Policy Research in Macroeconomics

Assessing the manifestos – the IFS fails the test

Share on facebook
Share on twitter
Share on linkedin

The “gold standard” for analysis of the economic policies of political parties was set by The Netherlands Bureau for Economic Policy Analysis (CPB), established by Jan Tinbergen, Nobel prize winner and, more substantial, member of the anti-Nazi resistance during World War II.  For decades before each election the CPB would assess the manifestos of all parties (see analytical review of the CPB).  Using its highly-respected and publicly available statistical model, CPB researchers would estimate the probable impact of party policies on the nation’s economy.

On 26 May the UK Institute of Fiscal Studies held one of its by-invitation-only meetings to announce its evaluation of the taxation and expenditure policies of the Labour and Conservative Parties (Liberal-Democrats received a few comments, SNP, the Greens and UKIP did not make the cut).  A person present who also had once attended a similar CPB event would have experienced, to reverse the famous phrase of Tom Paine, transition from the sublime to the ridiculous.

In place of the rigorous modelling in the tradition of Tinbergen, the IFS presentations gave the audience an exercise in arithmetic that is not fit for the task (in the 15th minute of his discussion of revenue estimates, Robert Joyce referred to “best educated guess”).  With some of Europe’s best macroeconomists, the CPB would estimate the impact of party policies on growth, employment, wages and inflation.  Setting its bar impressively low, the IFS restricted itself to educated guesses about budget balancing with no apparent link to broader economic effects.

The procedure used by IFS is bean-counting.  In the case of the Labour Party, expenditure plans are summed, then compared to the manifesto’s tax plans.  The IFS director reached the apparently devastating conclusion that proposed changes would not come close to their projected target for business tax increases, which in any case would prove damaging to the economy (see report in The Guardian).

Narrow as this bean-counting is, the IFS get it wrong.  Adding and subtracting is not the appropriate way to project tax revenue or public expenditure.  Both how much the government spends and the revenue it collects are strongly affected by the economy’s growth, which, as should be obvious, is itself affected by expenditure and revenue.  For that reason that the CPB, the US Congressional Budget Office use an interactive analysis (“model”) to estimate revenue and fiscal balances, which the IFS did not.

It may seem sensible to add up spending commitments, then subtract the revenue proposals and inspect the balance of the two.  Far from sensible, it is nonsense, because if realized the plans of both parties will have a substantial effect on the growth of the economy.  Making “educated guesses” about interactive effects is the stuff of tea break conversations not expert assessments.

The basic flaw in the IFS analysis of party manifestos, a flaw found in much of its work, is the organization’s ideological commitment to the balanced budget dogma.  The IFS’s deputy director, Carl Emmerson, betrayed this approach to macroeconomic policy when he said that the Labour Party was “pretending that everything can be paid for”, a statement that comes close to the metaphysics of “living within our means”.

As anyone trained in public finance should be aware, there is no economic reason why a national government need ensure that “everything can be paid for”.  As the IMF demonstrates in its guide to “sound fiscal policy”, governments, especially governments with national currencies, can finance expenditure either by revenue or borrowing.  The IMF guidelines specifically advise against “capping deficits”, recommending instead flexible expenditure rules consistent with the Labour Party’s rules on fiscal policy.  

Deficits themselves are not a problem.  Whether they are depends on several variables, their size relative to GDP, whether they fund government consumption or investment, size of the public debt, and level of unemployment.  Attempts to assess fiscal outcomes, be they from Conservative or Labour policies, are useless in the absence of an analysis of the economy as a whole.  

The IFS’s partial and static approach to fiscal outcomes is repeated for other issues in the 26 May event.  The assertion that Labour’s plans to raise business taxes might have harmful side effects was a glaring example.  In his comments the IFS deputy director asserted that increases in corporate tax are passed directly to workers and consumers, a position unsubstantiated theoretically and empirically (repeated by Robert Joyce in the vague assertion, “all taxes are paid by people”).  This blanket assertion that corporations “pass on” taxes is repeated on the IFS webpage.  However, it contradicts professional opinion (as explained in Brookings Institution working paper).

The IFS credibility for sober analysis suffered an own-goal on 26 May during its non-rigorous and biased assessment of the party manifestos.  A defender might argue that criticism of both the Conservative and Labour made the event “even-handed”.  However, the central message was that expenditure exceeding revenue is irresponsible, an ideological position that parrots the propaganda of the Conservative Party.  

Rigorous professional assessment of the economic impact of the economic policies of political parties would make an important contribution to informing voters.  To do this a research group must 1) accurately and fully specify each party’s policies; 2) introduce these into a professionally peer-evaluated macroeconomic model appropriate for the purpose (there are many to choose among); 3) made that model and the calculations it generates public knowledge; and 4) report the result in a non-judgemental manner.

The IFS assessment may have achieved the first.  As for the second and third, if there is a peer-evaluated model one finds no reference to it or the calculations it produced during the 26 May event or on the website for the event.  And fourth, accusations of dishonesty do not normally accompany a professional and unbiased approach.  Unless of course, the purpose is to grab headlines, which it did (“IFS makes damning assessment“).

John Weeks is Emeritus Professor of Economics, SOAS, and co-convenor of the network Economists for Rational Economic Policies (EREP).

Share on facebook
Share on twitter
Share on linkedin

2 Responses

  1. So many things to agree with here. Taking on the incidence brigade. Taking on the IFS’ firmly microeconomic view, at the expense of all-important macro.

    The one concern is that interactive modelling is a supreme tool for propagandists and demagogues – you can create any model you want, tweak the inputs, to get just the output you want. I would argue that instead of arguing too much about fine-tuning the numbers, it’s more important to emphasise politics, fairness, and democracy more firmly.

  2. Nice article. My only quibble is that while John Weeks correctly criticises the idea, seemingly promoted by the IFS, that a balanced budget is desirable, he doesn’t tell us exactly what the "non balanced" budget (i.e. the deficit) OUGHT to be. His argument would be more convincing if he actually gave an answer to the latter question, and the answer is not difficult. It’s thus.

    On the assumption that there is no desperate need for stimulus (or “anti-stimulus), i.e. assuming an average sort of year, the deficit actually needs to be about 1.5% of GDP. I set out the reasons for that figure here:

    http://ralphanomics.blogspot.co.uk/2017/05/top-uk-treasury-official-does-not.html

    In particular, see para starting “Assume inflation..” and subsequent 5 paras.

    1.5% is actually half way between the two main parties’ projected deficits for 2020 and 2022. So by an entire fluke, the great and good at Westminster have got the answer about right!

Leave a Reply

Your email address will not be published.

This website collects cookies and analytic data. To use our website you must consent.
For more information read our Privacy Policy.