Article first published on Advisor Perspectives July 24, 2012
Skeptics and devotees of technical analysis took notice last week when Albert Edwards, the closely followed investment strategist at Societe Generale, warned the S&P 500 was “on the verge of an ultimate death cross,” foretelling imminent major losses for the stock market. Edwards’ sense of doom is misguided. An ultimate death cross is mathematically impossible unless the S&P were to suffer an immediate and precipitous decline. Moreover, the signal would provide a positive outlook, if it were to occur.
The ultimate death cross is when the 50-month moving average (MA) of the S&P moves below the 200-month MA. The difference between these moving averages – the spread – must be less than zero for an ultimate death cross to occur. The spread will form a trough before the end of this year, irrespective of the level of the S&P over the next few months. This could be the harbinger of good news – a macro signal that the stagnation period since the year 2000 for the S&P is now finally over, and that a new secular bull market could commence.
For the period for which I have daily data (since 1965), the 50-month MA has always been above the 200-month MA and currently (July 20, 2012) the 50-month MA is 1,150.99 and the 200-month MA is 1,136.87. The difference between them, the spread, is 14.12, and it is declining. Fitting a second order polynomial to the spread over last six months, I found that the 50-month MA will approach the 200-month MA in a few months, but never crosses it to the downside. Below is a chart of this indicator, the green line, where one can see that it is possibly approaching a trough shape, as happened in 2006.
But let us assume that the S&P starts to decline from now onwards at a 30% annualized rate for the next year. One would then expect that the short-term 50-month MA would fall below the long-term 200-month MA, and then continue to fall lower and lower relative to the long-term MA, reflecting the continuous decline of the market.
This is not the case.
If one extrapolates the S&P forward in time based on this assumption, and then calculates the moving averages one finds that the spread reaches a low of 0.25 in November 2012, and then starts to increase again, and by the end of July 2013 reaches a level of 21.4 with the S&P at 998 (as shown by the red graph in figure 1).
Assuming the S&P remains range bound at the current level, then the spread reaches a low of 5.0 in the middle of November 2012, and then starts to increase again, and by the end of July 2013 reaches a level of 56.7 with the S&P at 1362.
The market would have to decline at an annualized rate of 40% in order for the spread to become negative, which would happen at the end of November with the spread reaching a low value of -1.3, but then would increase again to 11.2 by the end of July 2013, with the S&P at 903.
Unless there is huge decline in the market, the ultimate death cross will not occur for the foreseeable future.
Figure 2 shows the same graphs as figure 1, but over a longer time period, from 1870 to 2012. The data is from Robert Shiller’s monthly S&P 500 data series. There were only three occasions when 50-month MA moved from above to below the 200-month MA – in June 1894, November 1917 and May 1934. Two years after the first date the S&P was unchanged, and two years after the second and third dates the S&P was up 30% and 44%, respectively. A near miss was in June 1978 when the short-term MA came within 2.02 of reaching the long-term MA. That was the year when the secular bull market began, which ended in 2000. As demonstrated above the spread will make a trough at the end of this year, similar to what happened in 1978, which may be foretelling the start of the next secular bull market.
I do not share Doug Short’s view that this signal is “completely worthless as an indicator for the market or the economy.” Only on one occasion (1894) has this signal provided a neutral forecast for market direction; on all others it provided a positive forecast for the stock market. It was never a harbinger of doom, but actually was a bearer of good tidings – hence it is inappropriately named.
Edwards could have done the same calculations which I did and come to the same conclusions, however this would not have supported his bearish views. You can read more about Edwards’ bearish predictions here.