Month: November 2019

Some lessons from share market history

Shorter version published in Business Day, Friday 15th November

The market value of any business will surely  be determined by its economic performance. The most commonly applied and highly accessible proxy for performance are the earnings reported by its accountants and auditors. Cash dividends paid might be a superior indicator given how the definition of bottom line earnings has changed over time. Cash flows may be better still but are less readily available.

Robert Shiller provides 148 years and 1776 months of US stock market data. US S&P Index values, index earnings and dividends per share have followed a very similar path. The correlation between monthly prices, earnings and dividends, all up nearly 20,000 times since 1871, is close to one. [1]

Share Prices, Earnings and Dividends per Share (1871=100) Log Scale

f1

Source; Shiller Data, Investec Wealth and Investment

 

The P/E ratio for the S&P has averaged 15.76 over the long period with a low of 5 in 1917 and a high of 124 after earnings had collapsed in May 2009 while the Index held up to a degree. And dividends held up much better than earnings.

f2

Source; Shiller Data, Investec Wealth and Investment

 

If we run a regression equation relating the Index to earnings or dividends per share the residual of these equations (that what is not explained by the model) has a strong tendency to trend to zero- given enough time. Or in other words the price-earnings ratios have tended to revert to their long-term average of 15.8  over time but with variable lags.

The higher the P/E multiple the faster earnings must be expected to grow when they to make up for the low initial yield E/P and vice versa. Expected earnings drive current market prices. Surprisingly higher or lower revealed earnings will tend to move prices in the opposite direction. Earnings may catch up with prices or prices follow earnings. The move to any long-term equilibrium can come from either  direction, with advantage or disadvantage to shareholders.

Therefore be warned. Knowing that a PE ratio is above or below a long term average is not going to make you rich or poor speculating in the share market. The correlation of starting PE and returns realized over the next twelve months is close to zero.

The starting PE appears to become more helpful as a guide to investors when returns realized over an extended period are compared to it.  High starting PE’s are associated with generally lower returns and vice versa over three or five year subsequent windows. When we relate starting PE’s 36 or 60 periods before with returns realized three or five years afterwards over the entire period we do get a statistically supportive result. Choosing the right entry point to the market and waiting patiently for the outcomes would have been generally helpful to investors.

Examining the relationship between prices and earnings in the US does reveal some extreme cases. In the late forties and late seventies the market would have appeared as very cheap. But these were not good times for the US. The US was fighting and possibly losing a war in Korea. In the mid and late seventies the US was subject to stagflation- rapidly rising prices and slow growth. Also very un-promising times for shareholders.

However normality returned to the great advantage of those who did not share the prevailing pessimism and stayed in the market.  Between 1950 and 1953 the S&P PE crept up from seven to eleven times helping the total returns on the Index to average 22% p.a. In early 1978 the P/E was 8 times. Over the next three years the S&P delivered over 11% p.a on average as the US got its inflation under control.

By contrast in early 2000 at the height of IT optimism the S&P was trading at an extreme 33 times. Over the subsequent three years the S&P delivered negative twelve returns of -8.7% p.a. The IT bubble was only apparent after the event when expected earnings proved highly elusive.

S&P earnings collapsed during the GFC of 2008-09. From about $80 per index share in 2006 to less than $7 in early 2009. This sent the P/E multiple to 124 even as the Index fell sharply to a value of 757 by March 2009.  According to earnings the S&P was greatly overvalued. According to dividends that held up much better through the crisis, the market appeared as deeply undervalued. The dividend buy signal proved the right one as the economy recovered (unexpectedly) with lots of (unexpected) help from the Fed and the Treasury. The S&P Index since those dark days has that provided returns that have compounded on average at of over 13% p.a.

What is very different about the share market today in the US are the extraordinarily low interest rates- both long and short rates – that have surely helped drive the market higher as competition for shares from the money and bond markets fell away. What may appear as a demanding PE of about 22 times becomes much more understandable given abnormally low interest rates. Is this the permanently new normal for interest rates. Or will interest rates mean revert? And what is normal? Only time will tell.

