Sunday, 24 March 2013

A Socratic dialogue with @ToryTreasury

Yesterday I spotted a tweet from @ToryTreasury, who describes him/herself as "Official CCHQ voice for all things Treasury".
Ed Miliband's spending plans he repeated today would mean £33bn more borrowing this year - driving the deficit back up to double-digits.
I tweeted back
Out of curiosity, what multiplier are you using?
I wasn't looking for an argument. I certainly wasn't endorsing (or opposing) Labour's "five point plan for jobs and growth"; in particular, as I've said numerous times (eg here), I'm not at all enthusiastic about a VAT cut in current circumstances. But NIESR's views, or my own, weren't the point at all; I was genuinely curious about ToryTreasury's methodology.

Now, in order to assess the impact of a so-called "plan for jobs and growth", it is obviously necessary to assess the impact of tax and spending changes, how much it boosts "jobs and growth", what that does to tax and spending, and hence the impact on borrowing. That is exactly what the Office of Budgetary Responsibility (OBR) does with the government's tax and spending plans, as it has just done in the Budget.   The OBR has not done the same for Labour's plan; nor has NIESR. But apparently CCHQ and/or the Treasury had; I assumed that they had used some version of the OBR methodology, albeit presumably in a cruder fashion.

The first step in such an assessment, as the OBR makes clear, is the "multiplier"; what's the impact on growth of tax and spending changes. The OBR commendably sets out its assumptions on multipliers and hence the impact of tax and spending changes on growth. For example, it has stated (paras 3.22 and 3.23) that its best assessment, recognising the uncertainties, of the impact of the government's fiscal consolidation plan - tax rises and spending cuts - is that it has reduced GDP by about 2.4% up to and including 2012-13. It would be easy enough to apply a similar approach to Labour's tax changes if, as ToryTreasury clearly had, you make assumptions about the cost and timing.

ToryTreasury seemed rather reluctant to answer the question, but eventually the reply was: 
@jdportes OBR multipliers. They don't see any evidence to justify the higher multipliers you assume
Well, that's a perfectly good response. It made absolutely clear, I thought, that ToryTreasury had indeed done exactly the exercise I outlined above. And while I have in the past argued (along with the IMF and many others) that I think the OBR's multipliers are underestimated, I wouldn't expect CCHQ or the Treasury to agree.

Using OBR multipliers, of course, implies that "Labour's plans" (for tax cuts and spending increases, or at least slower spending cuts, relative to the government's plans) would increase GDP (especially any increased investment, for which the OBR assumes a multiplier of 1). So I asked how much, and this is where it gets interesting:
Jonathan Portes ‏@jdportes  You are using OBR multipliers! Excellent. So what's the increase in GDP that results from the £33 bn?
Tory Treasury @ToryTreasury ‏ @GDP would plummet due to a spike in interest rates and a loss of confidence in the UK, as peter mandelson warned this week
Jonathan Portes ‏@jdportes so you're not using OBR multipliers! But you just said you were. Did you apply OBR multipliers or not?

Tory Treasury ‏@ToryTreasury you really are a fantasist if you think £33bn more discretionary borrowing this year would not lead to a loss of market confidence

Jonathan Portes ‏@jdportes We're not talking about me! Positive multiplier on the £33 bn (as OBR), negative, or zero. Just answer..

Tory Treasury ‏@ToryTreasury  yr simplistic approach to multipliers ignores reality of market risk + loss of confidence - why yr recommended policy so dangerous

Jonathan Portes ‏@jdportes  You're back to talking about me! Simply asked whether your explicit statement that you used OBR multipliers was true or not?
I have omitted some of my other duplicative tweets, which simply firmly but politely repeated requests for an answer to my question; was ToryTreasury's tweet saying that they had used OBR multipliers true or not? ToryTreasury was neither prepared to reaffirm the original tweet, nor to withdraw or correct it, although their subsequent tweets make it clear that it was not in fact correct. 

So, having established that,contrary to his/her original tweet, ToryTreasury had completely ignored the OBR multipliers and methodology in their calculation, and was adopting a different approach, I tried again to establish what methodology they had used:
Jonathan Portes ‏@jdportes  There's nothing wrong with disagreeing with OBR - as you say, I do too -- but you should tell us if you do and explain why
Tory Treasury ‏@ToryTreasury as OBR say multipliers work OK for small changes in fiscal policy but can't account for market risk - clearly too complex fr you
Another error; it is wrong to say that the OBR thinks multipliers work only for "small changes".  As I say above, the OBR uses its multipliers to estimate the impact of the government's entire fiscal consolidation programme: a much larger change in fiscal policy than any conceivable interpretation of Labour's plan.  It seems highly probable that if the OBR were asked to assess the impact of that plan, they would start with their published multipliers, albeit recognising the uncertainties.  

But it's worse than that.  Without a multiplier assumption, ToryTreasury's original tweet is not just wrong, but simply meaningless, because it is not possible to calculate the impact of "Labour's plans" on the deficit.  ToryTreasury's actual argument is "any discretionary borrowing would be catastrophic because of [the credit rating agencies/Greece/bond vigilantes/the confidence fairy]."  Now I pointed out the flaws in this argument two years ago; and the empirical evidence since then has overwhelmingly confirmed my view, as the IMF, for example, says. But it's a legitimate, if faith-based, perspective. What is not legitimate is to claim that it is a basis for quantifying the impact of discretionary borrowing on the deficit.

In the end, the most succinct and accurate summary of ToryTreasury's position came not from me but from @RedEaredRabbit
@RedEaredRabbit We’re using OBR multipliers but we’ve multiplied them by -1 and it is inappropriate to use multipliers for this anyway.
Where does this leave us?  As I said above, I am not endorsing Labour's specific policy proposals; nor do I wish to rehearse yet again here the broader case for an alternative fiscal policy.  It is perfectly legitimate for ToryTreasury to use whatever assumptions it wants to assess the impact of Labour's plans on growth, borrowing and so on.  But making a quantitative statement without any underlying quantitative assumptions; then making incorrect statements about those assumptions (or their absence) and then refusing to correct such statements, means that any such assessment has little credibility.  

[Updated.  I have been asked to acknowledge that it's not inconsistent for ToryTreasury to argue, as they have done on twitter, that the positive direct impact on GDP, and hence on the deficit, is exactly offset by the negative impact of market reaction.  This is true.  There's nothing logically inconsistent about that at all, although it would in my view be a remarkable and economically implausible coincidence. However, if ToryTreasury is claiming that the £33 billion number in the original tweet was based on such an analysis, then they should be able to show their working - that is, explain, in quantitative terms, what both the direct and indirect impact on GDP, interest rates, etc, is, and why that translates to a net impact of £33 billion.  If they wish to do so, I'm happy to post it here.]. 

No comments:

Post a Comment

Note: only a member of this blog may post a comment.