GDP forecasts: Figures of speech

The IMF’s headline growth and forecast figures probably matter far less than they appear to. So why continue to make such a fuss about them? On April 14, Pierre-Olivier Gourinchas, the IMF’s chief economist, opened a press conference in Washington by announcing that global growth for 2026 would come in at 3.1 per cent, down from the 3.4 per cent the fund had forecast three months earlier. He blamed the ongoing war in the Middle East, the closure of the Strait of Hormuz and an energy-price shock that has lifted headline inflation back towards 4.4 per cent. The bigger country-level downgrades were in the Gulf. Iran’s economy is now forecast to shrink by 6.1 per cent, a revision of 7.2 points, and Saudi Arabia’s was cut from 4.5 to 3.1 per cent. Less exposed places fared better: the eurozone nudged down by 0.2 points to 1.1 per cent, America to 2.3 per cent. Coverage of the forecast the next morning was everywhere. Finance ministries from Rabat to Astana issued statements, television news led with the global figure and newspapers chewed over the ‘three scenarios’ (reference, adverse and severe) the fund now presents instead of a single baseline. Bond markets, which had been watching tanker traffic rather than IMF statistical appendices for several weeks, barely moved: emerging-market spreads were little changed, rates held and even the dollar was quieter than for most of March. Most of what the report contained was already in the market. Consensus Economics , which aggregates monthly views from more than 1,000 professional forecasters, had registered much of the downward revision by the second week of February. Central banks have their own models, as do most finance ministries, and teams at investment banks have been revising their 2026 numbers ever since the Red Sea and the Persian Gulf became expensive (or impossible) places to ship oil through. By the time the fund’s headline figure reached the wires, nobody charging real money to trade it was surprised. Of course, different readers want different things from the report. Bond managers treat its numbers as a consensus check and mostly have their own projections in place already. For finance ministers in developing economies, particularly smaller ones, the figures carry an institutional authority local research does not, and can be slotted into whichever argument is currently live in the cabinet room. Editors, meanwhile, need a headline, which the IMF reliably produces. A matter of ceremony The IMF’s forecast record is not flattering. A 2021 staff working paper by the fund’s own researchers found growth systematically over-predicted during downturns, with biases persisting across horizons. For several emerging market groups IMF projections are no better than a naïve model that assumes next year will look like this year. A 2019 Bloomberg analysis of 28 years of IMF projections found the fund within 0.1 points of actual growth only six per cent of the time, the average miss two percentage points. Gourinchas knows this, which is partly why the April report leads with three scenarios rather than a single baseline. If the single number is unreliable, three numbers at least do not pretend otherwise. In practice, the figures become points for negotiation. A minister pushing fiscal tightening will quote the fund’s growth trajectory. Her opposition flips to the downside scenarios instead. A rating analyst going into an emerging market sovereign review can nod at its inflation path without defending her own assumptions from first principles. Investment-promotion agencies put the figures on slides, gaining a multilateral imprimatur no local institution can match. The effect is felt most in the region this platform covers. On April 14, the same day the global forecast landed in Washington, an IMF mission in Tashkent raised its 2026 projection for Uzbekistan from 6.2 to 6.8 per cent, citing buoyant remittances, elevated gold prices and steady reform progress. The number has already done the rounds of domestic media and will almost certainly turn up in the next batch of sovereign-bond road-shows, where ‘IMF-confirmed 6.8 per cent’ carries a weight no local institution can replicate. The Eurasian Development Bank, incidentally, reached the same 6.8 per cent figure back in December . Markets had priced it. Beyond the headline Even if the forecasts work well as consensus markers, the broader obsession with GDP point-estimates is harder to defend. Simon Kuznets, who built the first modern national accounts for America in the 1930s, warned explicitly in 1934 that his aggregate should not be confused with welfare. Joseph Stiglitz’s 2009 commission, Mismeasuring Our Lives, said the same thing more loudly. Indeed, it might even be a good idea to ditch GDP as a measurement entirely . Again, the effect is especially distorting in the economies of emerging Europe, the Caucasus and Central Asia. Kazakhstan’s forecast 5.5 per cent growth in 2026 rests heavily on a handful of hydrocarbon fields with price trajectories unrelated to domestic policy. Armenia’s 12.6 per cent expansion in 2022 was largely the mechanical result of Russian nationals and capital relocating after the invasion of Ukraine; when the flow reversed so did the headline, though the underlying economy had changed less than the number implied. Georgia’s recent growth has leaned on tourism, transit and construction, and the country’s current-account deficit is flagged in every one of the IMF’s Article IV consultations. A country on five per cent headline growth can be doing several things at once, and the investment promotion agencies of the region know which of those things matter to investors. Much of the interesting work sits there, and the fund itself does most of its best thinking there too. The regional reports and Article IV papers receive a fraction of the attention the headline numbers do, and in aggregate they are much better guides to what is actually happening in any particular place. They are also where the fund’s economists tend to be most candid. The communications office, understandably, prefers a round-ish single number that fits a headline. Whether anyone outside a finance ministry pays attention to that headline in the way the fund pretends they do, on the evidence of April 14, is open to question. Photo: Dreamstime .