...

American criminologist Richard Rosenfeld calculated the threshold beyond which rising prices begin to turn into rising robberies. This fall, the West crossed it, and the first signal has already come from the stores.

On Friday, August 14, 2026, Donald Trump traveled to a police training center in Garden City, outside New York, to announce a figure the White House had long been waiting for. In 2025, the number of murders in the United States fell by 18.1 percent, and the murder rate dropped to 4.1 per 100,000 residents. The country had not been that low since 1955–1956. Violent crime fell 9.3 percent in a year, property crime 12.4 percent, robberies 18.5 percent and motor vehicle thefts almost 23 percent, to their lowest level in more than two decades. The president took the credit.

At the same time, other federal statistics painted a very different picture. The U.S. and Israeli strikes on Iran at the end of February and the paralysis of shipping in the Strait of Hormuz sent Americans back to 2022: a gallon of gasoline that cost $2.98 in late February topped $4.50 by Memorial Day. In September, annual U.S. inflation reached 3.7 percent, and in October gasoline was 37.4 percent more expensive than a year earlier. The eurozone posted 4 percent inflation in October, the highest since September 2023, with energy prices up 20.8 percent year over year.

Two lines on a chart, record-low crime and once again rising prices, cannot diverge for long. Or so argued a man who did not live to see the current triumph of police statistics. Richard Rosenfeld, a professor at the University of Missouri–St. Louis and former president of the American Society of Criminology, died on January 8, 2024, at the age of 75. Ten years before his death, he published a paper after which the conversation about economics and crime could no longer be conducted the old way. Today it reads like a forecast for 2027.

The Great Recession Puzzle: Unemployment Soared, but Robberies Fell

For half a century, criminologists looked for the link between the wallet and crime in one place: the labor market. The logic seemed ironclad: a person loses his job, then his income, then his footing, and sooner or later he reaches for a lock pick. The evidence refused to cooperate. Studies produced results so contradictory that in 1987 American criminologist Theodore Chiricos described the state of his field as a "consensus of doubt."

By the turn of the century, the doubt began to lift, though thanks to other indicators. As early as 1985, Philip Cook and Gary Zarkin showed that robberies and property crimes rise during recessions and fall during upswings. It later emerged that state per capita GDP and the index of consumer sentiment explained the dynamics of theft better than unemployment did. The new consensus looked solid: a bad economy produces thieves.

Then came the crisis of 2008–2009. Output collapsed, U.S. unemployment jumped to 10 percent, and consumer optimism fell to historic lows across the developed world. Crime, contrary to the models, did not budge, and in some places it kept falling. In Rosenfeld's sample of 14 countries, murder and robbery rates barely changed. Burglaries in eight countries rose 21 percent in 2008, but by 2009, at the bottom of the recession, they fell 9 percent.

Rosenfeld noticed a detail his colleagues had missed. During all previous downturns, especially in the stagflation era of the 1970s and early 1980s, prices had been rising. In 2009 they were falling, and economists were seriously worried about deflation. Ireland saw consumer prices drop 4.5 percent that year, so sharply that Rosenfeld had to exclude that data point from his calculations as a statistical outlier. The answer to the puzzle lay in the price tags, not in the unemployment lines.

Four Price Revolutions and One Post-Soviet One

This insight had a historical foundation that criminology had ignored for almost two decades. In 1996, economic historian David Hackett Fischer published "The Great Wave," in which he described four price revolutions in Western history: in the 14th, 16th, 18th and 20th centuries. Each began in an era of prosperity, was accelerated by population and demand growth, and was fueled by rising inequality. In their late stages, every one of them produced a surge of violence and theft, followed by a long period of price "equilibrium" in which crime declined.

Fischer was categorical about the 20th century. "Crime rose rapidly throughout the world from 1965 to 1993. These movements were very closely correlated with the rate of inflation," he wrote. The Great Depression, which had puzzled criminologists with its falling crime, fit the same pattern: prices declined in the 1930s, and crime declined with them. The famous paradox of the 1960s, "crime amid affluence," also ceased to be a paradox. American prices climbed steeply from the middle of that decade, and the crime rate followed.

