To people in education, the Wells Fargo scandal sounds eerily familiar.
• Top executives tried to increase performance by setting ambitious goals for opening new accounts, and attaching big rewards and penalties.
• Managers believed internal reports showing that the performance incentives worked.
• The incentives did work to some extent, in that the number of bona fide accounts increased.
• However, front-line workers were so afraid of failing to meet quotas that hundreds turned to falsifying results—for example, opening phony accounts.
• Customers were hurt by fees for accounts they never opened, the bank paid for unearned bonuses, and the whole enterprise, particularly its top leaders, were discredited.
This all sounds a lot like the Atlanta public schools test-cheating scandal of 2013. The superintendent relied on cash incentives for raising test scores and believed the hyped numbers she (and the public) were shown. Teachers and administrators in all schools tried hard to raise test scores, and some improved. But in many schools, fear and cynicism took hold so that test prep replaced other instruction and faculty and administrators falsified test scores. Real harm was done when parents were deceived about levels of student performance, and children were denied catch-up opportunities. The superintendent and some others ended up as defendants in the criminal justice system, and the once-celebrated Atlanta school system was back to square one.
But there is one big difference. In the Wells Fargo case, no one is saying that measuring and incentivizing results inevitably leads to fraud, or that business people just can’t stand up under any performance pressure.
Critics like Senator Elizabeth Warren aren’t questioning the need for outcome measures and performance incentives. What they are saying is that management was negligent, didn’t anticipate the possibility of cheating, and didn’t both look for it and create strong counter-incentives against it. Nobody is saying that banks shouldn’t follow their bottom line and link pay to productivity. That’s because something important is at stake—depositor and shareholder money—and it would be absurd to run a bank without measuring results closely and assessing people and programs on their contribution to the bank’s solvency.
In the Atlanta case, anti-reformers said something else: that testing and performance incentives are inevitably corrupt, that teachers can’t handle performance pressure—making fraud inevitable, that performance measures are deadly, so district and state leaders should avoid using them at all.
What was at stake in Atlanta couldn’t be more important—children’s current and future opportunities. But the anti-performance accountability crowd preferred something no banking critic would ever countenance: just trusting educators to do well.
Unlike today’s critics of Wells Fargo, yesterday’s commentators on Atlanta didn’t consider the possibility that top managers could assess whether set performance goals needed to be adjusted, whether reasonable security measures would prevent cooking the books, and whether claimed results were real.
Of course, Atlanta’s education managers could have done all of those things—for example by looking into performance claims by schools that had done nothing visible to improve instruction, taking answer sheets out of the school the day tests were administered instead of leaving them in the schools’ custody for weeks, and comparing reported test scores with an external standard like NAEP.
Wells Fargo is learning a hard and correct lesson—that performance incentives need to be realistic, that results must be checked, and that managers must question rosy results. In the analogous Atlanta case, many education leaders drew the wrong lesson: that school systems shouldn’t measure and reward performance at all.
That wrong lesson is still burdening state and local efforts to design accountability systems. The right lesson is that performance measurement and incentives are indispensable but cheating shouldn’t be easy. Accountability systems need to include test and answer sheet security and penalties for overuse of test prep and falsification of results. Then, the designers of state accountability systems would no longer have to tiptoe around using direct measures of student learning.
— Paul Hill
Paul Hill is Founder of the Center on Reinventing Public Education and Research Professor at the University of Washington Bothell.
This post originally appeared on The Lens, the blog of the CRPE.