JERRY BROWN can point to only one clear policy achievement since he became governor of California again last year. It has nothing to do with California’s ongoing fiscal crisis, which Mr Brown has so far failed to solve: this week he announced that a budget deficit he had previously estimated at $9 billion has swelled to almost $16 billion. Instead, it concerns the equally pressing disaster in the state’s prisons. Mr Brown’s sweeping reform, if it goes as planned, might alleviate this crisis and become a model for other states. But it is risky.

Excessive incarceration is an American problem. The country has about 5% of the world’s population but almost 25% of its prisoners, with the world’s largest number of inmates and highest per capita rate of incarceration. California eagerly participated in this trend of locking up ever more people. During Mr Brown’s previous stint as governor in the 1970s the state switched to more inflexible sentencing. It then spent another two decades adding “tough-on-crime” laws that kept extending sentences even for minor crimes.

The resulting prison-building boom, and rapacious bargaining by the prison-guards union, meant that state penitentiaries became the fastest-growing major cost in the state budget. California’s 33 prisons and associated camps therefore bear no small responsibility for the state’s recurring budget crises, and the resultant crunch on school and university funding.

Simultaneously, the prison craze also caused a humanitarian crisis as the places became overcrowded with inmates. The human misery stayed somewhat hidden because most state prisons are far away from cities, often in the dusty inland deserts. But at one point the prisons housed twice as many inmates as they were built for, and a lawsuit made its way to the US Supreme Court. Last year the court ruled that California must reduce its imprisoned population to restore humane conditions.

So Mr Brown had to do something. And he chose a policy that his wonks call “realignment”. This refers broadly to a reallocation of power from the state to counties. Specifically, it means shifting responsibility for punishment from prisons, which in America are state or federal operations, to jails, which are run by counties and their elected sheriffs.

Inmates already doing time in a state prison will stay there. But since October, anybody in California who commits a new crime that is non-serious, non-violent, and non-sexual (or “non-non-non”) has been sent to a county jail instead of prison. The main difference is that sheriffs and their deputies have much more discretion over how to deal with such offenders than state-prison wardens do.

Sheriffs can, for example, send troublemakers to mental-health treatment instead of jail. They can “flash-incarcerate” people for just a few hours. They can put them under home surveillance with a GPS monitor strapped to their ankle, or make them do community service and drug rehabilitation. They can refer them to vocational training so they can get jobs.

A similar change applies to everybody now released from state prison. Before October, all these people were automatically on “parole” (a state term). And about two in three parolees soon ended up back in jail, usually for technical hiccups, such as a missed meeting with a state parole officer. This revolving door of recidivism has now largely stopped, as former prisoners enter “probation” (a county term) and work with a local officer.

The flow of new admissions into the prisons has thus begun to dry up. Capacity is defined as one inmate per cell, or 80,000 inmates. And overcapacity is down to about 155%, which may not sound like an improvement but is. To comply with the Supreme Court, the number must fall to the oddly precise level of 137.5% by next year. Matthew Cate, California’s prison boss, says that the state has met the court’s first target for lowering prisoner numbers and might have to ask for only a short extension to satisfy the final count.

A new question, however, is whether realignment is merely shifting a humanitarian disaster from the state to its 58 counties. The reform will work in the long run only if the overall number of people behind bars in California declines. Optimists say that it will, because sheriffs can do so many things to release inmates safely. For instance, about three in four people in jail on any given day have not been convicted of anything, and are there only because they can’t afford bail, because bail amounts have risen so much. Many of these can be let out until their trial.

But pessimists worry that Mr Brown’s funding formula encourages the old and wrong approach. As part of realignment, the state currently allocates money to counties based on how many prisoners each county sent to state prison before the reform took effect. This formula rewards lock-‘em-up counties such as Kern, Riverside and Shasta. It punishes counties such as San Francisco or Alameda that already had modern rehabilitation practices. Since sheriffs can use the money as they please, some might simply build more jails, then fill them up again.

Alternatively, they might let jails become as overcrowded as prisons were. Merrick Bobb, the independent counsel of Los Angeles County, estimates that its jail population rose from 14,500 in 2011 to 17,000 this February and could grow to 21,000 by 2015 due to realignment. This would raise “inmate tension”, he notes, and thus violence in a jail system that is already under investigation for abuses of inmates (see article).

A lot will depend on how the state measures the progress being made by different counties. How do the changes affect crime on the streets? How ready are sheriffs to experiment? The overall cause for optimism is a gradual change in attitudes. In recent decades “America has seen prison as a place to throw people away,” says Mr Cate, whereas “Europeans see prison as place people will return from.” Californians, he thinks, have at last begun to reconsider the meaning of “corrections and rehabilitation”, which is what they’ve officially called prisons all along.