The harshest illustration of the pattern, however, came not from the West at all. In 1992, consumer prices in Russia rose roughly 26-fold, according to Rosstat. The murder rate, judging by mortality statistics, doubled in three years: from 15.2 per 100,000 residents in 1991 to 32.3 in 1994. The collapse of the ruble went hand in hand with the disintegration of the state, criminal privatization and the paralysis of the police, so it is impossible to isolate the pure contribution of prices here. My generation, which lived through the early 1990s in the post-Soviet space, understands Rosenfeld's mechanics without any regressions: when your salary went up in smoke within a week, the flea markets filled up with goods whose origins no one asked about.

Where Does a Shopper Go When Even the Discount Store Is Out of Reach?

Rosenfeld's main contribution was the mechanism. Earlier authors, from Fischer to Gary LaFree, explained the effect of inflation on crime through stress, distributional conflict and the erosion of trust in institutions. Rosenfeld did not reject these explanations. He added the dry economics of demand.

His reasoning goes like this. When prices rise faster than incomes, consumers begin to "trade down": first they leave midrange department stores for Walmart and dollar stores such as Dollar General, then for Goodwill and Salvation Army thrift stores. Retail statistics capture this shift reliably: in years of price pressure, discounters and charity thrift stores do noticeably better than midmarket stores, which lose customers in exact proportion to how far real incomes lag behind price tags. The question Rosenfeld posed sounds almost provocative: where does a person go when even Dollar General is out of reach?

The answer: to the market for stolen goods. The main feature of a stolen item is that it is cheap. Without that, such goods would have no buyers, since legitimate retail offers guarantees, returns, peaceful dispute resolution and no risk of arrest. Surveys show that price is what determines a person's willingness to buy an item of dubious origin. Rising demand for cheap stolen goods increases the profitability of theft and burglary, and supply follows demand. British researchers Sutton, Hodgkinson and Levi put it as succinctly as possible: without bargain hunters, the market for stolen goods would shrink enormously.

The scheme has a sequel that links inflation to violence. Deals involving stolen goods take place out of the state's sight: late at night, in alleys, in garages, in messaging app chats. If a buyer cheats a seller, there is nowhere to complain, neither to a customer service desk nor to a court. In such a "stateless" space, agreements are enforced with fists and guns, and the participants themselves, legally defenseless, become easy prey for robbers. Criminology long ago described this logic using drug markets as an example. Rosenfeld extended it to any shadow market that swells along with prices.

This leads to his central thesis: inflation should hit crime harder than unemployment does. Job loss affects a minority. Even in the worst years, the overwhelming majority of the working-age population remains employed, and the full effect of a layoff shows up only months later, when savings and benefits run out. Rising prices affect every shopper the day the price tag changes. They hit the poor hardest, but everyone stands in the checkout line.

The 2.67 Percent Threshold: A Number Central Banks Overlook

Rosenfeld decided to test the hypothesis on data from 13 European countries and the United States for 1981–2010. The sample was far from ideal, and the author admitted as much. European data series had to be stitched together from Interpol figures for the 1980s and two European sources for the following decades. At the seams between sources, the burglary series showed implausible breaks, sometimes by an order of magnitude, and only seven of the 13 European countries remained in that part of the study. The models accounted for unemployment, per capita GDP, urbanization, the share of young people aged 15 to 24, and infant mortality as an indicator of absolute poverty.

The result was inconvenient for a half-century tradition. Unemployment and GDP growth showed no significant association with any of the three types of crime. Inflation did. A 1 percent rise in prices was accompanied by a 0.112 percent increase in the burglary rate and, at the edge of statistical significance, a 0.071 percent increase in murders. When the United States, with its incomparably high murder rate, was removed from the sample, the effect of inflation became significant at the 5 percent level.

The most interesting finding concerned robbery. Here the relationship turned out to be nonlinear, U-shaped, just as Fischer had predicted with his "quiet beginnings" of price revolutions. As long as inflation is moderate, robberies barely respond to it and even decline. Beyond a certain threshold, each additional percentage point of price growth begins to push them up. Rosenfeld calculated this threshold: 2.67 percent a year. Average inflation in his sample first dropped below that mark in 1996 and did not rise above it until 2010. In other words, the entire era of the great decline in Western crime unfolded beneath a ceiling of 2.67 percent.

The size of the effect was not modest either. By Rosenfeld's calculations, without inflationary pressure, murders would have declined roughly a quarter faster and burglaries 15 percent faster. With inflation one standard deviation above the mean, about 7.9 percent a year, price growth explained 77.5 percent of the annual increase in robberies.

