VoxTalks Economics

For three decades years, software makers have paid bug bounty hunters to break their code so they can make it more secure. Recently, some software companies have decided that bug bounties programmes aren't useful any more: Curl shut its bounty in January, and Intel suspended its own in September. And the discovery of the PHP flaw that sent websites offline in October came from researchers working with AI, not from a bounty hunter's report.
Tobias Kretschmer (Imperial College London, CEPR) and his co-authors followed 86 programmes on Bugcrowd to discover how they work best. Do you run closed programmes, open only to invited hackers, or open them to all comers? And what happens now that AI is making a bug report cheaper to produce? Tim Phillips asks Kretschmer how well bug bounties are working.
The research behind this episode:
Zrahia, Aviram, Sarit Markovich, and Tobias Kretschmer. 2026. "Platform Openness: Evidence from Bug Bounty Programs." CEPR Discussion Paper 21970 (gated).
To cite this episode:
Phillips, Tim, and Tobias Kretschmer. 2026. "How well do bug bounties work?" VoxTalks Economics (podcast).

About the guest

Tobias Kretschmer is Professor of Strategy, Technology and Organisation at Imperial College Business School and a Research Fellow at the Centre for Economic Policy Research. He joined Imperial in 2025 from LMU Munich. His research spans competitive strategy in digital industries, platform markets, and the digital transformation of organisations.
Research cited in this episode

Bug bounties. Netscape launched the first technology bug bounty on 10 October 1995, offering rewards for flaws found in the Navigator 2.0 beta. Jarrett Ridlinghafer, a technical support engineer, proposed the idea after he saw enthusiasts finding bugs for free. In this paper's data, the average reward for a unique vulnerability is about $566 in private programmes and $209 in public ones. High-priority flaws average about $1,440.
Bugcrowd connects organisations with independent security researchers. It gave the authors anonymised records of more than 480,000 submissions to 2,483 programmes, from mid-2012 to mid-2021. The data predates the widespread use of AI-generated submissions, so the paper measures the effect of openness, not the effect of AI. The authors argue that AI sharpens the same problem.
Triage. Each submission is classified as rejected, duplicate, or unique. A unique report identifies a flaw nobody has reported before. Only the first valid report generally earns a reward, so a duplicate can be good work that arrived late. It must still be reviewed.
Wisdom of the crowd. Francis Galton's "Vox Populi" (Nature, 1907) reported that the median of about 800 guesses at an ox's dressed weight, made at a Plymouth livestock show, was remarkably close to the true figure. Eric Raymond gave software its version in "The Cathedral and the Bazaar": "given enough eyeballs, all bugs are shallow". Maillart and co-authors tested that idea on bug bounties in 2017. Kretschmer's point is that real attackers differ in their objectives and methods; a varied crowd resembles them more closely than one employee can.
Open and closed programmes. Anyone can join a public programme; a private one is by invitation. Bugcrowd and many other bounty platforms leave that choice to the software company rather than setting one rule for all. Many other platforms keep far tighter control, which makes this the governance decision worth studying. Kretschmer and co-authors set out the wider argument about platforms as organisations in 2022. Among programmes that never switched, public ones were 7.6% of programmes but drew about half of all submissions. Private ones accounted for about 59% of unique discoveries and nearly 80% of dollar rewards.
Comparing like with like. Firms choose their regime deliberately, so comparing open with closed programmes misleads. The authors focus on the 86 programmes that switched once from private to public, and compare them with programmes that stayed private. The main estimates are about 23 more submissions a week, nearly six more unique vulnerabilities, 10 more new researchers, and seven more duplicates. An alternative estimator built for staggered timing gives smaller figures: about 15, three, seven, and six. The effect is front-loaded. Weekly submissions averaged about five before opening, reached about 70, then settled near 20 within 10 weeks. Opening is a choice, not an experiment, and the authors say so.
Quality mix. This is the share of incoming submissions that are unique. It is about 16 percentage points higher in private programmes, and it falls by about 17 points when a programme opens. Opening produces more discoveries in absolute terms and a smaller share of useful reports.
Monitoring. The authors cannot see how much firms spend on review, so they count verdicts per submission. One more weekly verdict is associated with about $139 more in weekly rewards in public programmes and about $508 in private ones; this is an association, not a causal estimate. A rough calculation in the paper suggests that opening pays only when a programme's discoveries have fallen to fewer than about 3.5 a week, before review costs. Programmes that opened averaged about 2.2. The authors say they do not identify the best moment to open, and Kretschmer makes the same point in the episode.
Programmes that have closed. Curl ended its bounty on 31 January 2026 after paying more than $100,000 for 87 confirmed vulnerabilities since 2019. Its creator, Daniel Stenberg, reported that the confirmed rate fell from more than 15% of submissions to below 5%, which he blamed in large part on AI slop. Intel suspended its programme in September 2026, which paid up to $100,000 per flaw, and replaced it with one that pays nothing. It gave no reason. Google stopped accepting product vulnerability reports to its open source reward programme on 1 October 2026, a change it calls temporary.
PHP2Shell. Two critical flaws, CVE-2026-109618 and CVE-2026-109619, were patched in PHP on 21 October 2026. PHP runs about 69% of websites whose server-side language is known, and many operators took sites offline as a precaution. Calif found the first route to an attack during an audit under OpenAI's Patch the Planet programme, and Wiz and OpenAI reported further ones; all used AI tools. The PHP Foundation says a core developer first saw signs of one flaw in 2024, then the fix stalled, and it calls that delay "a structural failure" of process and resources. It reports 15 CVEs and 106 reports closed as invalid, duplicate, or non-reproducible since 1 July, and says "the models don't make the calls".
Open source software. Both curl and PHP are open source, maintained by small teams on which much of the web depends. Kretschmer argues that no single model fits. A firm may keep tight control of its core code and open the rest to its users. A related CEPR paper by Miklos Koren and co-authors, "Vibe Coding Kills Open Source", models how AI could weaken the user engagement that sustains these projects.
More VoxTalks Economics episodes

