Today sees the publication of the latest estimates for UK economic output for the first quarter of 2013. These are the final estimates for what is called Gross Domestic Product and no change is expected to the previous quarterly growth of 0.3% and annual growth of 0.6%. However I wish to take the opportunity to continue my occasional series where I peer under the bonnet of these numbers to demonstrate that the situation is much more uncertain and unreliable than many would have you believe.
What is GDP used for?
There are a multitude of uses for it from estimating both the size of and the rate of growth in our economy to an ersatz measure of well-being. Such matters attract a lot more attention when there is contraction followed by stagnation as we have experienced post credit crunch.
But also GDP is used as a comparison tool between countries and is also used as a measure of debt affordability when compared to national debts. GDP to national debt percentages are regularly quoted and have become something analysed ever more.
What is GDP?
Here we have the beginnings of trouble as there are in fact three ways of measuring it which use production or output, expenditure and income respectively. The standard measure and the one quoted unless otherwise stated is invariably output.
If you are wondering what happens if they do not give the same answer? Then be relieved that you are asking a practical rather than a theoretical question as this from the Office of National Statistics demonstrates.
The size and direction of the quarterly alignment adjustments in the first quarter of 2013 indicate that, for 2013 quarter one, the levels of both expenditure and income were lower than that of output.
What happens then?
Published GDP estimate – ONS only publishes one, balanced estimate of GDP, taking account of the data from all three approaches to a greater or lesser degree.
If that seems somewhat ad hoc and imprecise then it is. Although we find that in practice output is used and the following happens.
For all periods, the expenditure and income estimates are aligned to the headline GDP figure.
You can thank Nigel Lawson for that bit as he thought that three different numbers might scare the horses so to speak and gave instructions for there to be only one. I discovered this when doing some research into the income series which in the United States has offered useful insight and been a better guide at times than output.
So we end up using this in practice.
The production approach to GDP, known as GDP(P), is the sum of all production activity within an economy. In the form of an equation, this is described by:
GDP(P) = output – intermediate consumption + taxes on products – subsidies on products
If you think how do they measure that precisely? Then you are on the road to realising that presenting a quarterly economic growth number of 0.3% as the UK just has is an example of spurious accuracy. Especially if we remind ourselves that the other two methods (income and expenditure) gave lower answers.
The statisticians at the Office for National Statistics do their best to make the numbers more accurate over time. Although this reinforces the concept that the initial spot estimates which attract so much media attention are unreliable. Indeed today’s update has given us some examples of how it is a mistake to over rely on them.
One media obsession about a double-dip has gone (remember also the treble-dip hype).
GDP growth between Q4 2011 and Q1 2012 has been revised from a fall of 0.1% to flat, thereby removing the phenomenon of two consecutive quarters of negative growth.
Also we see another way in which the credit crunch continues to be a different experience in these updates.
In Q1 2013, GDP was estimated to have been 3.9% lower than the pre-financial crisis peak in Q1 2008. Previously GDP was estimated to have been 2.6% lower for the same period.
The peak to trough fall of the economic downturn in 2008/09 is now estimated to be 7.2%.
Usually recessions get revised downwards over time as the scale of the fall is reduced and occasionally eliminated whereas the credit crunch contraction estimate has just risen.
This theme of the credit crunch being something of a different experience is important as I believe that the considerable change that it represents makes economic data particularly unreliable right now. For example in the worries about UK productivity where falls have been recorded I think that a lot of care is needed. How sure are we that the statistics are accurate and reliable? One of the worst things we can do is make a policy response to an inaccurate number.
You might think that this era might not be the best time to make changes to the calculations and muddy the waters. However this is not so.
I discussed this in detail in my Notayesmanseconomics blog back on the 25th of October 2012. In essence inflation is measured in the national accounts via what are called implied deflators and the calculation of it was changed as the ONS explains below.
The second is replacement of Retail Price Index (RPI) series with Consumer Price Index (CPI) series in forming the deflators
This has this effect as I pointed out on the 25th of October 2012.
So on that a lower inflation measure for the same reading will give a higher growth measure.
So a change representing some 18% of the implied deflator was made (CPI already represented 5%) which will over time lead to higher numbers for UK real GDP and therefore recorded economic output and growth.
I am particularly reminded of this today as we note that in spite of such upward changes the credit crunch has been revised downwards.
Of Royalties and Research and Development
There are planned changes in these areas next summer to bring the UK in line with what has become in the statistical arena a chilling phrase “international standards”. The main change is in regard to research and development where something previously treated as a cost which is recorded as a benefit when it leads to new or higher output suddenly Superman style goes into a phone booth and comes out as an addition to investment.
So our economy will suddenly be larger and growth may be revised higher too. The obvious problem is that statisticians are discovering an apparently larger economy just as we are discovering it is ever more troubled!
For those who wish to look at the detail I discussed the changes in relation to the United States economy on April 22nd.
You may be wondering what this concept is so via the ONS let me explain.
In the National Accounts, the output of housing services comprises not only the services produced by rented dwellings but also those services provided by owner-occupied dwellings.
This is an uncomfortable concept as for example I do not rent my home in any way as I live in it and yet the rental value of it is apparently adding to the UK’s economic output and perhaps growth! However if we give the statisticians the benefit of the doubt here I note that even they have concerns.
The question then becomes how to determine the rent of a similar rented dwelling and match it to an owner-occupied one.
This has become a bigger issue over time as the size of imputed rents used in the calculations has risen. For example they were £117.5 billion in 2011 using 2011 prices but only £86.2 billion in 2007. This is slightly awkward as this means that they are apparently a growing part of our economy. Interestingly they are growing faster than actual rents collected in the economy as they rose by 21% rather than 36%!
The ONS has been investigating this area and has published this only this morning.
Throughout the year a large development project has taken place to change the treatment of ‘Imputed rentals on owner-occupied dwellings’ and ‘Maintenance and repairs of the dwelling’,
Okay and if we cut to the chase the result is?
As can be seen in Table 2 above, the net effect of the new method is not zero for total UK HHFCE (or for GDP). This is because the increase in imputed rentals expenditure is larger than the reduction in maintenance and repairs. (HHFCE) is household final consumption expenditure.
Let me put that into numbers. The impact of UK imputed rent has been raised by £17.6 billion in 2011 as the rise in maintenance was lower than the rise in rents.
This area has attracted my interest as when I looked into the UK’s inflation statistics there were problems with the rental series which has been changed. This seemed to bother officials much less than me as they used rents in the new measure of house price inflation called CPIH. But as we move to the wider implications it seems that issues here are popping up in our GDP statistics too.
It is not my purpose to destroy any faith that readers may have in economic growth figures as represented by GDP numbers. However the production of spot numbers each quarter gets much more analysis than their accuracy deserves and it is my intention to suggest that they are taken with a dose of salt. Unfortunately some of the changes that are being introduced mean that the dose of salt will need to be larger rather than smaller.
There was I believe something for us to learn from the title of this article from Willem Buiter who was my tutor for a year back in the dim and distant past.
I know I know nothing; but at least I know that