Rosenfeld later expanded his testing. In 2016, together with Aaron Levin, he analyzed U.S. data for 1960–2012, including unemployment, GDP, income and consumer sentiment in the models. Only inflation retained a robust short-term and long-term effect on property crime. In 2023, a city-level study he co-authored with Matt Vogel was published. The link between inflation and homicide proved significant and was only partly explained by rising acquisitive crime, and it was noticeably stronger in disadvantaged cities.

Rosenfeld stated the conclusion for monetary policy cautiously but unambiguously. By keeping inflation within a 2 to 3 percent range, the Fed and other central banks are, without intending to, engaged in crime prevention. In 2014, he added that developed countries appeared to have entered a new era of price equilibrium and had secured low and stable crime for a long time to come. Eight years later, that sentence was out of date.

The 2022 Dress Rehearsal

In June 2022, annual U.S. inflation hit 9.1 percent, a four-decade high. The eurozone reached 10.6 percent in October of that year, and Britain 11.1 percent. Rosenfeld, by then a professor emeritus and co-author of the regular crime trend reports of the Council on Criminal Justice, a nonpartisan Washington think tank, issued public warnings. "I think inflation is a public safety issue," he said in April 2022. His forecast was specific: property crime would go up because shoppers would be drawn to the black market.

The early data did not refute him. In the large cities in the Council's sample, larcenies and robberies rose roughly 20 percent in the first half of 2022. Murders, meanwhile, were already declining after their pandemic peak, and this divergence is an important detail we will come back to.

Shoplifting followed a path that seemed lifted straight from Rosenfeld's textbook. In the 21 American cities where the Council has maintained a comparable series since 2018, the average rate of reported shoplifting fell to 402.5 per 100,000 residents in 2021. Following prices, it rose to 446.1 in 2022 and 529.2 in 2023, and the increase continued in the first half of 2024. The contents of the thief's bag changed too: the median value of stolen goods rose from about $75 in 2019 to $100 in 2021. Deodorants locked under plastic covers at Duane Reade drugstores in New York, photographed in August 2023, became the visual symbol of the entire American debate.

Europe repeated the pattern with national variations. Dutch police recorded 40,000 shoplifting incidents in 2022, more than in any year since 2015. In Bavaria, the number rose from 35,528 in 2022 to 42,449 in 2023. Police in England and Wales recorded 530,643 shoplifting offenses in the year ending March 2025. That was a 20 percent increase over the previous year and the highest figure since the current recording system began in 2003. Convictions for shoplifting in the year to June 2025 reached 44,701, more than double the level of three years earlier. Among those convicted, the share of people aged 40 to 59 grew, while the share of young people declined. The profile of the British shoplifter grew older precisely when the cost of living became the dominant theme in national politics.

Then inflation receded, and theft receded with it. In 2025, shoplifting in the American cities in the Council's sample fell 10 percent, the first decline since 2021. England and Wales showed a decline of about 1 percent for calendar year 2025, to 509,566. Germany recorded almost 6 percent fewer simple shoplifting cases than a year earlier, 357,600, and Bavaria 38,394. With a lag of a year to a year and a half, the curves retraced the path of the price tags.

Where the Formula Comes Apart at the Seams

An honest analysis requires acknowledging that the inflation theory of crime does not explain everything, and the inconvenient facts are in plain sight.

The most obvious of them is homicide. In 2022, American inflation averaged 8 percent, yet murders began to decline and have fallen every year since. According to the Council's estimate, by 2025 their number in large cities had dropped 44 percent from the 2021 peak. If prices were the main engine of violence, 2022–2023 would have produced a new surge. It did not happen. Rosenfeld, to his credit, saw the weakness of this particular link himself: in his 2014 paper, the effect of inflation on murder was significant only at the 10 percent level, and most of its dynamics remained unexplained. The pandemic surge of 2020–2021 and the subsequent decline were driven by other forces: lockdowns, the protests after George Floyd's death, record gun purchases, a crisis of trust in the police and its gradual resolution.

The second inconvenient fact is car theft. The surge of 2022–2023 in the United States was largely explained by a viral social media video that taught teenagers how to steal certain Kia and Hyundai models without immobilizers. Inflation had nothing to do with it, and a researcher who mechanically attributed the rise in car thefts to it would be mistaken.

