Parsing the GDP estimates and calling for more realism about them
The SA economy has now recorded two quarters of negative growth: it is in recession. The decline in real GDP of -0.7% in quarter 2 came as something of a surprise, enough to send the rand significantly weaker. Which in itself makes the growth outlook even less promising given the implications of a weaker rand for more inflation and higher interest rates. That the rand was weaker for South African rather than global reasons was apparent in the performance of the rand compared to other emerging market exchange rates. The rand by time of writing, 16h30 on 4 September, was 2.85% weaker vs the US dollar and only slightly less, 2.2% down on our nine emerging market currency basket.
The logic in the market reactions to the surprisingly low GDP growth estimates seems clear enough – slow growth adds risk to the fiscal outlook. It may encourage the rating agencies to downgrade SA debt further so encouraging capital outflows form the RSA debt market, hence the weaker rand. Though it should be appreciated RSA dollar denominated five year bonds were already trading in junk territory before the GDP release.
Insuring RSA dollar debt against default required paying a risk premium of 180bps. at the beginning of August 2018. The emerging market crisis took this risk premium to 230bps by the end of August. It is now 360bps – up from 330bps previously. The equivalent Turkish risk premium is now 580bps compared to 333bps in early August, up a mere 8bps on the day.
The GDP growth numbers themselves require careful consideration, perhaps more than they have received by the market place. Not only did highly volatile agricultural output drag the quarter to quarter seasonally adjusted annual output growth rates lower by 0.8%, but more important, a decline in estimated real inventories reduced expenditure on GDP by as much as 2.9% at an annual equivalent rate.
By definition output (GDP) is made up of value added by the different sectors of the economy. This estimate of output is by definition identical to the expenditure on this output and the incomes earned producing goods and services. Supply determines demand and demand determines supply – and gives us the national income identity- that is GDP is equal to expenditure on GDP.
This expenditure is made up of spending by households and government on consumption of goods and services and spending by firms and government on additional capital goods- known as gross fixed capital formation (GFCF). To which additional or in the case of Q2 2018 reduced investment in inventories is added or subtracted. Exports less imports are then added to make up the estimate of Expenditure on GDP (see figures 2 and 4 provided by Stats SA). This shows the contributions to the growth outcomes of the different sectors and categories of expenditure. It should be noted that net exports (exports volumes grew faster than imports and so were a positive contributor to growth in GDP in Q2. Perhaps some of these mining exports were sourced from stock piles (inventories) rather than current output- enough to reduce inventories at the rate indicated?
Farming volumes while contributing about 2.5% of all value added in Q2 2018 are inherently variable, subject to drought and flood and may not have any seasonal regularity. The real output of agriculture forestry and fishing fell at an annual rate of 29% in Q2 and was down by 34% the quarter before. In Q4 2017 the growth was as much as 39% at an annual rate and 42% p.a. the quarter before that. For a better sense of sustainable growth rates it might be better to exclude agriculture from the GDP estimates
The run down in inventories had however a much more important influence on GDP growth than farming output in the second quarter as may be seen in the table above. Less invested in inventories reduced estimated GDP growth by 2.9% p.a in the quarter. Inventories fell by an estimated over R14 billion in constant price terms, at an annual rate. Enough to take the growth estimate into negative territory with such significant repercussions and despite the positive contribution of 3.7% p.a made by exports to the GDP growth recorded
The adjustments made for seasonal effects on the change in inventories were highly significant. It is very difficult to make economic sense of the very different estimates of changes in inventories when measured in current or constant prices, or when measured each quarter or alternatively at a seasonally adjusted annual rate. According to the statistics provided by Stats SA the actual quantity of inventories in Q2 grew by R6.7bn (at constant 2010 prices) in the quarter. In current prices and at an annual, seasonally adjusted rate inventories are estimated to have declined by R7.4bn in Q2 2018, (much less than the R14b decline when estimated in constant prices). Using quarterly statistics (not seasonally adjusted) the value of inventories, measured in money of the day, increased by R11.1bn in Q2.
It takes something of a leap of faith to accept and reconcile these very different estimates of inventory changes at face value. When growth in the other components of spending is at the understandably slow rates recorded in Q2 2108, estimates of investments- less or more – in inventories take on particular significance – as they did today on their release. They should perhaps be treated with much greater skepticism than has been the case.
Yet for all these reservations about the estimates of GDP growth, the weakness of household spending cannot be gainsaid. It is by far the largest contributor to expenditure on GDP and on GDP – equivalent to 59.4% of all expenditure in Q2 2018 at current prices. The reluctance of households to spend more is at the heart of South Africa’s inability to grow faster. Without a greater willingness and ability of households to spend more the economy and its output and incomes will not – cannot grow faster.
Household consumption expenditure declined at a 1.3% annualized rate in Q2 2018- taking 0.8% off the estimated growth of -0.7% p.a. The weaker rand and the higher inflation that will accompany the weaker rand will depress household spending further. It is surely inconceivable that interest rates could be raised in circumstances of this depressed kind.
The value of the rand is beyond the influence of interest rates. Surely as much is painfully apparent after events of the past few weeks. But interest rates do effect the ability of the households to spend more. If economic logic were to prevail interest rates in SA would be reduced not increased given the negative growth outlook. And if so the growth prospects would improve – not deteriorate. If so, it might lead to rand strength not further weakness. Weakness that comes with slower growth, as we saw today. 5 September 2018