March 6, 2007
Equities look overvalued, but where is the turning point?
Is the market turbulence of the last week telling us something or is it no more than “a tale told by an idiot, full of sound and fury, signifying nothing”? Some analysts are prepared not only to explain day-to-day movements in markets, but to predict them. I am neither clever enough for the former, nor rash enough for the latter. I am prepared, however, to make four statements: first, a period of market volatility is welcome; second, core equity markets do look overvalued; third, that this does not appear to be the case is due to the extraordinary condition of the world economy; finally, the big question is how long those conditions will endure.
Any long period of market stability encourages speculation. Taken to excess, such risk-taking, particularly when fuelled by huge amounts of borrowing, can create significant instability. At a time when asset markets are generally buoyant and risk premiums low, the need for a reminder of riskiness is valuable. It is far better, as natives of San Francisco must know, to suffer a series of mini-earthquakes than a long period of calm, followed by a huge one. Similarly, euphoria in markets is dangerous. From time to time it needs to be punctured, before bubbles reach the proportions seen in Japanese markets in 1990 and US markets in 2000.
The remainder of Martin Wolf’s column can be read here (FT.com subscribers only). Discussion from our guest economists is free.
Stephen Cecchetti Martin’s column reminded me of a recent encounter with a book entitled “The Complete Idiot’s Guide to Market Timing.” There are hundreds complete idiot’s guides. While you can get a guide to pet psychic communication, the financially related ones are generally pretty good. I think of them of as helping otherwise uninformed people to do smart things. But the market timing guide gets it backwards; it helps otherwise smart people do stupid things. I firmly believe that divining market turning points is a matter of pure luck. We should all be investing for the long term, and that means riding out the ups and downs of the market. Practicing what I preach, I have never sold any of my investments and I barely look at the periodic statements that come in the mail. As Martin says at the end of his column, “Forecasting [turning points] is for far cleverer and braver people than I am.”
Regardless of my own investment decisions, unraveling the sources and implications of equity market volatility does provide me with an intoxicating puzzle. In my role as financial system detective I have two comments, one regarding the level of equity prices today and the second about volatility. On the level, it is always straight forward to compute the approximate equity premium that can justify the current price earnings ratio. (The earnings yield, E/P, equals the risk-free rate minus the equity premium.) Using Bob Shiller’s cyclically-adjusted numbers (computed using a 10-year lagged moving average of earnings) we get that the current equity premium is around 1.5 percent. Given recent history, this is a respectably high number – during the dot com boom of the late 1990s, the implied equity premium for the U.S. market was negative. It is times like that when policymakers should consider defensive actions; but that argument is for another day.
Turning to volatility, here I must say that I am much more sanguine that Martin (or than Larry Summers in his December 26, 2006 column posted in the forum). Over the last few decades, financial markets have been remarkably resilient, absorbing an unrelenting series of large and small disturbances. But, as Jeremy Siegel first pointed more than a decade ago, most volatility is short term. Remember, someone who looks at equity prices only on the first of each year would have seen a gain of nearly 6 during 1987 – the year in which equity prices sustained their biggest single day decline in history. My conclusion is that over even the medium term, our industrialized economies are really very stable. And, more importantly, we have done a very good job of insulating the real economy from financial disturbances. This is partly a matter of building institutions and markets that efficiently allocate risk, and partly matter of having financial and monetary policy makers who understand how to do their jobs.
There will always be challenges, and we will continue have puzzles to work out. But in the end, I’m simply not worried. I’m not worried in the short run about increased volatility of the equity markets, and I’m not worried in the longer run about the impact that financial volatility will have on the real economy.
Posted by: Stephen Cecchetti | March 8th, 2007 at 1:58 pm | Report this commentAndrew Smithers: I naturally agreed with your admirable article. It helps that you make the point that value is not a guide to short-term market performance. If it were, markets would never get out of line with value. Value is, however, some guide to future returns - this is a necessary condition for a valid criterion of value.
I have two small points: (i) The long-term average PE is 14 rather than 15, even when boosted by recent high multiples. It is important to measure this by the geometric mean rather than the arithmetic one. If the arithmetic mean is used, the inverse of the average earnings yield is significantly different from the average PE but, with the geometric mean, they are the same. (The CAPE average is around 15, but this reflects the lower EPS of the past 10 years, compared to today.)
(2) The reason that the real earnings of companies has grown much more slowly than GDP is not because the index is underweight high return companies, but because the measure is per share and there is no reason why it should grow in line with GDP. The real return on equity appears to be stable around 6 to 6.5% and EPS growth will therefore be a function of plough-back. This must be so to conform with the Modigliani-Miller theorem that returns are independent of pay-out. If companies’ profits are correctly recorded, then EPS growth will be equal to the long-term return on equities, minus the dividend. Assuming 6.5% return and average pay-out of 50%, the dividend growth of EPS will be 3.25% and, if the pay-out ratio is 90%, it will be 0.65% p.a. EPS growth is independent of the growth of GDP. (Though I am yet to find an actuary who can understand this.)
In practice, we know from the data that dividend growth has been much less than that suggested by the pay-out ratio. This is because profits are habitually overstated. This is also shown by the difference between (a) the average earnings yield and the return on equity to investors (b) the return on equity to investors and to companies.
I recently summarised this in our report on depreciation (No 287), issued on 14th February, 2007.
Whether or not companies which are not in the index have higher returns than those in it is irrelevant. The returns on companies outside the quoted US universe cannot account for inconsistencies in that data universe.
