RIAs - Regulatory Impact Assessments or Results In Advance?

The Adam Smith Institute blog pointed to the new publication on Regulatory Impact Assessments (RIAs) from the Centre for Policy Studies (CPS) a couple of days ago, but this is a sufficiently important issue that I didn't want to simply let it go past because I hadn't had time to comment immediately.

All proposed legislation and regulation nowadays comes with an RIA attached. They are supposed to weigh up the costs and benefits of the proposals, to ensure that those proposals are not unduly burdensome.

The CPS publication entitled "RIAs: why don't they work?", prepared by Keith Boyfield for submission to the new Business Council for Britain, assesses RIAs and finds that they have failed in their objective. This will be no surprise to those who have suffered the effects of most legislation and regulation enacted since RIAs were introduced. It would be hard to argue that the burden of the state has fallen more lightly on the shoulders of its citizens in recent years.

Mr Boyfield provides an accurate diagnosis of many of the causes of the ineffectiveness of RIAs. They could be summarised as "no one in government wants anything more from them than a fig-leaf behind which to hide any flaws in a pre-determined set of measures". They are exercises in self-justification, not critical analysis.

Sadly, Mr Boyfield's prescription is to try to improve the process by which RIAs are prepared, rather than simply to scrap them. My objection is not that efficiency in government isn't important - it is vital - but that asking government (or its sub-contractors) to measure in advance the effectiveness of its own proposals is bound to end up with skewed results, however many refinements you introduce. Better to aim for clarity on the actual (rather than predicted) costs and impacts of government, and then let voters judge every five years or so whether they are getting value for money.*

Let's consider Mr Boyfield's suggestions in detail. He leads with:

Set up a central register, available on the internet from where draft and completed RIAs can be downloaded.

Have you ever read an RIA? They are impenetrable gobbledigook. Making it easier to find them en masse will not help to bring them under greater critical scrutiny. And anyway, why would you want all the RIAs together and separate from the documents to which they relate? Each RIA can usually be found on the same web-page as the proposal to which it relates. To the extent that they have any relevance at all, it is only in the context of the proposal, so this is the appropriate place to reference them, not some central register.

Introduce a scorecard system to ensure that RIAs have been properly drawn up. The Permanent Secretary or his or her equivalent in a government body or agency should be responsible for explaining a low score.

That should solve the problem of too much red-tape - applying bureaucratic measures to the assessment of the document that assesses the effectiveness of bureaucratic measures. And none of his suggested questions for the scorecard will prevent the people preparing the RIA from doing the minimum necessary to tick the boxes. The result will be RIAs that are no more effective but harder to criticise.

Bolster a new culture to reward better quality RIAs. Officials’ promotion prospects should partly hinge on the quality of their contribution to effective RIAs. Establish an annual awards events for the best – and worst – RIA of the year.

What is meant by "quality" and "effective" in these contexts? We know Mr Boyfield means "better", and "more critical". But amongst those who are paid to prepare them, these terms will be interpreted as "serving the purpose of justifying the measures", exactly as is already the case. Ultimately, it is unrealistic, and probably undemocratic, to think that civil servants or paid consultants are going to commit career hara-kiri by carrying out RIAs that undermine the legislative intent (right or wrong) of the elected government.

bring in an independent governmental body to ensure that RIAs are correctly carried out, and that they take account of best-practice techniques (this would follow the US and Dutch precedents

Oh good, another quango. And we've noticed just how independent-minded most quangos seem to be, haven't we? Their placemen are always criticising the political masters who decide the appointments to these lucrative posts. Why would the government turkey vote for Christmas by inviting genuinely independent criticism? It's not going to happen - if any supposedly independent body were introduced, you would know that the government had a way to keep them in line.

contract-out the RIA task to outside specialists, such as consulting firms or management consultants. This might prove particularly useful in those areas where RIAs have been seen as inadequate or where there is a particular need for a counter-balance to the sponsoring department.

Here's the rub. Would that be outside specialists like Keith Boyfield Associates, by any chance? And how many contracts will KBA be awarded if their RIAs are too critical?

Simplify the RIA process by stripping out the superfluous ‘tests’ that have been added over time, notably the so-called ‘rural proofing’ test and the requirement to assess whether a proposed regulation encourages ‘sustainable development’.

You say po-tay-toe, I say po-tah-toe, let's call the whole thing off. Why are these any more superfluous than the rest of the tests in an RIA? What Mr Boyfield regards as superfluous, others may (presumably did when these tests were added) regard as fundamental. The problem with cost-benefit analysis (CBA), which is effectively what an RIA is a form of, is that it is attempting to compare many things which either cannot be measured in the same units, or often cannot be measured at all. CBAs attempt to turn all sorts of subjective qualities into values in pounds, which can then be aggregated and compared, but once you include some aspects that cannot accurately be measured in pounds, you have a hard time justifying the exclusion of other things, simply because you don't think they are as important. I realise this is anathema to most (mainstream) economists, but my firm view is that not just RIAs but CBAs more broadly are irrational concepts, which serve mainly to provide employment to economists, and not to illuminate any debate.

Make post implementation reviews mandatory. An objective, quantified view on whether a specific regulation has achieved its goal and whether this has been done on a cost-effective basis is needed.

Who will determine the objective, quantified metrics for assessing the regulation? The metric for a hike in the minimum wage might be that the lowest-paid jobs were paying wages equivalent to at least the new minimum wage. The measure would very likely pass its assessment, regardless of whether this "improvement" was achieved by raising the wages of the lowest paid, or simply putting all those paid less than the minimum wage out of a job. Why is any rational government going to do anything other than ensure that the metrics used are ones that they are confident they will satisfy, and which can be spun as success?

Introduce a sunset clause to all regulations so that Parliament can assess whether a statutory intervention is still justified. Do its objectives remain justifiable? Can they be achieved and does the regulation provide the most cost-effective solution?

Subject to the same difficulty of enforcing an objective and independent metric of success. This will simply make Ministers a little more careful about defining their notion of success sufficiently flexibly that they cannot fail.

Improve parliamentary scrutiny of RIAs. Select Committees in particular should be able to draw on the experience and expertise offered by the National Audit Office (NAO) so that they can scrutinise the quality of the RIAs produced by each individual department.

According to Mr Boyfield, there are over 400 RIAs produced every year. Do MPs have either the time or the skills to scrutinize each one thoroughly? He might usefully remember Ludwig von Mises's warning:

"It would not substantially impair the power of the bureaucrats, if they were under the necessity of approaching Parliament for legislating on all these matters. Parliament would be flooded by a multitude of bills the contents of which would extend beyond the range of its competence. The Members of Parliament would lack both the time and the information to examine seriously the proposals elaborated by the staffs of the various agencies."

There is no way of overcoming the incentives of all involved in government to manage self-imposed "scrutiny" in such a way that it ensures a desirable outcome. Mr Boyfield's prescriptions are so naive that you would think that, as an economist, he had never heard of public-choice theory. A pity, because his criticisms of RIAs are well-aimed and important.


* Of course, the way that this clarity would historically have been achieved would have been through diligent and skillful questioning by opposition and media. Sadly, too few opposition politicians and journalists seem equipped nowadays to ask the right questions. It is an indictment of both groups that some of the most piercing questioning comes from the Government's own benches - from people like Frank Field and Bob Marshall-Andrews.



I notice that I provided a philosophical deconstruction of the use of CBAs, RIAs, EIAs etc back in February. I believe these arguments still hold water. And they link nicely to recent discussions on happiness research.