Should evidence determine policy?

There seems little to choose between evidence-based policy-making and policy-based evidence-making. See this discussion: "the idea that we've moved from ideology-based policy-making to evidence-based policy-making...is completely misleading, because the evidence-gathering process is itself value-laden." Life is so rich and complex, and the stakes so high, that so-called experts will always be able to find evidence that justifies whichever policies serve the interests of the highest bidder. So my answer to the question in my header is: no. And my suggestion is just as simple: outcome-based policy.

Emissions trading not working - who would have thought it?

No surprises here:

Emissions trading, the European Union hoped, would limit the release of harmful greenhouse gases. But it isn't working. The price for emissions certificates has plunged, a development that is actually making coal more attractive than renewable energy Source

Here's another idea: instead of using fossilised science, and then allowing corporate interests and their friends in government to dictate policy, why not target the outcome that we want to achieve? A more stable climate, defined in terms of some combination of human, animal or plant health indicators. And why not reward people for achieving this outcome, rather than for performing some activities that might, but actually don't, achieve it. See here for more.

Bad Policymaking

Ben Goldacre writes:

[F]or ... assessing causal relationships, intuitions are often completely wrong, because they rely on shortcuts which have arisen as handy ways to solve complex cognitive problems rapidly, but at a cost of inaccuracies, misfires and oversensitivity. Bad Science, Ben Goldacre (page 238)

This becomes a problem for policymaking under the current system, where government is often charged with identifying causal relationships and, if it gets them wrong, the consequences can be calamitous. For murky, complex social, economic and ecological relationships, we need a mechanism other than the one-size-fits-all, top-down approach that is a feature of the current policymaking system (and which can work well, when causal relationships are obvious). Central planning, a catastrophic failure when applied to economies, fails too in policy areas where there is a compelling need for diverse, adaptive approaches.

But central planning is precisely the approach we are taking in tackling extremely complex social and environmental problems: climate change, or warfare, for instance. They seem to be failing in much the same way as in the economy. There's little response to expanding knowledge or to changing circumstances. There's little diversity. Most important, failed approaches aren't terminated. The incentives are to maintain existing institutions, rather than to achieve the stated outcome.

That's where Social Policy Bonds could be a better alternative. They would reward people only if a specified outcome is actually achieved and sustained. A body that issues the bonds doesn't have to have an opinion about what causes a social problem. It just has to reward the people who solve it. Incentives are built into the system: only efficient solutions will be rewarded.

For more about Social Policy Bonds, please download my book (see right-hand column).

The triumph of process

It's a familiar story: in so many policy areas - health, education, the environment, for instance - adherence to process is more highly rewarded than socially desirable outcomes. So we have a blizzard of micro-targets combined with a disintegrating physical and social environment and a disengaged electorate. Policymaking itself is an arcane process, comprehensible only to those who are paid to participate in it or who are lobbyists for powerful interests.

And it's the same in the justice system, at least in the US, where "Six million people are under correctional supervision in the U.S.— more than were in Stalin’s gulags." Adam Gopnik explains:

accused criminals get laboriously articulated protection against procedural errors and no protection at all against outrageous and obvious violations of simple justice. You can get off if the cops looked in the wrong car with the wrong warrant when they found your joint, but you have no recourse if owning the joint gets you locked up for life. You may be spared the death penalty if you can show a problem with your appointed defender, but it is much harder if there is merely enormous accumulated evidence that you weren’t guilty in the first place and the jury got it wrong. The caging of America, Adam Gopnik, 'New Yorker', 30 January

Mr Gopnik's article goes on to point out that the large falls in US crime rates over the past three decades, especially in New York, have many explanations, few of which could be known in advance.

I think this well-written piece helps make the case for targeting outcomes, as I have advocated, whether or not by using Social Policy Bonds. Society is so complex that a single group of policymakers cannot know in advance with any certainty the underlying relationships between, say, prison sentences and crime rates. Or between spending on schools and literacy. Or between greenhouse gas emissions and the numbers of people killed or made homeless by adverse climatic events. Where cause and effect are clear - as say, between inoculation rates and disease - there is a strong case for government working to achieve a social target. Where it is not there is still a strong case for government setting the target and raising the revenue for its achievement. But, instead of trying to achieve it directly, it would, I think, do better to contract out the achievement to a motivated, diverse and adaptive private sector. Social Policy Bonds are one way in which this division of labour could be carried out efficiently.

