Gross Domestic Product, or GDP, is a measure of economic activity which captures how much the economy produces during a given period. It can be measured using an output, income or expenditure approach.
[GRAPH MOVEMENT – latest estimates of GDP growth: % growth on 12 months previously]
This chart shows the latest available estimates of the percentage growth rate of the economy compared with a year earlier, for each quarter since 1961. The higher the percentages, the faster the economy was growing relative to a year earlier. Negative numbers represent times when the economy was contracting.
These estimates are made with the advantage of having fairly full information available, at least for all but the most recent years. However, in order to generate more timely estimates, the Office for National Statistics publishes a first estimate of GDP growth just 25 days after the end of the quarter, based on available data. These estimates may be revised as more and better data becomes available with a time lag, or as methods improve, leading to more reliable, accurate estimates.
[GRAPH MOVEMENT – M+1 Estimates]
So how sharp is the trade-off between timeliness and reliability? One way to look at this is to see how successive revisions change the picture of how the economy has been doing.
This chart shows the first estimates of quarterly GDP growth – the M+1 or month 1 estimates – for each quarter between 1960 and the present. Now let’s see how the picture changes when the M+3 estimates become available – that is, the estimates published three months after the end of each quarter.
[GRAPH MOVEMENT – M+1 and M+3 Estimates]
These estimates are a little different, but not so far apart as to alter the interpretation of how the economy was doing over this period to any great extent.
So now let’s add in the later estimates for each quarter based on the further information that became available.
[GRAPH MOVEMENTS – M+6, M+ 12 and M+ 24 estimates are added successively]
And now the later estimates still, reflecting not only fuller information but also the effects of changes in the methodology by which GDP is produced.
[GRAPH MOVEMENTS – M+36, M+48 and M+60 are also added successively]
By the end of five years, most of the revisions that are going to be made have been made. So looking at all the estimates up to five years after the quarter to which they relates gives a pretty full picture.
What can we conclude? The lines are not identical, reflecting the fact that the quarterly growth rates were revised over time. In a handful of instances, the revisions make a difference to how the economy would have been interpreted.
[GRAPH MOVEMENT – Zoom in on 1988]
For example, the first estimates for GDP in 1988 showed a much weaker economy than later estimates revealed to be the case. This led to a review of the way GDP estimates were compiled and substantial changes to the methodology were implemented thereafter.
[GRAPH MOVEMENT – Zoom in on 1998-99]
There again, the early estimates also underestimated the strength of the economy in 1998 and 1999, at the time of the so-called South East Asia crisis.
There is also a suggestion that the extent of the revisions may be a little greater around turning points in the economy - when it is about to recover or to enter into a deceleration.
[GRAPH MOVEMENT – Zoom out and circle key years of change]
Estimating GDP is commonly agreed to be more challenging at times when the economy is changing direction.
[GRAPH MOVEMENT – Circles disappear]
Overall, however, the chart suggests that the early estimates supported an interpretation of the economy’s behaviour which was not very different from that which is implied by the latest estimates.
In turn, this corroborates the value of the early estimates. Because they cannot be based on full information, they are necessarily less reliable than the latest estimates. But, in general, they do appear to be good predictors of the eventual outcome.