How well does patent screening work? Mark Schankerman on another system that depends on experts sorting good claims from bad ones.
Will AI transform economic growth? Anton Korinek on what happens if AI starts to automate AI research.
Related reading on VoxEU

Platform regulation: What policymakers can and cannot learn from utility industries, a VoxEU column by Sven Werner and Tobias Kretschmer arguing that rules for digital platforms must account for the independent participants those platforms oversee.
Cybersecurity vulnerabilities and their financial impact, a VoxEU column showing that firms with high cyber exposure earn about 0.42% lower excess returns per month.
A brief look of the economics of open source software, a VoxEU column from 2011 by Neil Gandal on why volunteers contribute and how firms came to mix open and closed code.

Go to work, follow orders all day, and your living standards will rise. That is the deal behind the capitalist economy. In his forthcoming book, Cooperatives in the twenty-first century, Francesco Caselli (LSE, CEPR) argues that the deal has been broken for two generations. Pay at the top keeps climbing. Pay in the middle and at the bottom has barely moved.
The usual fixes for spiralling inequality are higher taxes on the rich and stronger unions. Francesco's answer is to change who owns the firm. In a worker cooperative there are no outside shareholders, so the profits stay with the people who produce them. In conversation with Tim Phillips, he explains which of the economic objections to cooperatives are red herrings, and which ones are more serious.
The research behind this episode:
Caselli, Francesco. 2026. "Worker-Managed Firms Can Curb Inequality and Protect Free Markets." VoxEU.org, 7 October.
Caselli, Francesco. Forthcoming. Cooperatives in the Twenty-First Century: How Worker-Managed Companies Can Curb Inequality and Support Free Markets. Cambridge, MA: Harvard University Press.
To cite this episode:
Phillips, Tim, and Francesco Caselli. 2026. "Cooperatives in the 21st century." VoxTalks Economics (podcast).

About the guest

Francesco Caselli is Norman Sosnow Professor of Economics at the London School of Economics and Director of the Macroeconomics and Growth Programme at the Centre for Economic Policy Research. He is a Fellow of the British Academy and of the Econometric Society. His research spans growth, technical change and inequality, the sources of economic development, the economic and political effects of natural resource wealth, and how voters select politicians.
Research cited in this episode