Japan is different. Oh really?

But the LDP [of Japan] shows the same intransigence that has been its stock-in-trade since it lost power in 2009. It vows to block the tax bill, even though raising the consumption tax has long been a plank in its own policies. Generational Warfare, 'The Economist', 28 January

Once more, we see the unimportance of outcomes to today's politicians. The right policy rejected because it's proposed by the people on the other side. We - that is, the entire democratic world - desperately need a new political system. One that targets outcomes that are meaningful to ordinary people, not politicians, corporates or government agencies. The old system is just not fit for purpose.

New Zealand opts for Mickey Mouse micro-targets

Cancer patients who need chemotherapy should receive it within four weeks of being assessed, under new [New Zealand] government health targets. The cancer target at present requires hospitals to provide radiation therapy within four weeks of assessment. From July 1, hospitals would also have to ensure that patients needing chemotherapy received treatment within four weeks. Source

It's a shame that the New Zealand Government has learned nothing from other countries' mistakes. Targets like this have nothing to do with the broader health of the population. They will be so manipulated as to become meaningless or, worse, divert resources into (apparent) compliance and away from health care. In the UK, for instance, we have seen ambulances delaying their arrival at hospitals, so that targets for seeing patients within four hours of arrival can be met. It's not difficult to imagine ways in which New Zealand's new chemotherapy target will be similarly gamed, at the expense of people's health. This is the sort of micro-management that did so much to cripple the Soviet Union. Private corporations with a narrow focus on a few accountancy ratios are prone to similar errors. In theory at least competitive pressure would ensure that the mismatch between targets and reality cannot continue to worsen indefinitely. (In practice, if the corporations are big enough, they subvert government and change the rules.) But when government applies these micro-targets, there's little to bring them back into line. High-sounding, well-meaning experiments like this are rarely terminated.

We need to bring government back to its core focus: if its goal is to improve the health of its citizens, that's what it should target. Let a competitive private sector work out how best to achieve that goal. Government can still set broad health targets, and it can, and should, raise the revenue to achieve those goals. But it cannot possibly keep up to date with science, nor respond adequately to changing events or diverse circumstances. Only something like a Social Policy Bond regime, where people are rewarded for being efficient achievers of meaningful targets, can do that.

Mickey Mouse micro-targets

Numerical targets, though they can never accurately measure everything of importance, are going to have to play a role in determining the efficiency and effectiveness of policy instruments. Many of our social and environmental problems can be attributed to (1) the private sector's use of accountancy measures, to the exclusion of anything else, in evaluating its own performance, and (2) the failure of the public sector to use any meaningful numerical targets.

Governments do use plenty of meaningless targets. Here's Theodore Dalrymple:

The Times Educational Supplement is Britain’s most important journal for the teaching profession. In the January 6 edition, it described the methods school principals use to deceive the official inspectorate of schools. The inspectorate’s reports, in the words of the TES, “are vital checks on the performance of schools, relied on and trusted by parents and those running and working in the system.” The precise extent of the principals’ cheating is, in the nature of things, difficult to measure. But once the principals know that an inspection is coming, many employ techniques such as paying disruptive pupils to stay home, sending bad pupils on day trips to amusement parks, pretending to take disciplinary action against bad teachers, drafting well-regarded teachers temporarily from other schools, borrowing displays of student work done in other schools, and so forth. It’s Gogol’s Government Inspector translated to the educational sphere. The Less Deceived, 'City Journal', 10 January

What government should be doing is targeting broad measures that are meaningful to real people. Real people, as opposed to government agencies or corporate accountants. In education government should be targeting, at the very least, functional literacy and numeracy.

