In my e-mail note of 8th April I argued that the bull market was “running on empty”. The argument was that the nominal level of equity prices could be viewed as the product of
- The quantity of money allocated to equity funds by investors, and
- ii. Equity fund managers’ desired ratio of money to their total assets (i.e., their “bullishness” or “bearishness”, or indeed their “liquidity preferences”, as Keynes would have put it).
(This formulation was a little naive and begs many questions, but it gets the ideas over.) My worry in that note was that in the USA the bull market from March 2009 had depended almost entirely on a fall in fund managers’ desired ratio of money to assets (i.e., on medium-term bullish sentiment). Money growth had been negligible at best since early 2009. On some measures – such as the old “M3”, no longer published by the Fed, but easily enough guesstimated – the quantity of money had in fact started to fall in late 2009.
Since peaking in late April, most equity markets (including the US market) are off by about 10% and are lower than at the time of the 8th April note. Of course many other factors influence the direction of markets, with alarm about the potential break-up of the Eurozone certainly being a major negative at present. However, I thought it worthwhile in this weekly e-mail to update the numbers in the 8th April note. The main point is straightforward, that the US banking system remains gripped in a regulatory vice which has stopped it from expanding. M2 is going sideways, while M3 – which consists of M2 plus other money balances (notably the “institutional money funds” which are the main form of cash for the USA’s long-term savings institutions) – continues to fall. So the US equity market is still “running on empty”. Perhaps most worrying of all, there is little evidence that the Federal Reserve’s research effort pays much, if any, attention to the money trends which I have been analysing.
Money and equity markets: some theory
The 8th April note contained quite a long discussion of the relationship between money and asset prices, and equity markets in particular. There is no need to repeat all that here, except to say that the analysis could be summarized in the proposition that
and that this could be brought closer to institutional realities by recognising that a high proportion of variable-income assets are in fact held by specialist fund management organizations. Specifically, the level of equity prices could be interpreted as the product of
- The quantity of money allocated to equity funds by investors, and
- Equity fund managers’ desired ratio of money to their total assets (i.e., their
“bullishness”, “bearishness”, or “liquidity preferences”, with the phrase chosen depending on taste).
If liquidity preferences are stable (and they are in the long run), the change in the nominal prices of variable-income assets (i.e., equities and real estate) is determined by the change in fund management institutions’ money holdings. The cyclical patterns are complex, but – roughly speaking – changes in the quantity of money held by long-term savings institutions are more volatile than the total quantity of money. Meanwhile short-run movements in equity and real estate prices reflect both fluctuations in the allocation of money to investment portfolios and changes in investors’ “liquidity preferences”. Given the wild swings in mood that characterise financial markets, it is hardly surprising that markets in equities and real estate are extremely volatile. Nevertheless, it remains true that – if liquidity preferences are stable – the change in equity prices should be similar to the change in the quantity of money “in the financial system”, while the change in the quantity of money in the financial system depends heavily on the change in the aggregate quantity of money.
Most of my analysis of these topics has been with UK data. But in 2006 and 2007, as a research associate of the Financial Markets Group at the London School of Economics, I spent some time checking whether long-run patterns like those in the UK also obtained in the USA. In some respects the US data were better, since they went back further (i.e., from 1952) and had fewer series breaks. My main findings were simple. Over the 53 ½ years from 1952 Q1 to 2005 Q3 the institutions’ total assets climbed by 187 times and their money holdings by 214 times. The annual compound growth rate of total assets was 10.3%, very close to that of money assets which was 10.6%. True enough, occasional changes in the ratio of money assets to total assets did occur, with the money/asset ratio varying from a low of 2.3% in 1960 to a high of 6.0% in 1984, but the change between fourth quarters (i.e., the annual change) never exceeded 1% of total assets.
Whatever one’s prior beliefs, this stability is surely impressive. It does not prove that a given x % change in long-only savings institutions’ money will be accompanied over the next two or three quarters by an identical x % change in their total assets, but it does argue that over the medium and long run changes in money assets and total assets are likely to be highly correlated.