The labour share. The labour share is the part of national income paid to workers as wages; the rest goes to the owners of capital as profit. Caselli's point is that profits flow mostly to people at the top of the income distribution, so when the labour share falls, inequality rises even if pay within firms stays the same. That is why he argues ownership, not executive pay, is the bigger part of the story.
Wealth taxes. Tim raises Thomas Piketty, who argues for higher taxes on top incomes and on wealth. Caselli agrees in principle, but says any victory is temporary because concentrated wealth funds the campaign to reverse it. For evidence on how people respond to a wealth tax in practice, listen to Marius Brulhart on Switzerland's cantons, linked below.
Worker bargaining power. The other standard fix, associated with labour economists such as Suresh Naidu of Columbia University, is to strengthen unions and give workers more power to bargain. Caselli puts it in the same category as wealth taxes; desirable, but vulnerable to the same political erosion.
Mondragon. The Mondragon group of cooperatives in the Basque Country of Spain is the best known example of worker ownership at scale. Caselli cites its rules that let members keep a stake in the returns to investment after they leave, which answers the worry that departing workers will vote against investing in a future they will not share.
WBO Italcables. When the owners of a steel wire plant in Caivano, on the outskirts of Naples, gave up on it, the workers bought it and ran it as a cooperative. Il Manifesto reported its reopening in December 2015. Caselli says the firm has since grown, invested in new equipment and exports worldwide, and uses it to question whether closures are always healthy creative destruction.
The residual claimant. The residual claimant keeps whatever is left after the bills are paid. In a conventional firm that is the shareholder; in a cooperative it is the workers. Armen Alchian and Harold Demsetz (1972) used the idea to explain why firms have owners who monitor effort. Caselli turns it round: when a slacking colleague costs you money, you watch your colleagues.
The compensating differential. A compensating differential is the pay people will give up for a job with other rewards, such as meaning or status; the idea goes back to Adam Smith's The Wealth of Nations (1776). Caselli says it is how cooperatives keep their most able members, and accepts there is a limit to it.
The talent drain. Gabriel Burdin (2016) found that in Uruguay's worker-managed firms, where pay is more equal, the highest-ability members are more likely to quit. This is the objection Caselli treats as serious, and his answer may surprise listeners who expect a defence of strict pay equality.
Pro-worker AI. Daron Acemoglu, David Autor and Simon Johnson argue that AI can be designed to complement workers rather than replace them, in CEPR Policy Insight 123, "Can We Have Pro-Worker AI?" (2023). Caselli argues that AI developers build what their customers want, and that cooperative customers would want something different.
More VoxTalks Economics episodes

Wealth taxes. Marius BrÃulhart uses Switzerland's cantonal wealth taxes to show what happens when the rich are taxed on what they own, the fix Caselli fears will not last.
Inequality and creative destruction. Richard Blundell on how to get innovation-driven growth without letting the churn widen inequality.
Related reading on VoxEU

Worker-managed firms can curb inequality and protect free markets. Francesco Caselli's column sets out the argument of the book and answers the classic objections to cooperatives.
Labour in the boardroom: The effects of codetermination on firm performance and wages. Simon Jager, Benjamin Schoefer and Joerg Heining find that giving workers seats on German company boards does little for wages, which suggests a minority voice is not the same as ownership.
How AI can become pro-worker. Daron Acemoglu, David Autor and Simon Johnson propose policies to steer AI towards complementing workers rather than automating them away.

The 29th Geneva Report on the World Economy examines three pressures on the international financial system at once: global imbalances, the status of the dollar, and the stability of bond markets. Two of its authors, Valentina Bruno (American University, CEPR) and Cédric Tille (Geneva Graduate Institute, CEPR), tell Tim Phillips why stocks, the dollar and US Treasuries all fell on Liberation Day in April 2025, why that could happen again, and what we should do about it. 
So is this a rupture, or a transition in the world's financial system?
The research behind this episode:
Bruno, Valentina, Steven B. Kamin, Angel Ubide, and Cedric Tille. 2026. "Changing Structure and Challenges of the International Financial System: Imbalances, Currencies, and Financial Stability." Geneva Reports on the World Economy 29. Geneva: International Center for Monetary and Banking Studies; Paris and London: CEPR Press. Free to download.
To cite this episode:
Phillips, Tim, Valentina Bruno, and Ceric Tille. 2026. "The challenges to the international financial system”. VoxTalks Economics (podcast).

