Accountability, transparency and their unintended consequences

Everybody is in favor of accountability. Certainly in a democratic society, people in government should be required to explain their actions. Even in real life -- where “accountability” is just a euphemism for “punishment,” as in, “Where is the accountability?" -- people who perform badly are expected to suffer the consequences.

And everybody also is in favor of transparency. What principle of democracy could be more basic than openness? And transparency, of course, makes accountability easier.

Well, a recent discussion during an executive education program for senior federal career managers, which I’ve been teaching, illustrated that it’s not quite so simple as all that.

Contrary to what everyone believes -- that is, every everyone except people who actually have experience managing organizations, as well as scholars who study management -- there may be tradeoffs between accountability/transparency on the one hand, and performance improvement on the other.

One participant noted that in order to improve how his organization was managed, he needed people working for him to be willing to provide information about problems -- to be willing to state they were "red," as the participant put it. However, to make that possible, he as a leader needed to create a safe zone in which subordinates could discuss problems without fear of exposure or punishment. As it turns out, these principles that everybody likes -- accountability and transparency -- actually create the incentive to hide information, and thus makes it harder to work to solve problems before they blow up.

This tradeoff has actually been studied by academics. Work by my colleague, Amy Edmondson, at Harvard Business School indicates that hospitals that do not punish people for medical errors actually improve their medical-error performance over time more than those that do. A non-punitive environment encourages people to reveal errors and allows learning about how the organization's processes should be restructured to reduce them.

However, in the punishment-oriented world of politics and the media, this approach is highly counter-intuitive, to put it mildly. We need an honest discussion of how best to balance these different considerations.

It is an issue federal CIO Vivek Kundra needs to worry about in terms of how he manages TechStat reviews of IT projects -- which, unfortunately in my view, are being labeled as "accountability" sessions.

On the one hand, persistently incompetent or seriously unethical behavior should, of course, have consequences. On the other hand, problems are often systemic and organizational, not the result of individual incompetence or evil, and the response needed is process re-engineering, not individual punishment.

One approach is for leaders to try to carefully separate performance measurement for purposes of organizational improvement through learning from individual performance appraisal for accountability purposes. The Army does that with its "after-action reviews." Another idea is to create a "non-transparency" window of several months -- conceptually analogous to the "pre-decisional" exclusion under the Freedom of Information Act -- for information about problems that subordinates provide. This window would allow the information to go up the chain and for management to develop early measures in response.

Given that "accountability" and "transparency" are like Mom and Apple Pie, few people want to acknowledge these tradeoffs. This is yet another barrier to improving the performance of government.

Readers, what do you think about that? Submit a comment below.