NIESR’s forecast: Matthew Parris’ astonishingly inaccurate critique

UPDATED 15/9/2013

The Times has now published the following correction (see bottom of column here):

Matthew Parris’s article on economic forecasts (Sept 7) stated that the National Institute of Economic and Social Research predicted, in February 2012, that “output was on course to contract by 0.5 per cent (a triple dip recession)”. In fact, NIESR’s February 2012 forecast was for output to fall by 0.1 per cent in 2012 and to grow by 2.3 per cent in 2013. NIESR’s August 2012 forecast predicted output would fall by 0.5 per cent in 2012, but grow by 1.3 per cent in 2013. In neither forecast did NIESR predict a “triple dip recession”

It is unfortunate that I was obliged to complain to the Press Complaints Commission to secure what should have been a straightforward admission of error from the Times. And it reflects particularly badly on Mr Parris that he chose to respond to a polite request by digging himself further into a hole. But the text of the correction is entirely satisfactory, and I consider the matter closed.

Matthew Parris had an interesting article about economists and economic forecasting in last week’s Times.  His main substantive point – that we’re not very good at forecasting turning points – has considerable force, although it’s hardly original. And, as I say here, it’s perfectly reasonable to ask forecasters, including NIESR, why we seem so far to have underpredicted the strength of the recovery.   But, in his references to me and to NIESR’s forecasts, he obviously hasn’t done his own homework, since he simply gets what we said wrong.. 

Mr Parris wrote:

 His NIESR think-tank said [a year ago] “deterioration in the UK economy has been more pronounced than even we expected”.  Mr Portes…had that February [2012] called on the Government to relax its plans for reducing the deficit, his think-tank warning that output was on course to contract by 0.5 per cent (a triple-dip recession). Unemployment, it said, would continue growing, to 8.9 per cent.

There’s not a single point in this paragraph that actually supports his argument, and most of it is just plain wrong. Let’s start with the first sentence.  Understanding it really shouldn’t be hard: the clue is in the (perfect, meaning “action already completed”) tense.  Here’s the full passage: 

The deterioration in the UK economy has been more pronounced than even we expected. We estimate that the ‘Jubilee effect’ reduced estimated growth in the second quarter by 0.4 percentage point, with a corresponding boost to be expected in the third quarter. This implies that underlying growth is still negative. But focusing on a ‘double dip’ distracts from the more important trend; the level of output has effectively been flat over the past two years.

So the sentence he quotes wasn’t, as he claims, an inaccurate forecast about the future; it was an accurate statement about the past. Either Mr Parris knew that, or he didn’t even bother reading what we said. Neither reflects well.

What about the numbers?  Well, here‘s the forecast we actually made in February 2012. It says:

The economy will contract very slightly (0.1 per cent) this year but grow by 2.3 per cent in 2013.

There is, needless to say, no mention at all of a triple-dip recession, since at that point we hadn’t had a double-dip recession!   The forecast also noted, as all our forecasts do, that the plausible uncertainty was much larger than that 0.1 percent.

I pointed out these errors to Mr Parris and the Times. His response was to add a note implying (although it’s still vague!) that he was referring to NIESR’s August 2012 forecast, from which the first sentence quoted above is indeed taken. That forecast says:

The economy will contract by 0.5 per cent this year, but grow by 1.3 per cent in 2013. Unemployment will peak at 8.6 per cent in 2013.

So that does have the 0.5% contraction, but for 2012 not 2013, although the unemployment number is different.  How did that forecast turn out? Well, partly because of data revisions, the GDP outturn for 2012 was somewhat, although not dramatically, better than that forecast (just over zero). But, with output actually shrinking in the last quarter of 2012, that wasn’t a particularly inaccurate forecast, and more importantly it is hardly relevant to the subject of the article, which is supposed to be economists’ failure to forecast the current recovery.  Nor, of course, did that forecast mention a triple dip recession either, since it still wasn’t an issue in August 2012, so this error remains uncorrected.    

So what did NIESR actually forecast about a triple dip recession later, when it did look like one was a theoretical possibility ?  Well, in February 2013, when this debate was at its height we said this:

Overall, the concern should not be whether or not the economy shrank slightly at the start of 2013 to fulfil the ‘technical’ definition of recession, or whether (as we expect) there is slight growth.

So rather than predicting a triple dip recession, we predicted the opposite. Moreover, we pointed out that even if there was one technically, it would be more or less irrelevant (and we made this clear even more explicitly here).

I assumed Mr Parris’ original errors were the result of sloppy research. But the cursory, and equally incorrect “correction”, even after the mistakes were pointed out, suggests that he’s now either deliberately trying to mislead his readers; or, probably more likely, too embarrassed to admit he got it wrong. Either way, whatever you think about economic forecasting, it’s terrible journalism.  I have complained to the Press Complaints Commission, and I expect a proper correction in due course.