About the guests

Valentina Bruno is Professor of Finance at the Kogod School of Business, American University, where she holds the Arlene R. and Robert P. Kogod Eminent Scholar Chair in Finance. She is a Research Fellow of the Centre for Economic Policy Research in the Banking and Corporate Finance programme. Her research spans macroeconomics and finance, with recent work on how the US dollar transmits global financial conditions. She has worked at the World Bank and been a visiting scholar at the Federal Reserve Board.
Cédric Tille is Professor of International Economics at the Geneva Graduate Institute, where he heads the Bilateral Assistance and Capacity Building for Central Banks programme. He is a Research Fellow of CEPR in the International Macroeconomics and Finance programme. His research spans financial globalisation, the dynamics of countries' external assets and liabilities, and the currencies used to invoice international trade. He sat on the Bank Council of the Swiss National Bank from 2011 to 2023 and previously worked for nine years at the Federal Reserve Bank of New York.
Research cited in this episode

The triple decline of April 2025. After the US announced sweeping tariffs on 2 April 2025, which President Trump called "Liberation Day", US stocks, Treasuries and the dollar fell together. Bruno argues that foreign investors did not sell their US assets; they bought protection against a falling dollar, and that hedging pushed the currency down. Much of their US equity exposure remains unhedged, so a new shock that starts in the US could trigger the same response.
Global imbalances. Large and persistent current account surpluses in some countries, mirrored by deficits in others. Imbalances were a major policy topic before 2008, faded after the global financial crisis, and are now back. Tille argues that the headline measures have risen only modestly; the concern is China, whose trade surplus is the largest ever recorded as a share of world GDP.
The China shock. Autor, David H., David Dorn, and Gordon H. Hanson. 2013. "The China Syndrome: Local Labor Market Effects of Import Competition in the United States." American Economic Review 103 (6): 2121-2168. Cheap imports from China benefited US consumers as a whole, but the job losses were concentrated in particular industries and places, furniture making among them. Tille uses it to explain why Germany's car workers see the current wave of Chinese competition differently from German consumers.
The coordinated policy package. International Monetary Fund. 2026. "2026 External Sector Report: Amid Rising Imbalances, the Case for Rebalancing." The standard prescription is US fiscal consolidation, more consumption in China and more investment in Europe. The report cites IMF estimates that a coordinated package would narrow global imbalances by 0.5% of world GDP over the medium term, compared with a widening of 0.4% under current trends. Nearly all the narrowing comes from China.
The renminbi debate. Brad Setser of the Council on Foreign Relations argues that China holds its exchange rate down and that a stronger renminbi is part of the solution. Tille notes that other economists treat the exchange rate as a sideshow to China's high savings and weak domestic demand.
The basis trade. Pension funds, insurers and asset managers often prefer to hold Treasury exposure through futures. Hedge funds take the other side; they sell the futures, buy the bonds and earn the small gap between the two prices, funded with overnight borrowing. In calm markets the trade links the futures and cash markets and helps the Treasury sell its debt. When lenders pull back, hedge funds must sell bonds at the same moment. The report cites estimates that Cayman Islands hedge funds alone held about $1.8 trillion of Treasuries.
March 2020 and the UK gilt crisis. In March 2020, the Covid shock set off a scramble for dollars in which even Treasuries were sold, and the Federal Reserve bought Treasuries to restore order. In September 2022, UK pension funds using liability-driven investment strategies had to sell gilts as yields rose, which pushed yields higher still, until the Bank of England stepped in. Bruno cites both as examples of safe assets at the centre of stress, where falling prices force more sales.
The 1970s and 2008. Bruno compares today with two earlier periods when imbalances, the dollar and bond markets came under pressure together: the collapse of the Bretton Woods system of fixed exchange rates in the early 1970s, when inflation also eroded bond values, and the years before 2008, when Asian savings and European bank lending flowed into US assets. In 2008 the stress sat in the banks. Today it sits with pension funds, insurers and hedge funds.
Exorbitant privilege. The phrase, usually attributed to Valéry Giscard d'Estaing as French finance minister in the 1960s, describes the advantage the US gets from issuing the world's main reserve asset; foreigners accept a low return to hold it. Tille points out that the US once ran a trade deficit while earning more on its foreign assets than it paid on its liabilities. That surplus on primary income has now gone.
The market-based trilemma. The old trilemma of international finance, associated with Robert Mundell and Marcus Fleming, says a country cannot have free capital flows, a fixed exchange rate and an independent monetary policy all at once. The report proposes a new one. Governments want to issue ever more debt, regulators want tight limits on bank balance sheets, and investors expect bond markets to work smoothly every day. You can have any two, but not all three all the time.
"Our currency, but your problem." John Connally, US Treasury Secretary, told European counterparts in 1971 that "the dollar is our currency, but it's your problem." The report updates the line; the dollar is no longer just a currency but the operating system of global finance.
Rupture or transition? The report opens with a remark by Mark Carney, Prime Minister of Canada, at the 2026 World Economic Forum, that the world is "in the midst of a rupture, not a transition." The authors test whether the same is true of the international financial system.
More VoxTalks Economics episodes