[1] http://www.econ.yale.edu/~shiller/data.htm

The US and the JSE since the Global Financial Crisis – a tale of success and relative failure.

A shorter version was published in Business Day on 1. november.

It is ten years since the Global Financial Crisis (GFC). R100 invested on October 2009 in the 500 companies that make up the most important equity index, the New York S&P 500 Index, would now be worth as R680, with dividends reinvested in the index. This is the result of extraordinarily good 12 month returns over the ten years that averaged  over 18% p.a. in ZAR and 13.4% p.a in USD.

Most other equity markets have not performed anything like as well. R100 invested in the JSE All Share Index or in the MSCI EM, again with dividends reinvested, would now be worth about R276. This is equivalent to an average annual return of about 12% from the JSE over the ten years.

Ten years ago the SA bond market could have guaranteed the rand investor 9% p.a for ten years. Realized equity returns of 12% p.a therefore did not fully reward investors running equity market risks – assuming a required equity risk premium of 4% per annum.

The strong outperformance of the S&P 500 began in 2013 and has continued strongly since. It reflects very different fundamentals. The S&P 500 delivered growth in index earnings per share in USD of 11.7% p.a. over the ten years.  By contrast the JSE delivered growth in index dollar earnings per share of only 1.46% p.a and 5.5% p.a in rands, over the ten years. Barely ahead of inflation.

 

f1

Source; Bloomberg Investec Wealth and Investec Wealth and Investment

Back in September 2009, US Treasury Bonds offered a guaranteed 3.4% p.a for ten years. A good yield by the standards of today. This meant, at that especially fraught time, investors would have required an average return of about 7.5% p.a. to justify a full weight in equities, assuming the same required extra equity risk premium of 4% p.a. Actual realized returns on the S&P therefore exceeded required returns by about a very substantial 6% p.a.

Time has proved that the risks of financial failure in the US were greatly exaggerated. The lower entry price for bearing equity risk ten years ago, reflected by Index values of the time, proved unusually attractive.

Dividends per JSE Index share by contrast with earnings have grown from the equivalent of R100 in 2009 to R324 in September 2019 while earnings per share have no more than doubled. The ratio of the JSE index to its dividends was 41 times in 2009- it is now only 27 times. The ratio of the Index to its trailing earnings per share was 16 times in 2009 – and is the same 16 times today. There has been no derating or rerating for the JSE. See figure below

 

The JSE over the past ten years. Values, earnings and dividends (2009=100) Price to earnings and dividend ratios.

f2

Source; Bloomberg Investec Wealth and Investec Wealth and Investment

The equivalent required risk adjusted return of the highly diversified S&P 500 Index today is a mere 5.8% p.a. on average. With long RSA bond yields now offering close to 9% p.a. the required risk adjusted return from the average company listed on the JSE is now at least 13% p.a.

It has become increasingly difficult given slow growth and low inflation for a SA based company to add value for shareholders by earning returns on its capital expenditure  of over 14% p.a.  . And SA business has responded accordingly by saving and investing less and paying out dividends at a much faster rate. This is not good news for business or the economy. JSE listed companies would be much more valuable if they could justify investing more and paying out less- as US companies have done.

They need encouragement from faster growth in their revenues and earnings and lower interest rates. Lower short-term interest rates are in the power of the Reserve Bank and if reduced could help stimulate extra spending by households. SA business and its share market also need the encouragement of lower required long-term returns. That is from the lower long-term interest rates that would come with less inflation expected. Less inflation expected (and consequently lower interest rates) means a growing belief that SA will not fall into a debt trap and print money to escape it- that could be highly inflationary. So far and after the MTBPS last week not obviously so good. The jury remains very much out on the ability of the SA government to manage its debts successfully and the cost of capital for SA business has become even still more elevated.