/ 26 May 2000

Blame poor growth on the gods

David Le Page

Economists are never short of detractors, and the latter have jeered loudly this week following the release of Gross Domestic Product (GDP) figures for the first quarter of 2000.

The derision follows the huge disparity between the predictions of most economists, who were expecting growth of between 2,5% and 3%, and the actual figure of 0,9%. This was the lowest figure since the 0,6% recorded at the end of 1998.

Most economists had predicted that there would be a decline from the 3,6% increase in the last quarter of last year, but few anticipated the scale of the decline.

Fingers were immediately pointed at the figures for agriculture, forestry and fishing, which declined 17,7% on the last quarter of 1999. That sector was hit hard by high rainfall and heavy flooding, as Statistics South Africa head Dr Mark Orkin was quick to point out. He also pointed to the consoling fact that GDP has now continued to expand through six successive quarters.

The agricultural sector is notoriously volatile. One has only to look at the figures for the last five quarters to see the evidence of that: 6,8%, 15,9%, 38,5%, 12,4% and finally the dread -17,7%.

But why, it is asked, did the economists not anticipate this in making their predictions for the quarter?

“Because we don’t have good economists,” laughs Dawie Roodt of PLJ Financial Services.

He’s frank about his failure to accurately predict the figure: “I do not have an excuse.” He predicted 2,6%, and also points to the volatile agriculture component as the fly in the pundit’s ointment.

But it is not only acts of God – floods, droughts and environmental vagaries – that make the agricultural figures particularly volatile, he explains.

There is also a mathematical reason, forced on to statisticians by the need to produce quarterly figures. Because production of all kinds – but particularly agricultural production – undergoes seasonal variation, quarterly GDP figures are “seasonally adjusted”.

This means compensating for the fact that certain crops are produced only at certain times of the year. Without making a seasonal adjustment, the figures would oscillate even more wildly.

Econometrix director Tony Twine has a more detailed explanation for the disparities, which springs from the way GDP figures themselves are produced.

Essentially there are two kinds of GDP figures – GDP at market prices and GDP at base prices.

Most economists deal in GDP at market prices, which include household consumption figures and government consumption figures. However, these aren’t extracted from a breakdown of production across sectors such as mining, manufacturing, transport and so on. So it’s impossible to compensate properly for the vagaries of agriculture.

But it is not the figures for GDP at market prices which have been receiving all the attention. Rather these economic indicators are for GDP at base prices, or the sum of value added in the 10 major sectors of the economy. So the economists and the statistics organisation are not comparing apples with apples; it’s more like Granny Smith with Royal Gala.

Even the latest figures are not final – the GDP statistic will remain a “living figure” for the next 18 months, subject to several amendments.

Only once all those amendments have been factored in, will it be possible to truly assess the accuracy of the economists’ predictions. But by then, the issue will be long dead. The markets do not look at 18- month-old GDP figures for direction.

Another considerable influence on the GDP statistics was a steep drop in the quarter- on-quarter growth of the manufacturing sector. In the last quarter of 1999 manufacturing shot up an extremely healthy 6,9%. But activity in the sector actually declined in rand terms through January to March, dropping to R33,48-billion from October to December’s R34,41 (in current rand terms).

However, the seasonally adjusted and annualised figures show manufacturing actually increasing 0,5% over this period.

Superficially, a 0,5% increase doesn’t look nearly as dramatic as the 17,7% quarter- on-quarter decline of agricultural production.

But since the manufacturing sector is six times larger than agriculture, what look like insignificant changes has substantial effects on overall GDP.

Ernie van der Merwe of the Reserve Bank points out that there was some levelling off of growth in the non-agricultural sectors, from an annualised rate of 2,5% in the last quarter of 1999 to 2% in the first three months of 2000.

But even if agriculture is removed from the equation, first-quarter growth was 1,9%, still somewhat down from the forecast zone.

Van der Merwe speculates that in the non- agricultural sectors the Volkswagen strike might have been one of the factors that hit manufacturing.

He points out that the maize crop in the third quarter is expected to be around 9- million tons, up from an average 6- to 7- million tons. That should pull up the agricultural production figures substantially, and with them GDP.

The Reserve Bank is positive about the effect of the low GDP figures, saying they should hold back inflation for the rest of the year while growth figures should rebound.

Growth in 1999 and 1998 was 1,2% and 0,6% respectively. The Bureau of Economic Research has revised down its estimate of GDP growth to 2,5% from 3% for this year.