That, of course, won’t be an unfamiliar lesson for many right-of-center policy types, or conservatives in general, but it’s always one worth bearing in mind nonetheless, because sometimes there are things we can do about it. It came to mind when I read this piece in Governing magazine, entitled “Ohio’s Bad Idea for Boosting Welfare-to-Work.” It caught my eye because I was interested in learning why a welfare-reform program with work requirements failed. Here was the program:
[In 2013] Ohio — for the first time in 18 years — started to see [its welfare] rolls climb, a function most likely of the effects of the Great Recession. The reversal got the attention of Republican Senate President Keith Faber, who late in the 2014 session added a provision to the state budget that would give welfare caseworkers bonuses for moving clients from welfare to work. Under the Faber plan, five counties will be chosen to participate (currently it’s not clear how they will be selected).
On its face, this seems like a good, old-fashioned and sensible run-government-like-a-business notion. Not only will it incentivize caseworkers to hustle, but it may illuminate some particularly successful tactics in finding folks jobs. It’s also not an unheard of notion. There are numerous welfare-to-work programs run by nonprofits that operate on a pay-for-performance basis. Providers get a certain amount upon evaluating a client’s skills and needs, a certain amount upon training, some more upon placement, and even further payments for clients who stay in their jobs for six months, a year, two years and so forth.
As the piece notes, there’s a track record of success for a lot of welfare-to-work programs, including ones that pay for performance — Ohio didn’t have a “bad idea” per se at all. But the Ohio plan didn’t quite consider their lessons:
The well-designed pay-for-performance plans run by nonprofits account for some key variables, including the skill level or employability of clients (in order to prevent “creaming”) and the economic environment in which the task is being tackled. It’s one thing to pay bonuses in areas with relatively healthy economies and clients who are pretty much ready to roll, it’s quite another in places where the region is struggling and the clientele is especially challenging. But Ohio’s deal is typical of how way too many state legislators think about “performance”: It starts and ends with the bonus. Legislators didn’t bother to ask whether caseworkers would cream. They didn’t wonder: Will they try to shove clients into the first available, if not altogether appropriate or promising job? Will caseworkers simply reject folks who really do need assistance because their job prospects seem particularly dim?
The Governing piece doesn’t have actual evidence of what’s gone wrong in Ohio, but surely there were some unintended, undesirable consequences for the conduct of caseworkers — and misallocation of bonuses taken out of scarce state resources, because counties were ranked by performance with no way to control for the different challenges they face. These issues may to some extent be intractable, but it seems a number of state and local welfare agencies have in fact devised effective, worthwhile incentive systems for encouraging welfare-to-work — they were just more complex than Ohio’s. For good reasons, conservatives distrust complex programs. They also distrust government interventions because they have unintended consequences.
But unintended consequences aren’t always unpredictable. Many government programs create somewhat predictable distortions, and there is usually precedent for them. There were predictable distortions that would emerge from this incentive program (maybe ones of uncertain magnitude, though), just as there are predictable distortions that come from having means-tested welfare programs in the first place, such as their boosts to implicit marginal tax rates (though again, we can’t be quite sure of that effect’s magnitude). And sometimes, as in Ohio, the best answer to these somewhat predictable second-order problems is a little more complexity — incentives that take into account a county’s economic situation, that enable caseworkers to devote resources as people need without being swayed too much by the chances they will end up in a job, etc. If we want a welfare system, and we want it to encourage work, it’s going to take some tolerance for complexity.
Of course, local governments are best-suited to devise such complex solutions, because of the number of factors involved and the relative ease with which they can fix programs if they failed to consider unintended consequences. Saying Ohio really should have considered lessons from systems elsewhere is not the same as saying Ohio should be forced to do the same thing they did.