Anybody who thinks US share prices are driven by liquidity should be buying heavily now. Sadly, though, they would be wrong to do so.
Latest figures from the US Federal Reserve show that the M2 measure of the money stock has risen by 22.7 per cent in the past 12 months, the biggest increase since 1943*. If money drives up share prices, therefore, we’re heading for a fantastic bull market.
Except for one problem: it doesn’t. Quite simply, there’s no link between growth in the money stock and changes in US share prices.
One fact alerts us to this. It is that share prices are much more volatile than monetary growth. Since 1971, the standard deviation of annual changes in MSCI’s US index has been five times that of annual changes in M2. It is rare for something stable to be a major cause of something volatile.
Also, there’s just no correlation between the two. Since 1971 the contemporaneous correlation between annual changes in MSCI’s US index and in M2 has been minus 0.05. Since 2000, it has been even more negative.
Nor is M2 a lead indicator of equity returns. Since 1971 the correlation between annual M2 growth and annual changes in share prices in the following 12 months has been minus 0.19. That’s statistically, but not perhaps economically, significant. If anything, stronger monetary growth leads to shares not rising but falling – as happened in 1973-74, 1981-82 and in 2001-02 for example.
There are (at least) two reasons for this lack of correlation. One is simply that shares are driven by countless things other than money, such as earnings, expectations of these (whether correct or not) and attitudes to risk. Traditionally, economists have distinguished between real variables (such as output) and nominal ones such as money. Shares are to a large extent a real variable.
The second is that the phrase “money supply” is horribly misleading. In fact, the money stock is more often determined by demand for it. This is why its strongest growth often occurs in bad times such as 2008-09 – because that is when investors are nervous and want safer assets. Which is why there is a negative contemporaneous correlation between M2 growth and share prices. And if shares don’t bounce back quickly after such bad times, there’ll be no correlation between M2 growth and subsequent equity returns either.
Put it this way. The main reason why monetary growth in developed economies generally has been so strong recently is that the lockdown closed shops and restaurants causing us to build up our bank balances whether we wanted to or not. Why should this fact alone lead to rising share prices?
Money, then, doesn’t matter.
Or does it? My chart seems to suggest otherwise. It plots three-year changes in share prices against the ratio of the money stock to prices three years previously. There seems to be a correlation. For example, high ratios of money to share prices in the early 80s, 2003 and 2009 led to shares doing well, while low ratios in 1972, 2000 and 2006 led to them falling.
Sure, the link’s not perfect. The money-price ratio completely failed to predict the tech boom, for example. But it seems that money does matter to some extent. When Americans have lots of cash and few equities, they rebalance their portfolios towards equities, driving their prices up. And when they have little cash and lots of equities, they rebalance against equities causing prices to fall.
There are, however, two problems with this account. Even if it is true, it does not justify buying equities now because the money-price ratio is still quite low.
The other is that it might not be true. There’s another explanation for the pattern in my chart. It’s simply that shares sometimes overreact, rising too much and falling too much to levels from which they subsequently correct themselves. This means that share prices relative to any stable upward-trending variable will tend to predict returns – be it the ratio of prices to earnings, GDP ir whatever.
We can test this by replacing the ratio of money to share prices with the ratio of a time-trend to prices. If it’s money that drives prices, the money-price ratio should better predict subsequent returns than the trend-price ratio does.
But it doesn’t, at least not much. Since 1971 the correlation with subsequent three-year returns has been only slightly better for the money-price ratio than for the trend-price ratio – at 0.3 versus 0.23. And since 2000 the two correlations have been the same.
My chart, then, tells us that shares over-react and subsequently correct, not that money drives prices.
We should not, therefore, regard the boom in monetary growth as a bullish signal. This isn’t to say the market is heading for a fall: that depends a lot upon the course of the pandemic. It just means that you need to find other reasons for optimism.
* Current records don't go back that far. I'm using Friedman and Schwartz's Monetary Trends in the US and UK.