The third problem is the data itself. Britain's record for shoplifting partly reflects retailers' greater willingness to report it to the police. The Crime Survey for England and Wales, based on population surveys, estimated the number of thefts in the year to March 2025 at 2.8 million, nearly unchanged from the previous year. The Association of Convenience Stores counted 6.2 million thefts in its segment alone. The gap between reality and police records is enormous, and yet the trend in recorded crime is not the same as the trend in crime. The Council on Criminal Justice sample, on which almost the entire American debate relies, is not nationally representative, and the authors themselves honestly warn about this.

Of all the sources this article is built on, I trust the FBI and national statistical agencies the most and industry associations the least. There is a very specific reason for that choice.

Lobbyists, Prosecutors, Candidates: Who Profits from the Fear of Theft

In April 2023, the National Retail Federation, the main lobby of American retail, together with the consulting firm K2 Integrity, released a report claiming that "nearly half" of the industry's $94.5 billion in losses in 2021 was due to organized retail crime. Politicians and numerous media outlets picked up the figure. In December 2023, the federation retracted the claim, acknowledging that an analyst's conclusion had been erroneous. The federation's own annual survey showed that all external theft, organized and unorganized, accounted for about 37 percent of losses. The rest came from employee theft, accounting errors and logistics.

The exaggeration had obvious beneficiaries. Retailers used it to explain shrinking margins and store closures to investors, and they demanded more police and tougher penalties from the authorities. The Combating Organized Retail Crime Act, introduced in Congress in 2023, called for creating a dedicated coordination center within the Department of Homeland Security. In November 2024, California voters approved Proposition 36 by more than a two-thirds margin, toughening penalties for repeat theft and partially reversing a 2014 reform. The British government's Crime and Policing Bill eliminates the £200 threshold below which shoplifting was treated as "low-value."

Rosenfeld warned against precisely this kind of response. In his 2014 paper, he explicitly called America's war on drugs a model of how not to deal with the market for stolen goods. By his logic, sharply tougher penalties for theft and for receiving stolen goods drive the market even deeper underground and increase the "utility" of violence in disputes between shadow sellers and buyers. Instead, he proposed making stolen goods technically more costly: tagging electronics with chips, making resale harder, stripping stolen items of liquidity.

America implemented part of this program almost without noticing. Since June 27, 2023, the federal INFORM Consumers Act has required online marketplaces to verify the identity of high-volume third-party sellers. The digital flea market, where stolen goods used to vanish among millions of listings, has become more transparent. How much this contributed to the drop in shoplifting in 2025, no one has yet measured. Rosenfeld's successors face a far more useful task than yet another argument over which party "defeated crime."

Meanwhile, that argument is in full swing. Trump credits his policies, including the deployment of the National Guard in Washington, for the drop in murders. Council on Criminal Justice head Adam Gelb responded with academic coolness on the day the FBI data were released: "Crime rates are falling in cities with very different public safety strategies, economic conditions and political leadership." In his view, the uniformity of the decline across the map should discourage anyone from giving credit to a single leader, party or program. The irony lies elsewhere. The November U.S. midterm elections, by a widely shared assessment, were decided largely by discontent over prices, not by crime. Voters, it seems, sensed a connection that politicians never put into words.

The Strait of Hormuz as a Crime Factor

The 2026 oil shock differs from the 2022 inflation in its origins, but for Rosenfeld's theory what matters is the price tags, not their causes. In March, Brent topped $100 a barrel for the first time since 2022. By mid-September, prices had climbed as high as $107.80, and in October above $111. According to the Washington Institute for Near East Policy, 37 vessels were attacked in the vicinity of the strait in October, more than in any other month since the war began. November hopes for a deal between Washington and Tehran pushed oil below $100, but the strait never fully reopened.

The pass-through of the energy shock into consumer prices is following a familiar script. In Britain, September inflation jumped to 3.5 percent, with motor fuel up 29.8 percent year over year. "Inflation rose notably in September, reaching its highest annual rate in almost a year," Grant Fitzner, chief economist at the Office for National Statistics, noted on October 21. On November 5, the Bank of England raised its rate to 4 percent and published a forecast showing inflation approaching 5 percent by mid-2027. The European Central Bank raised rates in June and September. In Italy, October inflation reached 5.6 percent, in Spain 4.9 percent and in France 3.8 percent. Germany, at 2.8 percent, looks like an exception only because of a temporary cut in the fuel tax.