In other words, English-speaking fund managers in the USA behave in much the same way as their English-speaking counterparts in the UK. (Surprise, surprise.) When deciding on investment strategy, have a target ratio of cash (and/or liquid assets) to total assets in mind. If their actual ratio differs from the target ratio, they take action to restore the target ratio. For example, if the institutions’ money-to-assets ratio is lower than target, they try to sell equities and increase cash. But – to the extent that the institutions buy and sell from each other – these transactions do not change cash holdings for the industry as a whole. (Buyers are exactly matched by sellers.) Equilibrium is regained not by a fall in the institutions’ hands, but by a decline in the value of stocks and shares relative to the unchanged quantity of money. That raises the cash/asset ratio, as desired. (The message for market participants is to worry about a fall in share prices when the cash ratio is low, i.e., when most participants are bullish, and to place bets on a rise in share prices when the cash ratio is high, i.e., when most participants are bearish. Extreme high and lows in the cash ratio usually mean something for future market movements.)
What message for American equities at present?
1. Sharply falling money balances of long-only US savings institutions
The central feature of the American monetary environment has not changed in the last few weeks. The regulatory pressure on banks is causing them to limit risks and shed assets, resulting in negligible growth or even outright declines of broad money. In the three months to 10th May M2 rose, but only at an annualised rate of 1.0%. (The annual growth rate is now a mere 1 ½%.) The characteristic types of money balances in the hands of long-only savings institutions are institutional money market funds and securities repurchases. (Many of the securities repos are arranged with investment banks rather than commercial banks.) Neither of these are in fact included in M2, but they were part of the old “M3” measure which the Fed discontinued in early 2006. Data are still prepared for both types of money, with the number for institutional money market funds supplied weekly to the Fed by the Investment Company Institute of America and securities repo information provided as part of the Fed’s quarterly flow-of-funds.
There is nothing to say on securities repos relative to the situation a few weeks ago, but what about institutional money market funds? They peaked in May 2008, when they stood at over $2,500b. and were almost 15% up on a year earlier. Their high value at that date may partly reflect a diversion of business towards them, as the “shadow banking system” (i.e., the marvellously named “non-bank banks” and similar) disintegrated in 2007 and early 2008. Since then they have been falling relentlessly in value. It is quite reasonable to attribute this partly to fund managers’ bullishness and partly to the very low rate of return now available on cash. To some extent the fall in amount of money held in the financial system has been offset by rises in companies’ money holdings, because of heavy corporate fund raising in the last year or so. Nevertheless, the aggregate amount of money cannot be changed by switches of money between different kinds of bank account. In the background a fundamental reason for the fall in the value of institutional money market funds is therefore that the aggregate quantity of money has stopped growing. (If the aggregate amount of money is constant and the money balances of other holders [such as industrial and commercial companies] have been rising, the money balances held by the financial institutions must fall.) The alarming feature of the chart below is the apparent acceleration in the rate of decline in recent months. In the year to November 2009 IMMFs were down marginally by 2.3%. By contrast, in the year to April 2010 they dropped by 22.8% and in the three months to April they fell at an annualised rate of almost 37%.
The chart shows the annual % rate of change in institutional money funds since the mid-1980s. (Before then they were too small to matter.) Two obvious features are
1. the extreme volatility in this type of money holding (which is also matched in the UK in the large ups and downs of financial sector money, see pp. 56 – 74 of my 2005 IEA pamphlet Money and Asset Prices in Boom and Bust) and
2. the severity of the recent fall.
A clear message from the chart is that over this 25-year period, if institutions’ liquidity preferences (i.e., ratio of cash to assets) had been constant, the fluctuations in institutions’ money balances would themselves have led to great swings in equity markets. In fact, the relationship between the change in institutional money funds and the change in equity prices is not good, because investor confidence is far from stable from year to year. At any rate, the fall in institutional money funds in the last few months has no precedent in the lifetime of this kind of money balance. By itself this development is undoubtedly bearish for equities and other asset markets. (However, to repeat, a mechanical relationship is not to be expected. Investment funds are balancing money not only against equities, but also against bonds, real estates and various other more miscellaneous asset classes. Meanwhile equities can also be purchases by companies and individuals, and so broad money as a whole needs to be tracked.)
