Easter is the bane of those who attempt to measure the temperature of an economy. Without a good fix on current activity it is very difficult to forecast the future. The trouble with the Easter festivals is that unlike Christmas celebrations, they come at different times of the year. An early Easter for the average retailer will add to sales in March and reduce them in April and vice versa when Easter falls in April.
For motor dealers the opposite is true. For some reason, obscure to us, probably due to the regulation of their hours of trading, they stay closed on public holidays and Sundays. In other words, unlike your ordinary retailers who stay open on holidays for the convenience of customers and to the advantage of their part time employees, the motor dealers lose trading days over Easter.
This makes the essential seasonal adjustment more difficult to estimate. For retail sales in South Africa, the Christmas influence on spending at retail level, combined as it with the summer holiday effect on spending is very large. For the average South African retailer December month sales on average account for 35% more than the average month. To get a good idea of how good or bad retailers have done in December compared to past Decembers or to November, sales revenues of the average retailer have to be reduced (divided by) a factor of 1.35. For the motor dealer new vehicle unit sales have to be scaled up by 0.85 (ie divided by a factor of 0.85).
Over the longer run March and April on average have proved to be a slightly below sales months for the average retailer: the scaling factor is 0.98 or 0.97. But life is more complicated for the motor dealer. March is usually an above average month, with a scaling factor of 1.08, and presumably March becomes an even stronger month when Easter does not reduce showroom hours as they did this year. Meanwhile April, presumably because Easter usually but not always falls in April, is a below average month with a scaling up factor of 0.87.
This year, with Easter in March, will be a more difficult year to interpret sales trends for the motor dealers and perhaps also for retailers generally. To get at the underlying trend in sales and sales volumes we would have to scale up for the motor dealers and scale down for the orinary retailers by more than usual, but just how much would be a matter of some guess work. We will have to wait for sales in April to be confident in our measures.
Estimates of retail sales provided by Stats SA are only up to date to February. As we show below, the estimate of sales volumes in February were encouraging, suggesting that , on a seasonally adjusted basis, it was a better month for retailers than January 2013. On a seasonally adjusted basis retail volumes declined by 1.75% in January compared to December 2012 and grew by 2.7% in February compared to January. February volumes, compared to February 2012, were up 7.4%.
However despite this pick up in February sales volumes, extrapolating recent trends, appropriately seasonally adjusted and smoothed, suggests that the growth in retail volumes will continue to slow down marginally over the next 12 months. However this forecast growth in retail sales volumes can still be regarded as satisfactory. Real growth is predicted to be 4% in February 2014. With retail inflation currently running at a 4.6% year on year rate and predicted to rise to 5.2% in February 2014, this suggests that retail sales in current prices may be running at a close to 10% rate this time next year.
What the hard numbers say
We do however have actual vehicle sales volumes for March from the National Association of Automobile Manufacturers (Naamsa). These must be interpreted with caution. We also know from the Reserve Bank the value of its notes in circulation at March month end. We use these hard numbers to compile our up to date Hard Number Index (HNI) of economic activity which is an equally weighted combination of the real note issue and new unit vehicle sales. As we show below, this Index compares very well in its turning points with the delayed Business Cycle Indicator provided by the Reserve Bank. The Index, updated to March, indicates that the economy continued to grow faster in March but that the rate of forward momentum was more or less constant and maybe slowing down.
Values above 100 indicate economic growth. The Index was helped by strong growth in the note issue. This growth too was influenced by the early Easter and the spending intentions associated with it. The demand for cash is itself a coinciding rather than a leading indicator of economic activity. Households hold more cash when they intend to spend more on goods and services. However the advantage of measuring the note issue is that it provides a much more up to date indicator of spending intentions than spending itself. Spending, for example at retail level, is an estimate made from a sample survey, not a hard number, and moreover is only available with a lag. It will only be well into May before we can update our estimate of retail spending.
The close statistical relationship between growth in the note issue and growth in retail sales at current prices is shown below. Both series are on a slower growth trend and are predicted to remain so.
Should such negative trends in domestic spending materialise, more aggressive monetary policy would surely be justified. High rates of inflation that threaten the inflation targets have inhibited such monetary policy responses to date and may continue to do so. However, high rates of inflation cannot be ascribed to excessive domestic demand for goods and services. The trends moreover suggest that the growth in demand will be slowing down, rather than speeding up. The recent inflation in SA have had little to do with excess demand and much more to do with weakness in the rate of exchange and so the costs of imports that reflect also global commodity price trends. These trends, for example in the US dollar price of petroleum, suggest less rather than more inflation to come from this source (independent of exchange rates).
The problem for an inflation concerned Reserve Bank is that there is little predictable connection between interest rates and the exchange value of the rand and therefore very little direct influence the Bank can exert on inflation rates. Higher interest rates, if they implied slower economic growth, might well discourage capital inflows and encourage capital outflows, so weakening the rand and thus add to inflation, even as higher interest rates and a weaker rand discourage domestic spending.
Lower interest rates, where they boost economic growth, might in turn attract portfolio flows to the JSE and lead to a stronger not weaker rand. Faster growth with less inflation then becomes a highly desirable possibility.
Inflation targeting, without being able to predict the direction of the rate of exchange when policy action is undertaken, makes little sense. It may come to pass that the Reserve Bank accepts that the most it can hope to do with its monetary policy is to stabilise domestic spending, without regard to the outcomes for inflation. Recent policy actions by the Reserve Bank strongly indicate that in practice the Bank is following a dual mandate – targeting growth as well as inflation.
If only the rand would behave itself in the months ahead (implying no upward pressure on inflation rates) this dual mandate could lead it to lower interest rates. Recent movements in short and long term interest rates indicate that the money and bond markets are according a higher probability to a reduction in the repo rate over the next 12 months. Brian Kantor