In the United States, a bureaucratic detail distorts the picture. Because of last year's federal government shutdown, the Bureau of Labor Statistics did not collect prices for October 2025 and was therefore unable to publish annual inflation for October 2026. A monthly increase of 0.4 percent following 0.6 percent in September speaks for itself. The last available annual figure, for September, was 3.7 percent. The country has not seen the Fed's 2 percent target in nearly six years.

Let us put the numbers together. Rosenfeld's threshold: 2.67 percent. United States: 3.7 percent. Eurozone: 4 percent. Britain: 3.5 percent, with a prospect of 5 percent. Italy: 5.6 percent. For the first time since 2023, the Western world has once again crossed the line beyond which, according to the late criminologist's model, rising prices begin to push robberies upward.

The first signal has already arrived. In the Council on Criminal Justice's July report on the first half of 2026, nine of the 13 tracked offenses declined. Murders fell 18 percent, robberies 17 percent, carjackings 47 percent and residential burglaries 13 percent. Shoplifting, by contrast, rose 4 percent and was the only category to exceed its level in the first half of pre-COVID 2019. That half-year coincided only with the first months of the oil shock, when gasoline was just beginning its climb from three dollars a gallon to four and a half, and the fall inflation figures of 3.7 percent in the United States and 4 percent in the eurozone were still far off. A 4 percent increase over six months cannot yet be called a trend, and I would not draw far-reaching conclusions from it. Nevertheless, the crack has appeared exactly where the theory predicts.

Scenarios for 2027: A Respite or a Reversal

What happens next depends more on the strait than on the police.

The first scenario is a lasting ceasefire and the restoration of shipping in the first quarter of 2027. Brent falls toward $80, and inflation in the United States and the eurozone returns to 2.5 to 3 percent by the end of 2027. In this case, shoplifting will rise moderately, robberies will continue to decline, and politicians of both American parties will go on claiming credit for things that have nothing to do with them.

The second scenario is a protracted "gray" war with periodic attacks on tankers, with oil fluctuating in a $90 to $110 range and inflation in Europe and Britain holding at 4 to 5 percent throughout 2027. Here, Rosenfeld's model predicts a reversal. First shoplifting and petty theft will rise, then the decline in robberies and burglaries will stall. Murders will probably respond more weakly and later, although in poor neighborhoods, if his work with Vogel is any guide, the link will be stronger.

The second scenario has a worse version. Central banks, frightened of a repeat of 2022, tighten monetary policy more than the situation requires, and the economy slides into stagflation. Then unemployment will be added to price pressure, and for the first time since the early 1980s the West will face both crime-producing factors at once. It was exactly this combination that lay behind the crime of the Carter and early Reagan eras.

My forecast can be checked. The Council on Criminal Justice's year-end report for 2026, due out in January 2027, will show an increase in shoplifting for the year after the 10 percent drop in 2025. Police in England and Wales will record more shoplifting offenses in the year ending June 2027 than the record 530,643, if British inflation does approach 5 percent. U.S. murders will continue to decline for 2026, and skeptics will declare Rosenfeld's theory refuted. They will be wrong: his model never promised that price tags kill as quickly as they push people to steal.

One question remains that I cannot answer. The market for stolen goods has moved from flea markets and garages to online marketplaces and messaging apps, and laws like INFORM have partially bled it dry. Has inflation retained its former power under these conditions? Rosenfeld dreamed of surveys that would track the prices of stolen goods directly. No country has such data to this day.

The Next Crime Report Will Be Published by the Bureau of Labor Statistics

In the final paragraph of his 2014 paper, Rosenfeld suggested that developed countries would not need income support measures for the poor in case inflation broke through its target range for a long time to come. The world, he wrote, appeared to have entered an era of price equilibrium. He was wrong about the timing; about the mechanism, it seems, he was right.

This leads to a conclusion politicians will not like. The key people in the fight against theft in 2027 will be not police chiefs or the authors of mandatory minimum sentencing laws, but the members of monetary policy committees in Frankfurt, London and Washington. They themselves will not accept this role: their mandate is limited to prices and employment.

On December 10, 2026, the day after the Fed's last meeting of the year, the Bureau of Labor Statistics will publish the November consumer price index. Traders will watch Treasury yields, and analysts will look at the lines for gasoline and groceries. Police chiefs ought to be looking there too. The plastic covers over deodorants that appeared in American drugstores in 2023 may be needed again in 2027.