MACROSCAN. Why we should not trust World Bank forecasts

According to the World Bank, developing economies are now facing a “structural slowdown” that may well last for years.

In a recent report (‘The global economy in transition: Global Economic Prospects, June 2015) the Bank argues that since the possibilities of growth in major emerging markets are so diminished, they are likely to go from being drivers of global growth to becoming a drag on it.

As it happens, this reiterates the argument made in the previous edition of MacroScan that growth in developing countries is now less likely to be a buoyant contributor to the global economy, that will be forced to rely once again on the US.

The World Bank analysis is based on forecasts done by the World Bank staff, which predict that developing countries as a group will grow more slowly in this year but that growth will pick up again in the following year, 2016.

Forecasting is the economists’ equivalent of crystal-ball-gazing, with probably as much (if not lower) success in predicting the future.

However, the profession accords economic forecasting much greater legitimacy, on the grounds of being based on statistical models.

Yet, even a cursory examination of the actual track record of such forecasts should give rise to much greater scepticism.

Consider just the forecasts of the World Bank alone, for the time being. Given how uncertain the world economy is at present, it may be unfair to expect any great reliability of forecasts made several years in advance, since so much can change.

So let us give benefit of doubt by looking only at the most recent forecasts --- those made a year before the relevant year and those made in the middle of the relevant year, and see how correct the projections made by the World Bank have been in predicting the actual GDP growth according to its own subsequent estimates.

Chart 1 provides data on World Bank projections and actual rates of GDP growth for the high-income countries as a group, while Chart 2 provides data on developing countries and economies in transition taken together as a group.

Three things stand out immediately from these two charts. First, the World Bank’s own forecasts change dramatically over the course of the year, with significant changes in the forecasts for the current year compared to those made even one year previously.

Second, overall the forecasts are poor predictors of the actual rates of GDP growth, even when they were made in the middle of the relevant year when presumably much more data are available to feed the forecasting model.

Third, often, even the direction of change was completely missed, with significant deceleration of growth in years when acceleration was predicted, and vice versa.

Such variation in the direction of change has occurred across forecasts made in different time periods for the same year, as well as when compared to the actuals.

Indeed, the differences between forecasted and actual GDP growth are so large as to be embarrassing in several cases.

In 2008, when the global economy was plunged into massive crisis, the extent of the disruption was already quite apparent to many observers by 2007 and certainly by the middle of 2008.

Yet the June 2007 forecast for 2008 suggested no major deceleration of the global economy, and even as late as June 2008 the World Bank was actually predicting only a slight slowdown for the rich countries and an acceleration of growth for developing countries, compared to its own previous projections.

Needless to say, the eventual reality was rather different.

By the next year the Bank adjusted its forecasts to factor in the global crisis, but once again it seriously underestimated the impact of the crisis, as global growth turned out to be even lower.

But thereafter it did not correctly anticipate the rapid recovery of developing countries even as it had excessively high expectations of growth in the developed world.

In subsequent years, the volatility of its own forecasts, and the differences from the ultimate growth, remained significant.

Of course, forecasts for such large aggregations such as the world economy, or rich countries and developing countries as a group are necessarily difficult.

So let us consider the World Bank’s forecasts for some of the more prominent economies.

Table 1 provides the equivalent data (previous year forecast, current year forecast and actual GDP growth according to the World Bank’s own subsequent estimates) for selected countries, for the recent five year period from 2010 onwards.

Here too, similar results are evident. This was not a period marked by the extreme volatility of the Great Recession, and it should be easier to project GDP growth for major economies about which quite a lot is otherwise known.

So it could be expected that the forecasts would correspond more to the eventual GDP growth that the World Bank itself has recorded for individual economies.

But this does not appear to be the case: while the differences between the consecutive forecasts and between them and the actual GDP growth are not as marked as for the previous period, they are nonetheless quite notable, and cast major doubt on the reliability of the World Banks’ forecasting ability.

There is no very clear pattern in terms of over- or under-estimation, but some instances are striking.

Thus in 2010 and 2011, the World Bank sharply underestimated economic expansion in Argentina even as it overestimated expansion in South Africa.

In other years, it has not been able to anticipate continued recession in the Eurozone or the sharp slowdown in Russia

We have considered only World Bank projections, but the same is just as true of IMF forecasts or OECD projections. Thus, one examination of IMF growth forecasts for India found that the margin of error varied from 10.5 per cent to as much as 46.8 per cent between 2006 and 2013.

One may ask, so what if these forecasts are not particularly accurate? The unfortunate point is that, whether inaccurate or not, these forecasts matter quite a bit, especially for developing countries, because they affect not only media coverage but more crucially that complicated thing summarised as “investor expectations”, as well as, quite often, the decisions of credit rating agencies.

And very often, by influencing expectations, they can end up having some impact on eventual outcomes, even if they have got it wrong to begin with.

Since the World Bank — and indeed the other agencies — do not reveal the details of their forecasting models or the assumptions they make, it is hard to critique them other than in terms of their results.

But it is possible that one reason why they tend to be so off the mark is because they have partial or even wrong assessments of the mechanisms and factors that operate within national economies and globally.

They argued for “delinked” emerging markets that could serve as alternative growth poles even when it was evident that rich economies’ markets remained crucial drivers for developing country exports.

Now they argue that developing countries will drag down global growth, even as it is clear that reduced demand from Northern markets has led to reduced production in China, in turn affecting Chinese trade with other developing countries.

Whatever may be the reason for the wrong estimates peddled by the World Bank and other organisations, what is clear is that it may be counterproductive to put too much reliance on them.

COMMENT NOW