The government set up the Office for Budget Responsibility to keep Treasury decisions non-partisan. Now ‘dynamic scoring’ is bringing back politics by the back door
The chancellor has a new weapon in the battle of political ideas. It’s called ‘dynamic scoring’. An approach to assessing the costs of budget measures, it uses the latest economic knowledge, state-of-the-art economic theory and the best available computer technology. What’s more, it tends to show that tax cuts aren’t as costly to the public purse as conventional assessments suggest. So what’s not to like?
Sadly, George Osborne’s new toy risks undermining an emerging political consensus about the appropriate limits of politics.
Some questions are legitimately political – questions of values, about which there is no ‘right’ answer. Others are, at least in principle, matters of fact. But political debate too often dominates the latter. Values start to shape ‘facts’ which, freshly minted, are deployed to buttress arguments in favour of the values that spawned them.
When it comes to assessing something as important as the economic impact of different tax and spending policies, there should be no doubt that it’s best to leave the politics out. For example, many on the Left struggle to accept that returning to a top income tax rate of 50p might conceivably reduce the amount paid by the well-heeled. Small state devotees, on the other hand, have tended to argue that spending cuts in pursuit of deficit reduction boost economic growth. Neither of these views is sustained by the bulk of evidence.
So establishing the Office for Budget Responsibility has been one of this government’s most important – and hopefully enduring – institutional reforms. Part of the watchdog’s role is to keep the Treasury honest when it comes to assessing its own tax and spending plans. As an effort to depoliticise questions that are not legitimately political we should all welcome the OBR as evidence of progress. Its immediate acceptance by all the main parties is an encouraging sign of its likely durability.
More’s the pity, then, that the chancellor has recently been talking up a series of analytical papers the Treasury is writing to showcase its new dynamic scoring approach to determining the budgetary cost of tax cuts. The first two papers have looked at the costs of cutting fuel duty and corporation tax – two of the government’s big giveaway policies.
Under conventional scoring, the assessment of tax cuts must take account of how behavioural changes among those affected will alter the cost of the change. So if duty on fuel is cut, the anticipated revenue loss must reflect the fact that people will drive more because fuel is cheaper. That reduces the cost of the cut compared to a situation where drivers’ behaviour is assumed not to change.
Dynamic scoring takes the idea of accounting for behavioural responses a stage further, and then some. A complicated computer simulation of the entire economy seeks to trace all the subsequent behavioural changes people might make in response to the duty cut. Paying less for fuel, drivers now spend more on other goods, increasing VAT receipts, for example. And the resulting higher demand raises investment and employment across the economy, boosting corporation and income tax receipts. A chain reaction of these and thousands of other such relationships brings the entire economy to rest in a new place, with a substantial chunk of the lost fuel duty revenue offset.
If all this sounds devilishly clever that’s because it is. Far too clever, in fact. Two big problems lurk behind the endless reams of computer programming and mathematical equations that are the stuff of dynamic scoring.
First, to trace the myriad effects of a tax cut, such models demand that you input mathematical answers to, as the Institue for Fiscal Studies put it, ‘virtually every question, theoretical and empirical, that has ever been asked in economics’. One wonders why the omniscient economists answering these questions would need such a model in the first place. In reality they rely on a huge amount of guesswork. Politically biased judgments easily fill the evidence vacuum.
Second, while dynamic scoring might show the cost of the fuel duty cut to be less than previously thought, at some point the remaining shortfall still has to be made up if the government’s budget is to be sustainable. Raising another tax to balance the books, the same logic of dynamic scoring works in reverse: you need to increase the headline rate substantially more than conventional scoring might suggest, once all the damaging reductions in economic activity have been factored into the dynamic revenue estimate.
The Treasury’s dynamic assessments sidestep this problem by assuming that this alternative tax is one that has no deleterious economic effects. In effect, the results depend on government imposing an economically magical but practically impossible type of tax. Unsurprisingly, no tax of this sort has been proposed. In the real world, tax rises that have been imposed – such as the 2.5% VAT rise in 2011 – are far from economically harmless. Dynamic modelling of a fuel duty cut funded by a VAT hike wouldn’t give the government much to shout about.
Its capacity to cast tax cuts in a good light means that dynamic scoring has been promoted vociferously by politicians of the Right in the United States since the days of Reaganomics. And the cause is increasingly being taken up by British Conservatives. But the principles behind it and the risks of politicisation it entails are not the preserve of tax cutters. How much impact additional government spending on education could have on boosting the productivity of the UK workforce, for example, is hugely uncertain. Mix some politically motivated assumptions with an economically harmless source of taxation to fund much higher education spending, and big government enthusiasts can just as easily bend such reasoning to their own political ends.
Attractive in principle, dynamic scoring models can only ever echo the complicated and often shaky assumptions their masters feed into them. Unavoidably vulnerable to bias, but shielded from external scrutiny by its complexity, the approach threatens to politicise economic evidence. That would be the path to a diminished democratic system, and it should be resisted.
Ian Mulheirn, a consultant with Oxford Economics, is writing in a personal capacity