The Dollar Anchor Is Slipping. Tarek Hassan explains why the dollar became the world's safe haven and how tariffs weaken it.
Addressing Global Imbalances. Gita Gopinath and Philip Lane on why today's imbalances put more risk on government debt and non-bank lenders.
Stablecoins and Global Imbalances. Gilles Moec on how dollar stablecoins could pull global savings into US Treasury bills.
Related reading on VoxEU.org

Changing structure and challenges of the international financial system: Imbalances, currencies, and financial stability, a VoxEU column in which the four authors summarise the report.
Geopolitical tensions and international financial fragmentation: The 28th Geneva Report on the World Economy, a VoxEU column on how geopolitical rivalry is pushing global finance to split along political lines.

Who gets the kids? Courts aim to decide custody in the child's best interests, but nobody has had the data to test how well those decisions work.
Gordon Dahl (UC San Diego, CEPR), Costas Meghir (Yale, CEPR) and seven co-authors built the data themselves. They collected more than 56,000 Swedish custody rulings, scanning paper files by hand where no digital copy existed, and linked them to records of parents' earnings, prescriptions and children's school tests. 
It seems Swedish judges do not agree on how often to award joint custody. Some award it in about 60% of cases; others in about 20%. But that decision makes a huge difference to the lives, and happiness, of both parents and children. They tell Tim Phillips whether joint or sole custody has, on average, better outcomes.
The research behind this episode:
Canessa, Stella, Gordon Dahl, Anna Hasselqvist, Costas Meghir, Susan Niknami, Marten Palme, Helmut Rainer, Olof Rosenqvist, and Pengpeng Xiao. 2026. "Life After Divorce: Effects of Joint Custody on Parents and Children." CEPR Discussion Paper 21780. 
To cite this episode:
Phillips, Tim, Gordon Dahl, and Costas Meghir. 2026. "Life after divorce." VoxTalks Economics (podcast). 

About the guests

Gordon Dahl is Professor of Economics at the University of California, San Diego, and a Fellow in the Labour Economics programme of the Centre for Economic Policy Research. His research spans the family, public policy and the labour market, including how family income shapes children's achievement, peer effects among co-workers and family members, the effects of incarceration on employment, and how welfare use passes between generations.
Costas Meghir is the Douglas A. Warner III Professor of Economics at Yale University and a CEPR Fellow in the Labour Economics, Development Economics and Public Economics programmes. His research spans labour supply and welfare programmes, marriage markets and how resources are shared within households, informal labour markets, and early childhood development.
Research cited in this episode

The best interests of the child standard. Since the 1970s, most custody law has told judges to decide in the child's best interests rather than weigh the parents' claims. Most countries now read this as keeping both parents involved, which is why joint custody is the default in many jurisdictions. The standard leaves wide discretion to judges, and Meghir says judges get little systematic evidence on what works. Elizabeth Scott and Robert Emery examine why the standard has persisted despite this criticism in "Gender Politics and Child Custody: The Puzzling Persistence of the Best-Interests Standard" (Law and Contemporary Problems, 2014).
The tender years doctrine. Before the mid-19th century, English law treated children as their father's property. The Custody of Infants Act 1839, passed after a campaign by Caroline Norton, let mothers petition for custody of children under seven, and the presumption that young children belong with their mother spread across Europe. Dahl raises it to show how far the legal default has moved in less than two centuries, from father, to mother, to both.
Random assignment of judges. Families who end up with joint custody differ from those who do not, so a simple comparison cannot separate the effect of the arrangement from the effect of the family. In Sweden, custody disputes go to judges at random, and judges differ in how often they award joint custody. Comparing families who drew a lenient judge with families who drew a strict one isolates the effect of the ruling itself. Economists have used the same design to study prison sentences, for example Jeffrey Kling in "Incarceration Length, Employment, and Earnings" (American Economic Review, 2006). The method only measures effects for cases where judges could disagree; it says nothing about cases where every judge would reach the same decision.
Swedish court archives. Custody rulings are public records in Sweden, but no national database holds them. The authors requested files from all 48 district courts and received documents from 45. Digital archives covered cases from about 2010; for earlier years, the team visited court archives and scanned 19,491 paper rulings by hand. Dahl estimates that 80% of the project's time went on collecting data.
Swedish administrative registers. Statistics Sweden registers cover the whole population and record earnings, employment, family links and place of residence. The authors linked these to prescription records, which show who takes antidepressant, anti-anxiety and ADHD medication, and to records of intimate partner violence. This is why the research could follow mothers, fathers and children for years after each ruling.
National tests and school quality. All Swedish pupils sit the same national tests in Swedish and mathematics in grades three, six and nine. The authors use these scores to measure children's achievement, and build a measure of school quality from how much better or worse a school's pupils score than their backgrounds predict. In the paper, joint custody raises test scores by 0.47 of a standard deviation, which Meghir likens to about two extra years of schooling.
More VoxTalks Economics episodes

