At the Treasury Select Committee yesterday, in between being quizzed by MPs angry at the extent of the pre-budget leaks, George Osborne revealed that he was planning to enable the OBR to make greater use of so-called "dynamic scoring" when they prepare their economic forecasts – but only when examining tax plans. It may not sound it, but if he made that change, it would be a huge coup for the right.
Dynamic scoring is the practice of including behavioural changes in economic forecasts. Without it, in a static model, changes to revenue after something like a tax hike are facile to calculate. If VAT at 17.5 per cent takes in (say) £175bn, then VAT at 20 per cent will take in £200bn. Similarly, if you are trying to calculate the effects on government revenues of spending £100m buying new trains for the east coast main line, then you can conclude that it will cost £100m – you buy the new trains, everyone carries on riding the line, and nothing changes.
What a static model says is that people are blind to the effect that government actions have when they are choosing what to do. Clearly, that is untrue. But it is also very easy to calculate, which is why it has stuck around for so long.
Dynamic models include all these changes of behaviour. So when looking at a dynamic model for rail investment, you can take into account the fact that nicer trains will make people more likely to travel, boosting ticket revenue; and when looking at a dynamic model for VAT rises, you bear in mind the fact that as prices rise, people buy fewer things, so that an 11 per cent rise in VAT won't lead to an 11 per cent rise in VAT revenue.
The problem for economists is that dynamic models are a lot harder to calculate. Not only are there far more variables to take into account, but you open yourself up to more avenues of attack from political opponents. The combined effect of all the assumptions you must make means that your conclusions can be deeply flawed if you get too many wrong, and, as the IFS says (pdf), can "[open] the door to large controversies if these guesses are made – or perceived to be made – in a politically biased way." While static models contain assumptions that are obviously false, dynamic models still contain assumptions, which can still be just as false. And unlike static models, we can't know which ones these are until it is too late.
So why does George Osborne want to introduce dynamic scoring into OBR models? Does he believe that the government's economists have uncovered a breakthrough understanding in behavioural responses to government intervention?
A clue can be found in HMRC's report into the 50p tax rate, which found that dropping the rate down to 45p would cost far less than expected, because of expected behavioural changes. Indeed, this sort of dynamic thinking is the cornerstone of the Laffer curve, the economic theory that, past a certain point, increasing tax rates reduces revenues.
What Osborne will be hoping is that allowing the OBR to make its models more dynamic still will strengthen the rationale for scrapping the 45p rate entirely, as well as for cutting corporation tax all the way down to 20p by the end of parliament.
Yet while the OBR will be "informed" by these new dynamic models when it comes to tax policy, they won't be taking account of the same information when examining the justification for government investment, where, far from strengthening Osborne's preferred policy, it would weaken it. If dynamic scoring is indeed implemented in this narrow fashion, Osborne will have scored a stealth victory for future Tory policy.