I have covered the forecasting abilities of Fed and Riksbank. It is at best modest. Riksbank has improved its forecasting ability but still a long way to go. I have always doubted the ability of central banks/experts to forecast financial variables. But to see very similar performance for even macroeconomic variables is surprising (though not any more).
It is difficult to do similar analysis for India. However, we can do an analysis for the various organisations that do forecasting for growth levels in India. Before its quarterly monetary policy review, RBI releases a report on Macroeconomic and Monetary developments.
In the report there is a chapter on real economy (see towards the end) which publishes forecasts given by various agencies for the ongoing financial year. The forecasts are given for only expected growth in Overall economy and sector-wise growth (agriculture, industry and services). RBI has been releasing these estimates since July 2005-06.
I just did some analysis and here is how the numbers look:
| Overall | Agriculture | Industry | Services | |
| 2005-06 | ||||
| July | 7.0 | 3.3 | 7.4 | 8.3 |
| Oct | 7.0 | 3.1 | 7.5 | 8.1 |
| Jan | 7.2 | 3.3 | 7.9 | 8.6 |
| Actual | 9.2 | 6 | 8 | 10.3 |
| 2006-07 | ||||
| Apr | 7.6 | 2.7 | 9.1 | 9.0 |
| July | 7.7 | 2.5 | 9.0 | 9.3 |
| Oct | 8.0 | 2.1 | 9.7 | 9.5 |
| Jan | 8.3 | 2.3 | 9.5 | 10.1 |
| Actual | 9.6 | 3.5 | 10.7 | 11.8 |
| 2007-08 | ||||
| Apr | 8.3 | 2.9 | 9.2 | 10.1 |
| July | 8.5 | 2.8 | 9.4 | 10.4 |
| Oct | 8.7 | 3.3 | 9.4 | 10.4 |
| Jan | 8.8 | 3.4 | 9.5 | 10.4 |
| Projected by CSO | 8.7 | 2.6 | 8.6 | 10.6 |
The actual numbers have been picked from CSO. The agencies have underestimated the growth potential right along. The average estimates have been rising every quarter. Infact in Jan - 2008 the average estimates are higher than what CSO has projected. In sector-wise analysis the trend is similar. If we look at median or mode, the analysis is very much the same.
On checking the standard deviation it is around 0.3-0.5 for overall growth indicating there are diverse views amidst agencies. There is higher standard deviation in some quarters in sector-wise forecasts.
As far as number of agencies are concerned involved in forecasting, it has improved. In July 2005 there were 8 agencies and in Jan 2008, there are 15 agencies. Though, mostly agencies prefer to give overall growth forecasts but skip sector-wise forecasts. As a result, we have more number of data points for overall growth numbers.
August 8, 2008 at 2:08 pm
[...] from anchoring bias and rely much more on the past trend than new development. their Infact, I have shown earlier that forecasters in India do not revise their forecasts even amidst high growth and as a result [...]