1 • 2011
As China’s importance in the world economy has increased, so too has the level of interest in forecasting its economic performance. Consensus Economics, which specialises in collating economic forecasts, currently reports on forecasts of China’s key economic indicators produced by nearly 20 different international research organisations. At the same time, however, doubts have been raised about whether China’s statistical data are able to give a realistic picture of the huge country’s economic situation. For example, a revision of China’s national accounts in 2005 resulted in a full 16.8% upward adjustment in GDP, and some smaller adjustments have been made since then. Such a situation confronts both users of the data and forecasters with hard choices as they face the task of forecasting future developments.
The Bank of Finland Institute for Economies in Transition (BOFIT) publishes its own forecast for China twice a year and has in progress several research projects that deal with forecasts of the Chinese economy as well as forecasting in a slightly broader context. The studies involved will be reported soon in the BOFIT Discussion Paper series. These include a study by Aaron Mehrotra (BOFIT) and Jenni Pääkkönen (Government Institute for Economic Research, VATT) that compares China’s published GDP statistics with several other macroeconomic indicators. The purpose is to examine whether the economic dynamics portrayed by the official GDP statistics accord with other types of coincident indicators. The alternative indicators are coalesced by using factor analysis to combine data on a broad array of macroeconomic variables. The estimated factors are then compared with published GDP growth figures for 1997–2009, a period that included the Asian economic crisis of the late 1990s, the global economic crisis of 2008 and 2009, as well as China’s own robust growth performance, especially in the second half of the last decade.
The authors found that the published GDP growth figures are generally in line with the alternative indicators and that the discrepancies tend to be very short-lived. This suggests that there are no systematic errors in China’s GDP growth figures. However, the significance of the different factors changes over time, which is probably a sign of ongoing structural change in the Chinese economy. The study also compares the picture of Chinese economic dynamics given by the factor analysis with those of two published coincident indicators - China’s own official one and that of the US Conference Board. It became clear that the three depictions of economic conditions in China are very closely aligned.
In another study, Iikka Korhonen (BOFIT) and Maria Ritola (Demos Helsinki) examine how forecasts of GDP and inflation are carried out in different research organizations in the large emerging economies. It appears that in many countries the so-called consensus forecast (average of all forecasts) for GDP growth is unbiased, ie the average forecast error is zero. But this is not the case for China, where GDP growth forecasts have been notably overly cautious during the peak years. The forecast error - outcome versus consensus forecast of March of the same year - is about two percentage points for the last decade or so, but clearly larger for the years 2002–2007. The forecasters were unable to foresee China’s highly robust economic growth during that period. On the other hand, the forecasts were better on average during the international economic crisis (see chart).
Accuracy of forecasts of China’s GDP growth
Sources: Consensus Economics and Chinese authorities
According to the results for the indicator models, China’s national account figures accurately relate economic developments. For this reason, the systematic forecast errors cannot be explained by statistical problems, at least not for the years 2006 and 2007, when the errors were the largest. It may be the case that the forecasters placed too much trust in the possibilities of economic policy for restraining economic growth and avoiding economic overheating. Correspondingly, the inflation forecasts for 2007 and 2008 were below the outcomes.
Iikka Korhonen