GDP – An inherently unreliable guide to steering the economy
GDP (Gross Domestic Product) data is awaited every quarter with acute anxiety by UK politicians and economists to see if the long-awaited dawn of growth is being achieved. They are now down to looking at 0.1% changes in GDP from one quarter to the next with the same intensity as the Romans examined the entrails of chickens and other sacrificial animals.
Accuracy of the GDP calculation
GDP can in principle be calculated in three ways:
(a) by adding up all the items of expenditure incurred by households, central and local government, net business investment, charities, and less net imports;
(b) by adding up the incomes of individuals excluding benefits, government income and incomes of corporations with allowance for stock appreciation;
(c) by summing up the added value of each business, self employed individuals, government departments and charities.
Generally the three methods do not yield the same total. The discrepancies are referred quaintly as “statistical discrepancies” or sometimes even more obliquely as “balancing factors”.
The key point though is that each method involves a huge amount of estimation – guesstimation in many cases – which is why the results are revised so often. The NSO (National Statistical Office) can get a fix some numbers through PAYE and VAT receipts for instance, but others such as business capital expenditure are “lumpy” while few households keep records down to 1%. In any event business and household expenditures have to be estimated on a sampling basis – in many cases coming down to inspired guesswork and “feel”.
What have been the sizes of statistical discrepancies in the past?
In 1987 we had a discrepancy between (a) and (b) of 1.6% of GDP – after several revisions in 1988. In 2007 the discrepancy given by the NSO[1] was much smaller – 0.04%. Though not disclosed it is likely that this reduction has been achieved by averaging the results of methods (a) and (b) which should abolish the discrepancy altogether.
Does it make sense to judge the economy’s true behaviour on a quarterly basis by these methods?
The answer has to be no. Trying to steer the economy by looking at 0.1% changes over three months, when it is doubtful if any of the constituent numbers can be known to within 1%, more likely 5% for some, is a waste of time and effort. The data which really matters and which can be known on a month to month basis are: the claimants count for unemployment benefit, the purchasing managers’ index (PMI), and the trade data. Moreover the major OECD countries produce this data on much the same basis, so reasonable international comparisons can be made if required.
Effects of Government “cuts”
Besides the accuracy of the source data problem, there is the fundamental issue of the make-up of the GDP calculation. The UK government is committed to major reductions of some £25 billion in its staff costs corresponding to the loss of 60,000 or about 10% of public sector employment.
As the GDP calculations stand (and it is the same for all countries) these reductions will reduce the GDP on the income measure (b) by the amounts saved, unless the displaced 60,000 (many of whom will take early retirement) find employment. £25 billion is 1.6% of GDP – so even if say 40% of civil servants enter the private sector, the income GDP will show a reduction of 1%, which is about the size of the hoped-for “growth” in 2012. So the very success of reducing the deficit in this way will automatically reduce the appearance of growth in an economy undergoing a much sought-after reduction in its fiscal deficit[2].
A better Guide to Steering the Economy
It would therefore be a useful guide to seeing the real progress of the economy if the National Statistical Office published a private sector GDP – say PSP (Private Sector Product). After all, much of the public sector depends on taxes from the private sector to exist at all.
Such a calculation would force some major institutions such as universities and museums to decide how much of their activities were in the private sector and how much in the public sector – but this would be no bad thing in itself we believe.
Jobs implication of the goods trade deficit
Reducing the monstrous goods deficit of £8.6 billion monthly, i.e. over £100 billion per annum or 7% of GDP (in a recession!) ought to be the number-one target of this or any government. £100 billion equates to about one million jobs in modern manufacturing. Every time a retail chain orders up £100 million of foreign goods, it is effectively costing at least 1,000 jobs somewhere in the UK manufacturing and agricultural economies. Conversely every retail buying manager who replaces £100 million foreign supply by UK produced goods to the same value repatriates around 1,000 jobs for our citizens.
If we can fix the balance of trade problem, growth in GDP will be there automatically without having to gaze obsessively at the detailed entrails of the NSO statistics.
[1] Annual Abstract of Statistics 2007 Table 16.2.
[2] This is similar to the effect on GDP of a manufacturer reducing his prices by x% say. Unless his production increases by at least x%, the result of his hard work through efficiency gains – benefiting the consumer – will be to actually reduce GDP and “growth”. Conversely a lawyer who increases his fees by y% will automatically add this amount to the GDP as visibly happened in the $250 million O. J. Simpson trial in the USA.
