So you 'banned the box.' How did that work out?

It's hard to think of a more well-intended public policy than "ban-the-box" laws. They're designed to give a modest second chance to folks who have made mistakes and gotten into trouble with the law.

But according to fascinating (if frustrating) research presented in Minneapolis this month, ban-the-box policies are themselves having trouble with the law — the law of unintended consequences, that is.

We have here a vivid case study in the complexities of trying to transform economic and social conduct by decree. It may be useful to reflect upon amid many calls for regulations on wages, prices, rents, terms of employment and more. Swelling confidence in the power of good intentions, particularly among urban progressives, was apparent once again last weekend at the DFL Hennepin County convention, which declined to endorse veteran DFL officeholders County Board Member Peter McLaughlin and County Attorney Mike Freeman in a nod toward ever-bolder progressive agendas.

Ban-the-box research confirms the basic truth that people facing restrictions and mandates will make adjustments to protect their own interests. The result will be change — but not necessarily the change those issuing the restrictions and mandates had sought.

According to the work of Rutgers University economist Amanda Agan, ban-the-box does improve the prospects of getting a job interview for people with criminal records. But it does so at the cost of "encouraging racial discrimination" that "could harm black men, in particular those with no records … ."

To the considerable extent that ban-the-box policies have been promoted as a way to combat racial gaps in income and employment, this is an especially disagreeable side effect.

Agan discussed her findings — from an experiment she and colleague Sonja Starr from the University of Michigan conducted in 2015 — at a May 11 conference hosted by the Minneapolis Federal Reserve Bank's "Opportunity and Inclusive Growth Institute," established to investigate solutions for racial disparities.

Agan and Starr conducted an experiment (apparently the only one so far in any of 11 states with BTB laws) to see how the policy was working in New Jersey and New York. Just before and after the box was banned, they submitted some 15,000 online applications for fictitious young male job-seekers, some with nonviolent criminal records and others with no rap sheet. Paired applicants were similar in work experience, education, criminal record and so on, but were given names chosen to strongly suggest each applicant's race (e.g., Darnell Washington vs. Ryan Schmidt).

Then the researchers sat back to see which of their make-believe applicants got what are called "callbacks" for further consideration.

Before BTB took effect — when applications still asked about criminal convictions — discrimination against those with a record was clear. But regarding pure racial bias, results were not altogether terrible.

Among "white" applicants who disclosed a criminal record, only 8.7 percent got a callback. Among "blacks" disclosing a conviction, 8 percent got a callback. Applications indicating no criminal background did much better, whatever the apparent race.

Blacks with no record got called back 13.4 percent of the time, while for whites with no record it was 14.1 percent.

But once BTB took effect, the racial disparity in callback rate soared.

Employers suddenly lacking any information on job-seekers' criminal histories called back black applicants overall just 10.3 percent of the time — more than 23 percent less often than they had called back no-record blacks before BTB.

Meantime, 14.8 percent of all white applicants now got callbacks — a higher rate than no-record whites had enjoyed before the box was banned.

The researchers calculate, once all the complications are accounted for: "Before BTB, white applicants to employers with the box received 7 percent more callbacks than similar black applicants, but BTB increased this gap to 43 percent."

What happened to produce this sixfold increase in apparent racial discrimination? Agan and Starr say the most plausible explanation is that some employers, denied direct information about criminal histories, used apparent race as a "proxy" for a higher probability of criminal background and became more reluctant to call back "black" applicants.

The big losers are black applicants with no criminal records — surely an ironic and mortifying unintended effect.

Agan and Starr are at pains to note that race-based discrimination is illegal whatever its rationale. And they calculate that employers appear to significantly overestimate the statistical degree to which a young black male is more likely than a young white male to have a record (although there is a statistical difference).

But noting that tighter enforcement of hiring discrimination laws is "easier said than done" they suggest "policy makers … seek alternate strategies to increase hiring of people with records."

In an imperfect world, unwelcome facts must sometimes be faced. Good intentions are not enough, as the late University of Minnesota scholar John Brandl liked to say. If it turns out that "banning the box" encourages racial discrimination, then the box, counterintuitively, turns out to have been a useful tool for reducing racial discrimination. Its elimination may be harming young black males who, as the saying goes, have done everything right.

That sounds like a policy that deserves reconsideration.