Governance quality, measured

Anyone assessing a country reaches for the indices, and they are worth using with an understanding of what they actually measure.

The main limitation

Most widely quoted governance indicators are perception-based: they aggregate the views of experts, business surveys and risk analysts. Perception is a real thing and it moves slowly, because people update opinions long after facts change.

That produces a systematic lag. A country that reformed its procurement three years ago is still scored partly on the reputation it had before.

Where perception and experience diverge

Surveys that ask people whether they personally paid a bribe in the last year produce a different picture from surveys asking whether corruption is a problem. In Ukraine the experience measures have improved faster than the perception measures, which is exactly the pattern you would expect where the reforms have been real but the reputation has not caught up.

It is also the pattern you would see if reforms were cosmetic and only the encounters changed. Distinguishing between the two requires looking at the mechanisms, not the scores.

What I would look at instead

Verifiable operational data: procurement participation rates, the proportion of single-bid tenders, time to obtain a permit, audit overturn rates, court enforcement times.

These are boring, published, and cannot be moved by opinion.

The fair use of indices

As a trend line over years rather than a score in a year, and always alongside something measured rather than reported.

That the indices lag is why my own experience often disagrees with the ranking: perception measures the reality of two years ago. I look at four things instead — the time to register a company, the refund date, the customs clearance time and the court calendar. All four are measurable and all four show up on my invoices. The fair use of an index is for comparison, not for a decision.