Helping fathers to acknowledge paternity. Anna Raute on what happens to mothers and children when fathers do not formally recognise their child, and which policies change that.
Love, money and parenting. Fabrizio Zilibotti on what economic research tells us about parenting, and about raising children who thrive.
Related reading on VoxEU.org

The effect of joint custody on marriage and divorce, a VoxEU column by Martin Halla using US state reforms to show that joint custody laws raised marriage rates without a clear effect on divorce.
The effect of parental divorce on children's long-term outcomes, a VoxEU column using Austrian data to estimate how divorce affects children's education, employment and health as adults.

Since March 2022, it has been harder to buy an iPhone in Russia, because Apple chose to stop selling there. So did Prada, Renault, Starbucks, Nike and more than 1,000 other firms. These private sanctions added to the official sanctions imposed by 45 countries.
Beata Javorcik (EBRD, Oxford, CEPR) and her co-authors wanted to know if private sanctions did anything more than generate good headlines. To do this, they matched millions of Russian customs records to the trademarks on the goods, so they could tell an iPhone from a Samsung phone and a Gucci scarf from a Gucci bag. Would imports of goods under voluntary sanctions stop? Would the brands cheat? Would importers find an other way to get goods into the country through the back door? She tells Tim Phillips what private sanctions on Russia have achieved.
The research behind this episode:
Chupilkin, Maxim, Beata Javorcik, Alexander Plekhanov, and Aleksandra Peeva. 2026. "Not Wearing Prada: Do Private Sanctions Reinforce Official Sanctions?" CEPR Discussion Paper 21774. 
To cite this episode:
Phillips, Tim, and Beata Javorcik. 2026. "Not wearing Prada." VoxTalks Economics (podcast).

About the guest

Beata Javorcik is Chief Economist of the European Bank for Reconstruction and Development, on leave from the University of Oxford, where she is Professor of Economics and a Fellow of All Souls College. She is a Fellow of the Centre for Economic Policy Research in the International Trade and Regional Economics programme. Her research spans foreign direct investment, shocks to production networks, trade under sanctions and the evasion of import duties.
Research cited in this episode

