The IMF versus ‘The French’

Unlock the Editor’s Digest for free

Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.

Update: The IMF has clapped back, scroll down for their response.

The headline is clickbait, sorry. Panmure Gordon’s Simon French has a snappy drive-by on the IMF’s UK forecasting approach today, suggesting it stands for “Increasingly Mediocre Forecasts” (miaow!).

He writes:

We have defended the economic forecasting record of the International Monetary Fund (IMF) in recent times. Accusations of UK bias are unsubstantiated and economic forecasting is hard work to get right. Getting things wrong is an occupational hazard amidst a range of unpredictable shocks, and we have made our fair share of mistakes in recent times.

However, the latest IMF economic forecast for the UK economy is full of holes —three most obviously. These call into question the validity of their latest economic assessment. The UK economy has its fair share of challenges right now, but the opinion-shaping IMF are not providing much sober analysis in their latest World Economic Outlook. The forecast looks drunk.

Your can read mainFT’s write-up of the latest World Economic Outlook here, or, if you’re a kind of sicko, read the actual thing here.

French has three main boeufs with the IMF:

1) The IMF didn’t make use of the Office for National Statistics’ cOnTrOvErSiAl new GDP data (released on September 1st), despite its latest forecasts closing on September 25th:

Whilst it is questionable whether it would materially change the forward growth path for the UK economy, this suggests that the conditional parameters for recent productivity growth, service sector activity, pandemic scarring and economic momentum are understating the UK’s economic position relative to other major economies. This revision changed the UK economic narrative overnight, so its omission from the IMF’s assessment is a serious deficiency.

2) The IMF assumed a peak Bank Rate of 6 per cent, despite markets pricing in a 5.35 per cent top when forecasts closed. French acknowledges that estimates for the peak have shifted a lot since the last WEO, but says the lack of explanation for the change is troubling.

3) The deviation between the IMF’s inflation forecasts and those of the relevant central bank is markedly bigger for the UK than other countries:

It might seemed pernickety, but as French notes, it’s the IMF’s World Economic Outlook — people do actually read it (or read about it). He concludes:

This has a blowback for investment decisions and the cost of capital. Unfortunately, when it comes to the IMF’s latest assessment of the UK economy it has clear and significant deficiencies. Of course, the IMF may yet be proved correct — but their methodology has not given them the best chance.

Update 12/10/23: The IMF has responded, claiming French made two factual errors in his analysis and (after we initially went to pixel) clarifying their peak rates calculation.

Regarding the CPI forecast spread, they said:

Factual Error 1: The [figure]—which underlies Mr. French’s statement that “The deviation between the IMF’s inflation forecasts and those of the relevant central bank is markedly bigger for the UK than other countries”—has a basic data error. It compares our forecast for average inflation in 2024 (at 3.7 percent) with the BoE’s Q4 forecast (at 2.5 percent). A like-for-like comparison between Q4 inflation forecasts would show only a 0.1 p.p. difference (with our Q4 forecast at 2.6 percent). It would therefore appear that his assertion is driven by an incorrect handling of data, rather than anything to do with our comparative forecasts.

This does seem to be an error on French’s part. Here are the relevant tables:

On the use of older GDP data, the IMF says (our emphasis):

Factual Error 2: While the ONS indeed pre-announced its historical data revisions and provided an indicative headline GDP series on September 1, the revised national accounts data were only published on September 29, after our September 25 cut-off date for the forecast. Importantly, official estimates for quarterly GDP components and revised estimates for 2023Q1 and Q2 were only included in the September 29 data release (as also acknowledged by the ONS “The next Quarterly National Accounts release on 29 September 2023 will incorporate these revisions into our official estimates in line with our National Accounts Revision Policy”). We faced a trade-off between imputing our own estimates for historical GDP components just a few weeks ahead of the official release on September 29, or continuing to use pre-revision GDP data. We opted for the second approach (and mentioned it clearly in the WEO country notes) in order to avoid adding to confusion by publishing what might then have become a third set of historical GDP numbers for the UK.

We don’t agree that this is a factual error on French’s part — clearly this represents a difference of opinion on approach.

The IMF added (our emphasis):

While the ONS revisions are not part of our forecast, we have conducted a preliminary analysis thereof. The revisions do have a major impact on the level of UK GDP, which is higher by about 2 percent. However, we do not see the new data as having a significant impact on our near-term growth forecast. Thus, the notion that incorporating the revisions would have significantly changed our growth forecast is misleading.

This is interesting (and, as far as we can see, new) information!

Finally, on the rate path, we have this clarification offered by the the IMF’s chief economist (after we went to pixel):

Again, interesting and new information! The WEO simply said:

The short-term interest rate path is based on market interest rate expectations.