RDP 2002-05: Real-Time National Accounts Data 4. Conclusion

The difficulty of obtaining accurate estimates of quarterly national output within a few months is evident from the immense scale of the collection, analysis and reconciliation task involved in assembling these data. It is not surprising, therefore, that over the past three decades changes to output growth estimates on a scale significant to analysts have not been uncommon – although there is evidence that the degree of uncertainty surrounding initial output estimates may be lower now than in the past.^[20] Initial mismeasurement has also frequently been quite persistent, often largely remaining even several years after the period being measured.

These observations have an obvious cautionary implication for the operation of monetary policy. As Orphanides (2000) put it for the US case, they indicate ‘the profound importance of appreciating the information problem for successful policy design’. Clearly, it is vital to remember that our uncertainty about the economy derives not merely from the difficulty of forecasting the future paths of key variables such as aggregate output, but also of knowing where those variables are now and where they have been in the recent past.

Footnote

This would be consistent with the significant improvements which have been incorporated in the compilation of the national accounts over the past decade. Most notably, these have included the switch to chain-linking in 1998 and the introduction of benchmarking of the accounts to annual supply-use tables. [20]