Official sanctions on Russia. After the full-scale invasion of Ukraine in February 2022, 45 economies, including the US, the EU and the UK, imposed sanctions on Russia. They cut selected Russian banks off from the SWIFT payment messaging system, immobilised the reserves of Russia's central bank, limited Russia's earnings from hydrocarbon exports and banned exports of arms, dual-use goods, advanced technology, industrial equipment and luxury goods. By the end of 2022, around 40% of product lines were at least partly covered. The luxury ban explains some odd entries. Ski suits and sports clothing priced above €300 are sanctioned, and champagne is banned while prosecco is not. Enforcement sits with several bodies: in the US, the Office of Foreign Assets Control and the Commerce Department; in the EU, individual member states; in the UK, the Office of Financial Sanctions Implementation and the Office of Trade Sanctions Implementation.
Private sanctions and public opinion. Oliver Hart, David Thesmar and Luigi Zingales surveyed 2,915 Americans, a representative sample, in May and June 2022. Of these, 61% agreed that doing business in Russia is like being an accomplice to the war, and that firms should sever ties whatever the consequences. Javorcik raises the study to explain why firms with no legal obligation to leave Russia chose to go anyway: reputation, the threat of boycotts, pressure from investors, employees and business partners, and sometimes the values of managers. Hart, Oliver, David Thesmar, and Luigi Zingales. 2024. "Private Sanctions." Economic Policy 39 (117): 203-268.
"The business of business is business." The phrase is a popular summary of the view Milton Friedman set out in a 1970 essay in the New York Times Magazine, "The Social Responsibility of Business Is to Increase Its Profits." Hart and his co-authors found that only a minority of their respondents treated the decision to leave Russia as a purely commercial one.
The Yale list. A team at the Yale Chief Executive Leadership Institute, led by Jeffrey Sonnenfeld, has tracked what more than 1,500 major companies said about their Russian operations since the invasion. To qualify, a firm must have revenue above $100 million, be majority owned from outside Russia and Belarus, and have meaningful operations in its home country, in Russia and in at least one other country. The team grades firms from full withdrawal to "digging in," and keeps a separate record of firms that reneged on their promises to leave. Javorcik and her co-authors use the list to classify the owners of 1,657 Western trademarks. Only 11 firms moved from strict private sanctions to doing little or nothing.
Customs data with trademarks. The paper draws on transaction-level records of Russian imports from 2016 to 2023, about 12 million records in 2022 alone. Each record shows the product code, the value, the quantity, the trading partner and the trademark. The products are classified at the six-digit level of the Harmonized System, which has more than 5,600 categories. The trademark field contained more than 1.2 million spellings; the authors used a large language model to standardise them, which merged 600 versions of "Samsung" into one.
Trade diversion through neutral countries. Countries that do not participate in sanctions, including China, Turkey and the UAE, do not monitor re-exports to Russia. After the invasion, the share of imports arriving via traders in these neutral jurisdictions rose by an additional 16 percentage points for brands whose owners announced private sanctions. The same authors traced a related route in "The Eurasian Roundabout: Trade Flows into Russia through the Caucasus and Central Asia" (Chupilkin, Javorcik, and Plekhanov 2026, European Economic Review).
Tariff equivalent. Trade economists often express a barrier to trade as the ad valorem tariff, a tax set as a percentage of the value of the goods, that would cause the same fall in trade. On this measure, private sanctions against Russia equal a tariff of between 9% and 84%. The range is wide because the answer depends on the trade elasticity, which measures how strongly trade responds to a change in costs.
The Clark Center Forum surveys. On 8 March 2022, the Kent A. Clark Center for Global Markets at Chicago Booth asked its panels of leading economists if the sanctions already in place would lead to a deep recession in Russia. Of the respondents, 93% of the US panel and 94% of the European panel agreed. A deep recession did not follow. In June 2025, a majority of the European panel agreed that, on the evidence of the past 10 years, Western-led economic sanctions do not substantially deter target countries. Javorcik, a panel member, agreed, citing trade diversion and sanctions circumvention.
The UN vote. On 2 March 2022, 141 countries in the UN General Assembly voted for a resolution that condemned Russia's invasion and demanded its withdrawal from Ukraine. Javorcik contrasts this with the 45 economies that impose sanctions. The effectiveness of sanctions depends on how many countries take part.
More VoxTalks Economics episodes

Do sanctions work? Moritz Schularick takes the long view of the evidence on sanctions of every type, and asks if policymakers will impose the ones that bite.
Trading around geopolitics Giancarlo Corsetti, Banu Demir and Beata Javorcik on whether Turkish exporters filled the gap in the Russian market, and an earlier sighting of the balloon.
What impact have trade sanctions had on Russia? Dzhamilya Nigmatulina uses railway shipments, company accounts and procurement records to measure what official trade sanctions did inside Russia.
Related reading on VoxEU

Beyond mandated sanctions: Western firms' voluntary departure from Russia, a VoxEU column by Tetyana Balyuk and Anastassia Fedyk on how shareholder pressure drove Western firms out of Russia.
Export bans that weren't really bans: How Russia kept importing military goods, a VoxEU column by Lisa Scheckenhofer, Feodora Teti, David Torun and Joschka Wanner on the gaps, transit loopholes and third-country routes that kept military components flowing.
Economic consequences of Russia's invasion of Ukraine: Views of leading economists, Romesh Vaitilingam's March 2022 report on the Clark Center survey that predicted a deep Russian recession.