Evidence based policymaking is something which, in public at least, most public servants now aspire to. Think tanks on both sides of the Atlantic have taken to using the term over the last two decades to describe policymaking which is rooted in evidence, data, and research rather than driven by ideological reasoning. As one might imagine, economists – as social scientists – tend to be especially keen on the notion.
But whilst evidence based policymaking sounds good in the abstract, in practice it can often be tricky. Especially when the evidence is unclear or disputed.
The decision by the Federal Trade Commission (FTC) this week to ban non-compete clauses is a case in point.
Non-compete clauses, which seek to ban employees from working for a rival business or starting their own competing firm for a certain period of time after their employment, have a long history. By some reckoning, the first legal record of such a clause ending up in court dates back to Dyer’s Case in 1414, when a former apprentice’s master sought to prevent him from operating on his own. Nowadays though such restrictive covenants have become increasingly common in employment contracts. The FTC reckons that some 30 million Americans – around one in five workers – are subject to them. So the FTC’s decision is certainly a material development.
As in much of economic policymaking, there is a trade-off at play. Restricting employees from seeking work with rivals decreases labor mobility and almost certainly reduces the effective bargaining power of labor as a whole, potentially reducing wage growth. On the other hand, firms argue that such clauses are important to protect trade secrets and to prevent firms from effectively paying for the training of staff at rival enterprises, In the absence of non-compete clauses, investment in staff could well be lower.
The FTC, which voted for the new rule by three to two, has made some bold claims about the likely results.
In their view the change will boost aggregate wages $400-488bn over the coming decade, as well as reducing healthcare costs, increasing the rate of business formation, and lead to more, not less, innovation. Lina Khan, the FTC’s Chair, set out the reasoning this week in a useful thread on X (formerly Twitter). She argued that “robbing people of their economic liberty also robs them of all sorts of other freedoms”,
The decision though is not without its critics. In another useful thread on X, Brian Albrecht– chief economist at the International Center for Law and Economics – set out the case against the change. His argument mostly rests on the lack of good evidence about the likely impacts, especially when it comes to senior executives.
In January last year, the Clark Center’s own US Economic Experts looked into the evidence around non-competes. The results do not make for reassuring reading for anyone looking for a definitive answer.
Asked whether banning non-competes would increase wages in any given industry, 49% of respondents (weighted by confidence) were uncertain, with 39% agreeing. On the question of whether a ban on non-competes would raise innovation, 42% were uncertain and 35% agreed. And asked whether a ban would lead to measurable reduction in investment in staff training 42%, again weighted by confidence, agreed and another 42% were uncertain.
In other words, the experts are distinctly unsure whether or not the various benefits of the change which the FTC expects to see will actually emerge whilst a little more sure, although still fundamentally uncertain, that some of the potential costs are real.
One reason why the experts were so uncertain was the whole host of competing potential effects at play. As Larry Samuelson noted on the wages question – banning non-competes should enhance the bargaining power of workers but a reduction in staff training could well make employees less valuable in the longer run. Or, as Oliver Hart, argued on the innovation issue – ideas might move more easily between firms but equally firms might have fewer incentives to innovate in the first place.
The real problem is the lack of evidence. The United States is something of a patchwork quilt when it comes to non-compete bans at the state level at present, but few such rules have been in place for long. The empirical evidence is simply rather limited, especially when it comes to the effects on those at the higher end of the income scale where non-competes and often lengthy periods of ‘gardening leave’ have long been the norm.
The FTC’s decision is a major one which will almost certainly have measurable impacts on a wide range of economic variables. The silver lining, for empirical economists, will at least be that the changes that result will be quantifiable. In this case, the policy itself will help to create the evidence that future decision-makers can later draw upon. This is far from the idealized world of evidence-based policymaking that regulators often aspire to.
