GDP growth estimates and recession probabilities
With the decision to leave the EU, there has been much discussion about whether the UK will enter recession. This blog provides some information of the frequency of recessions in recent UK economic history, and puts into context the probability of recession in the aftermath of the referendum.
The Office for National Statistics (ONS) has today released its second estimate of second quarter GDP providing more information on the performance of the economy prior to the referendum. To see the first evidence of the post-referendum performance we will have to wait until the preliminary estimate for the third quarter GDP to be published on 27 October 2016.
The ONS' second estimate puts GDP growth at 0.6 per cent, unchanged from the preliminary estimate published at the end of July. It suggests that economic growth was robust pre-referendum. However, NIESR's monthly GDP series shows that most of the growth is due to a strong performance in April, with the economy falling back somewhat in the subsequent two months.
Data since the referendum are still sparse and largely of 'soft' evidence such as surveys of business sentiment and consumer confidence. On balance, we think the data indicates a deterioration in economic performance, but it will be some time before the 'hard' or official data are published. We expect the economy to slow to around 1 per cent per annum growth in 2017, from 1.7 per cent this year. However, we estimate the probability of a recession (at least two consecutive quarters of GDP declines) occurring sometime before the end of 2017 to be around 50 per cent.
To put this into context, figure 1 shows quarterly real GDP growth since 1955. In this time period there have been seven periods that fit this definition of recession, highlighted with grey bars, with an average length of just over ten months. The economy has been in recession for one tenth of the time.
By contrast, GDP has contracted in 44 out of 245 quarters, or 18 per cent of the time. There are many examples of a single quarter of contraction. These can often be explained by one-off factors such as poor weather conditions, or in the case of the final quarter of 2012, a return to ‘normality’ following a period where GDP growth was boosted significantly by spending associated with the Olympic Games.
However, the probability of a second consecutive negative quarter of negative growth, conditional on a negative quarter of growth, is 26 per cent. In other words, if there is a negative quarter of growth then there is a one in four chance of recession. This puts the NIESR view of a evens-odds bet of a technical recession into context. It would be consistent with a slightly shorter odds than a one-in-ten event. And even if there is a contraction in GDP this quarter, our call is against historical patterns.
So far the ‘soft’ data suggest a deterioration in performance and our modal forecast is for a decline in GDP in the third quarter of 2016. The signal we are able to extract from data releases in the coming months will guide our adjustments to this modal forecast so that we are rationally responding to new information, minimising the risk of introducing bias into our forecast process. Those who make near-certain forecasts after such an unprecedented event would do well to consider the evidence first.
 Monthly data can be volatile and so should be treated with caution. Nonetheless, the within-quarter profile for GDP is of particular interest given the economic and political uncertainty surrounding the referendum.
 The ONS’ preliminary estimate is based on only 44 per cent data. The remainder of the estimate is derived from a nowcast by the ONS.