2. M3 measure of broad money also falling
Because companies and individuals are also significant participants in US capital markets, aggregate broad money measures need to be monitored in addition to money holdings exclusively in financial institutions. As noted earlier in these e-mails, since February 2006 the research company Shadow Government Statistics has compiled a guesstimate of “M3” – a significantly broader money measure than M2 – from a variety of official sources. (The types of money balance in M3 but not in M2 are the “institutional money funds” already discussed, large time deposits, some repo balances and Eurodollar deposits held by US residents.) The data for M3 give a feel for the type of money balance that companies and individuals, as well as financial institutions, are balancing against equities and other assets in their portfolios. The chart below borrows from the Shadow Government Statistics information
Like the previous chart, this one has to be described as disturbing, perhaps even very disturbing. On the SGS numbers, M3 peaked last June and has been sliding ever since. There can be no dispute that the dominant reason for this contraction is the official pressure on the banks to raise capital/asset ratios and to shrink their risk assets. Many banks are complying with the regulatory edicts against their better judgement and resent the demands to shed profitable assets. At any rate, in the year to April M3 was down by 4.7%, in the six months to April it fell at an annualised rate of 8.2% and in the three months to April it fell at an annualised rate of 9.6%. On the face of it, the money contraction is intensifying.
In the Great Depression from 1929 to 1993 M2 fell from peak to trough by about 40%, i.e., at an annualised rate of decline typically of a somewhat more than 10%. The “M2” of the early 1930s is similar in conception to the “M3” of today. In the last few months the fall in M3 has been proceeding at a similar rate to that seen in the USA’s Great Depression.
Do American policy-makers care about the money slump?
So far the American recovery has been rather disappointing. It has failed to lower unemployment and has owed much to the ending of inventory rundowns rather than a revival in final sales. But there has been a recovery of sorts, with the overwhelming majority of market commentary and macroeconomic forecasting expecting it to continue and even to gather pace in the second half of 2010. The two charts above represent a challenge to the conventional complacency. For all the difficulties in the relationship between money and the economy, a great deal of evidence argues that expenditure and output do not generally grow at above-trend rates when real money balances are being squeezed. Until now business surveys have been good, even quite positive about the US economy, and the Conference Board’s leading indicator index has been upbeat. But the latest Conference Board LLI was a small drop and such sectors as residential real estate and house-building, classic advance indicators of economic activity, have been skittish at best in recent months. The recent fall in the stock market is also itself negative for spending.
The falls in broad money, and the steep drop in financial sector money balances, are certainly negative by themselves for US asset markets. Is a double-dip recession possible? The answer has to be “yes, of course”. In fact, for all the problems in the money/expenditure relationship, if M3 continues to fall at about ½% – 1% a month (and this is a big “if”), a double-dip recession is a virtual certainty. However, the likelihood has to be that – once disappointing macroeconomic signs emerged – the Federal Reserve would embark on another round of securities purchases. The Fed might even see the merits of large-scale and deliberate purchases of perhaps $1,000b. of government securities, rather like the UK’s quantitative easing, with the explicit purpose of boosting broad money by 5%, 10% or whatever. Fed economists and officials don’t think in these terms, but needs must when the devil drives.
The big problem here is intellectual, that American economists close to the heart of policy- making do not incorporate the quantity of money, in any shape or form, in their analysis of the macroeconomic scene. Krugman and Stiglitz – who are listened to with enormous respect in Obama’s White House – despise traditional monetary theory and have a religious aversion to any mention of the quantity of money. Summers is little better. All three of Krugman, Stiglitz and Summers believe in a primitive fiscalism, that if the budget deficit expands by 10% of GDP that represents a 10% “demand injection” which should lead to rapid output growth and falls in unemployment. So far they have been wrong. The USA has had a massive fiscal stimulus – or rather what its proponents call a “stimulus” – and the unemployment rate is stuck close to its peak. There is no sign from their statements that any of this very influential Keynesian trio has had a rethink. (Has Obama started to see through them?) However, in their general attitude towards traditional monetary theory, most economists at the Federal Reserve are little better than Krugman, Stiglitz and Summers. In fact, hardly anyone occupying a prominent position in senior US macroeconomic policy-making is worried about the money slump.
This is not the 1930s. No country is on the gold standard and, in extremis, money can be created easily by QE-type operations. But policy-making is in chaos as a result of economists’ various confusions. The American money slump – along with the risk of a Eurozone break-up – argues that investors ought to hold above-average cash weightings in the next few months, at least until the situation becomes clearer.