Evidence vs effectiveness

The row between Norm and Conor about liberal intervention has descended into mere semantics. For me, though, it raises an interesting question: could it be that there’s sometimes a trade-off between the evidence for a policy and the effectiveness of it, so that evidence-based policy-making can be self-defeating?
Here’s what I mean. Immediately after he seizes power, a dictator will be weak; his enemies will be alive and uncowed, and he won’t have had time to build a large client base. So he’ll not be in a position to commit huge atrocities. In this period, using force to remove the dictator will be relatively easy - you’ll find lots of internal allies and little loyalty to the new man - but there’ll be little evidence base to justify such intervention.
However, after many years in power, the dictator will have killed or exiled his critics, will have plundered his country to buy weapons, and will have accumulated a loyal following. There’ll then be lots of evidence to justify liberal intervention, but it’ll be difficult for the policy to succeed, not least because the people who might have helped build a liberal democracy are lying in unmarked graves.
This poses a dilemma for liberal interventionism. When there’s enough evidence to justify the policy, it’s (sometimes? always?) hard to implement it. And when the policy is feasible, there isn’t sufficient evidence base to win widespread support for it.
Now, I say this not to take a view about liberal interventionism. I do so instead to raise the question: how large is the set of policies that are subject to this trade-off between evidence and effectiveness?
I can think of a couple of examples. In investing it has for years been a cliché that if you wait for evidence of an economic recovery before buying shares, you’ll have missed the rally: events since the spring have vindicated this view.
Central bankers have a similar problem. If they wait for concrete evidence of inflation before tightening monetary policy, inflation will already have taken hold and big policy moves will be needed to force it down.
At this stage, an intelligent Conservative (if this is not an oxymoron) might reply as follows:

Fiscal policy is another example. Of course, there’s no evidence now that bond markets are fretting about the deficit. But if we wait for hard evidence of their concerns, it’ll be too late. Bond yields will have shot up astronomically: we know that, when markets move, they move a long way very fast. In this situation, we’ll need humungous cuts in public spending to convince the markets to continue financing the deficit. In order to prevent this, we should cut spending a little now just to remind the markets that we’re taking the deficit seriously and so retain our credibility. Credibility is like virginity; it takes a little effort to keep it, but a lot to regain it.

This is, of course, completely different from the actual argument the Tories are using. But what’s wrong with it? Could it be that my lack of concern about the deficit is mistaken not because I’ve ignored the evidence but because I’ve paid attention to it?

Comments

[Phil]

A row with Norm has descended into ill-tempered self-righteous pedantic nit-picking? I'm shocked, shocked!

[Luis Enrique]

there's evidence as in evidence that "we need to do something" and there's evidence as in "what we propose to do is likely to work"

I think you're talking about the former ... a case in point is whether a foresighted politician / bureaucrat in 2006, attempting to forestall the financial crisis, would have had the evidence needed to push through the necessary reforms.

["descended" into semantics? I thought it was very helpful to clarify how various terms are being used. Most of these arguments boil down to people talking at cross purposes]

[Leigh Caldwell]

I suspect your argument is the same as the one the Federal Reserve is using, to Scott Sumner's horror.

By making really sure that everyone knows they'll be really really serious about keeping inflation down, they probably won't have to bother actually doing it.

I suppose the category of policy to which this hypothesis applies is those with multiple equilibria and self-reinforcing dynamics (positive feedback). In these cases, small nudges to the initial conditions can dramatically influence the outcome - hence the impact of making policy early, before there is evidence to do so.

The clear commitment of world governments to using overwhelming firepower against the financial crisis seems to have worked; to a lesser degree their similar commitment to fighting recession. Unfortunately, those successes mean that we'll never have any evidence that such drastic actions were needed. So the governments which took them will be more criticised for running up debt than praised for averting catastrophe.

[Luis Enrique]

Leigh,

you raise a related problem: evidence based voting - what evidence voters have available with which to evaluate politicians. To use the same example as above, if some farsighted genius had managed to forestall the financial crisis (let's move this hypothetical story to the USA, where it works better), it would have probably involved a mini-recession (higher interest rates, less house-building, lower banking profits etc.) and all voters would have seen was the pain this politician caused, and not what was averted. Most likely, the genius would be voted out, next election.

[Diversity]

The world's fiscal deficts have to come down - sooner or later. The trick is to maintain confidence in the interval. Confidence is about fear - fear of the unknown. The future is unknown and fearful. So you construct a story to make appear what is to come more familiar and reassuring.

Announce a serious looking programme of expenditure reductions extending over several years. Keep the timetable "realistically" fuzzy at the edges. Some of your reductions will happen as those administering programmes find ways to save. Some savings will run late. Events will have "forced" some new expenditure. Improve the accounting as necessary to show convincing evidence of progress on expenditure restraint.

At the same time set out serious options for tax increases. Next year set out the timetable for the chosen options. Postpone parts of the options in later years as necessary to avoid putting a crimp in the economy.

Do all that intelligently and markets will give you great credit for successful and consistent elimination of a serious defict. Moreover, you will have kept the economy on a relatively smooth recovery path, improved your public expenditure mangement and lessened the net economic cost of your tax system. Accumulated public debt will be high, but that is something we can take our time growing out of.

In nudge terms, you have announced early, as intent: - much as central bankers sometimes talk early about the need to tighten monetary policy in the hope of avoiding some of the real tightening. You then actually act as the current sitiuation appears to require; revising your announced view of the future as seems most likley to avoid triggering damaging outcomes.

To return to where Chris started the discussion, the appropriate policy is rather like the jaw-jaw with talk of sanctions which is the preferred option to armed liberal intervention; preferred because the damage likely to be done by what often becomes a quasi-colonial war is not the preferred outcome. Similarly, talking an economy out of a recession is likely to do much less damage than strict fiscal discipline until we are out of it.

[jameshigham]

there’s sometimes a trade-off between the evidence for a policy and the effectiveness of it, so that evidence-based policy-making can be self-defeating?

That's well put, Chris, though the details may be argued over.