By Ana Galvão
Business cycle phases – expansions and recessions – are usually classified based on the historical path of macroeconomic variables, in particular real GDP. UK GDP is computed by the Office for National Statistics (ONS). Over the last 30 years, the first estimate of UK quarterly GDP has been published around 20-40 days after the end of the reference quarter. This first estimate is computed using a largely incomplete dataset as survey data for all months in the quarter can be incorporated only with a delay. As datasets are updated, ONS revises early estimates of real GDP, and later estimates are more accurate measures of GDP than earlier ones.
Business cycle phases are classified after turning points – peaks and troughs – are identified. A recession phase starts in the quarter after a peak and ends in the quarter of a trough. In our ESCoE Discussion Paper, we use a standard algorithm to classify UK business cycle turning points. The algorithm relies on simple rules. A first requirement in identifying a recession phase is the observation of two-quarters of negative growth. As the algorithm is applied to the latest release of UK real GDP (which was 2019Q1 at the time the research was done), we find that the last identified UK recession phase was between 2008Q2 and 2009Q2, a period associated with the Great Financial Crisis.
Readers of economic and political commentary may remember, however, the discussion in 2012 about the UK ‘double-dip’ recession. The discussion was supported by the evidence that the first releases of quarterly GDP growth for 2011Q4 and 2012Q1 were both negative. We provide evidence that if we use ONS’s real GDP data as it was published in 2012Q4, we find a peak in 2011Q3, leading to a recession phase starting in 2011Q4. As ONS recomputed real GDP values based on more updated datasets, the evidence of this recession phase has vanished.
In our Discussion Paper, we provide evidence that some economic indicators may have endorsed the recession phase from 2011Q4, while others may not. Indicators related to construction activity and retail orders were supporting an expansion phase in 2012, while consumer confidence would indicate a recession phase instead.
Our research suggests that one should be cautious in defining a peak using only early estimates of UK GDP growth. This is because initial GDP estimates are based on an incomplete dataset, as ONS requires time to complete all surveys, and these may be substantially revised as the dataset is updated. In evaluating business cycle phases in real time it is therefore important to gather information from alternative indicators of economic activity to observe whether they can support a similar picture to the initial GDP estimates.
ESCoE blogs are published to further debate. Any views expressed are solely those of the author(s) and so cannot be taken to represent those of the ESCoE, its partner institutions or the Office for National Statistics.