of moderate increases. Higher revenue productivity, however, is not a free efficiency gain, and the adjustment becomes less effective as wage floors rise.

The chat logs of AI companies are often used to measure what work AI actually does. A Bick, A Blandin, D Deming, & T Schumacher argue however that chat classifiers cannot see a user’s occupation and therefore classify use into a few generic activities and misattribute AI use across occupations.

Figure: Illustration of the classification chain for a chat asking for help analysing trends in data and showing how a task misclassification becomes an occupation misclassification.

The chat logs of AI companies are often used to measure what work AI actually does. This column uses a nationally representative US survey that links generative AI use to workers’ detailed tasks and compares them with task shares derived from Anthropic, Microsoft, and OpenAI chat data. The four sources disagree sharply, largely because chat classifiers cannot see a user’s occupation and therefore classify use into a few generic activities. Chat logs are informative but on their own can misattribute AI use across occupations and cannot substitute for measurement anchored in who the worker is.