The system is rigged

David Brooks asks Where are the liberals?:

Americans...don't trust the federal government. A few decades ago they did, but now they don't. Why don't Americans trust their government? It's not because they dislike individual programs like Medicare. It's more likely because they think the whole system is rigged. .... [R]ent seeking groups are dispersed across the political spectrum. The tax code has been tweaked 4428 times in the past 10 years, to the benefit of interests of left, right and center. 'International Herald Tribune', Asia Edition, 11 January

This malaise is common to all the western democracies. Interestingly, Mr Brooks mentions sugar subsidies, which benefit just a few wealthy individuals, while "imposing costs on millions of consumers". It's the persistence of such subsidies, in the face of decades of evidence pointing out their disastrous economic, distributional and environmental impacts, that makes one despair about whether our governments can ever reform themselves. And, if they can't, then where is the initiative going to come from?

Perhaps the most benign impetus for reform would come from a shift toward rewarding outcomes, rather than, as now, the specific interest groups - public or private sector - that currently seem to run government. Social Policy Bonds would subordinate all government funding to meaningful results. Under a bond regime only the most efficient achievers of social and environmental goals would receive taxpayer funding.

Clearly the current system is losing the consent of the majority of the people it's supposed to serve. Perhaps it's time to try Social Policy Bonds. My book suggests how a transition to a bond regime need not be too drastic, but could be gradually managed.

Selective memory

From The Great Derangement, by Matt Taibbi:

[US] national politics was doomed because voters were no longer debating one another using a commonly accepted set of facts.

Life and society are so rich and complex that we can easily extract evidence that supports (or appears to support) virtually any outlandish claim we want to make - or its complete opposite. There are plenty of funds and incentives to make the effort worthwhile. Sometimes these exercises are cynical. But often not. We can see this very clearly when looking at the current financial crisis: respectable commentators put forward coherent arguments for much more, or much less, stimulus spending.

Social Policy Bonds would bypass such arguments and all their attendant cynicism by targeting desired outcomes, rather than what people think, or are paid to say they think, are the means of achieving them. It would be up to holders of the bonds to identify and exploit the relationships between cause and effect, which is so difficult to do in our complex society. By focusing on outcomes, policymakers could concentrate on working out exactly what we want to achieve, rather than get bogged down by the interest groups, vested or not, who have their own agenda.

If this sounds far-fetched, we need look only at climate change to see how the debate has been effectively side-tracked into an expensive and ineffectual irrelevance by doing things the conventional way: trying to prove something to the satisfaction of people who oppose doing anything before taking action. A bond regime would instead be rewarding people for achieving certain specified goals, or a combination of them, which could be expressed as a wide array of physical, financial and social variables. It would be up to the investors in the bonds to work out the relationships, and they would be rewarded for doing so and for continuing to do so, until our goals have been achieved. In this case, as in others, there need be no general (and often impossible to achieve) agreement on the facts before taking meaningful action.

Half full

It's nearly the end of the year, so time to sum up my view of where policymaking in general and Social Policy Bonds in particular are headed. The latter is easy to summarise: I have had a few expressions of intellectual interest in Social Policy Bonds, but the bond concept remains untested. More generally though, I think the concept of payment for results is gradually gaining ground in government circles. This is generally a positive trend, though I have reservations. Chief amongst these is that the 'results' targeted are rarely outcomes that are meaningful to ordinary people. More often they are outputs of existing organisations, as measured by criteria intrinsic to that organisation. So results targeted include things like 'savings made' or (increasingly) number of employees sacked. In recent history it has usually been the case that well-meaning intentions to improve social and environmental outcomes founder on the solid, immovable rocks of existing institutions and their ways of doing things.

But I remain optimistic that something like Social Policy Bonds will eventually be issued. One reason is that society and the environment are becoming increasingly complex, so that existing methods are becoming useless. We only need to look at climate change for a spectacular and costly example of failure to manage human affairs. The other reason is that the existing system's failings are becoming obvious to all. Our economic system has been gamed to benefit the one percent - or 0.1 percent - and our political systems throw up uninspiring candidates for whom the concerns of most of us and the environment appear well down the list of things to worry about. It's clear that the current system cannot continue. I remain hopeful that someone, somewhere, in the public or private sector, will issue Social Policy Bonds for a worthwhile goal, and enable the concept to be tested, discussed, refined and eventually deployed to solve some of the world's urgent social and environmental problems.