Posted by: Andrew Smithers | March 8th, 2007 at 2:13 pm | Report this commentBy invitation - John Thomson, RIA Capital Markets: The anchor for the valuation of equity markets in the last year has been the belief that the US economy is going to have a soft landing.
Hence, the concern in the last week about Greenspan’s ‘one third probability’ of a recession. I would put the probability at closer to two thirds, simply because the imbalances are so much more significant in the current economic cycle (balance of payments surpluses and
deficits, private and public sector debt levels, and so on).
I agree that the lead will be provided by US corporate earnings. UK equities are currently fair value relative to historic PE and gilt yield
relationships. Longer term they should continue to outperform bonds.
However, the period of slower US profit growth and world economic growth will affect sentiment in the UK in the next year or two. I would be a
short term seller of equities into any strength from current levels.
(John Thomson is currently Chairman of RIA Capital Markets, an agency broker specialising in bonds. His background is equity fund management.
Posted by: FT Economist Forum | March 8th, 2007 at 3:28 pm | Report this commentPreviously he was Chief Investment Officer of Standard Life and CEO of Stewart Ivory. The above views are personal and should not be taken as
‘advice to investors’.)
Willem Buiter: There’s an old saying I don’t know the source of: “the prices that move don’t matter and the prices that matter don’t move”. Financial asset prices and commodity prices tend to be lumped into the first category, earnings of workers and the prices of other services and processed industrial goods into the second.
I believe that if we interpret ‘don’t matter’ as ‘don’t have much influence on real economic activity’, that is, on the demand for and supply of currently produced goods and services (including labour services), the statement would seem to be an increasingly good approximation to reality. Stock markets boom and crash with little effect on the real economy. The US housing boom and bust likewise appears to have had but a limited impact on consumption and investment demand outside the housing sector. The price of oil goes from $10 to $70 and back to $50 without leaving much of a mark on the economic landscape. I believe that the decline in equity prices of the past three weeks, and the widening (at last!) of credit spreads will likewise turn out to be, from a macroeconomic perspective, ‘much a do about rather little’. Why does real economic activity appear to have become increasingly uncoupled from the behaviour of asset prices?
One reason is that most financial assets are held by households and by institutions (on behalf of households) that are not liquidity-constrained. When their financial wealth changes, the incremental impact on consumption is small – at most that of a real annuity rate of return. If households are dynastic and care about their descendants, the ‘marginal propensity to spend out of financial wealth’ could be as little as three percent – rather less than the ‘marginal propensity to spend out of disposable income by liquidity-constrained Keynesian consumers, which may be close to 100%.
There are additional effects on spending of changes in some financial asset prices, like equity. If the price of equity is not just a measure of the value of ownership claims on existing capital but also of the value of increments to that capital stock ( in economese, if ‘Tobin’s average q’ is also ‘Tobin’s marginal q’), then higher equity prices boost capital investment by firms. Such effects are, however, very hard to find in the data.
Only a small subset of the gross stocks of financial instruments outstanding are direct claims on ‘outside assets’, that is, assets that are in positive net supply. For a single national economy, this would include claims on the earning of real assets (land and physical capital assets, goodwill etc.) and net financial claims on the rest of the world. The question of whether claims on the government should be treated as ‘outside assets’ is best left to another discussion. Changes in the prices of financial instruments that are direct claims on outside assets do represent net wealth effects in the aggregate. But the consumption and investment impact of such asset price changes are small.
Apart from these ‘outside assets’, an obvious reason why the effect on consumption demand of a change in financial asset prices may be even less than would be implied by the status of financial wealth owners as ‘permanent income’ consumers rather than Keynesian consumers, is that most changes in financial asset prices don’t change aggregate financial wealth. Most financial claims are ‘inside claims’, that is, paper or disembodied electronic claims by one private party on another private party. Such claims are, by definition, in zero net supply. All credit derivatives and asset backed securities fall into that category. When the price of such inside claims falls, there is a lump-sum redistribution from the party that is long the claim to the party that is short the claim. There is no effect on aggregate private financial wealth. There may of course still be effects on consumption demand if the response of the loser does not mirror that of the winner, but little is know about the direction and magnitude such ‘redistribution effects’. From a macroeconomic perspective changes in the prices of inside financial assets bring about zero sum wealth effects that can be safely ignored most of the time.
Increased financial asset market volatility and uncertainty (as reflected in risk premia) may boost precautionary saving and deter investment. The point is granted in principle, but very hard to find in the data in practice, except when faced with truly dramatic financial blow-outs of the kind that is seldom seen.
Of course, for the players in the markets for inside financial assets, asset price changes are a matter of life and death precisely because they are zero sum, regardless of whether they are up or down. For every winner there is a matching loser. Fortunes are made and lost when the price of any financial derivative changes in either direction. The private, individual importance of changes in the prices of inside assets does not, however, translate into macroeconomic significance, unless there are massive asymmetries in the response of winners and losers to their gains and losses.
Much financial wealth is owned by extremely rich individuals and by financial institutions that are well-capitalised. Increasingly, the portfolio allocation decisions (and innovation in the design of new, complex derivatives) seems to be motivated not by the desire to diversify risk but by the desire to take on additional risk. This may be motivated either by a belief that the market overprices risk and that the investor in question is smarter than the market. It may also be motivated by risk-loving rather than risk-averse mind sets. Some very rich investors appear to approach investment decisions with the attitude of the gambler who loves the thrill of taking on additional risk, even if the expected returns do not justify this. The money that is made and lost ‘betting on the ponies’ or gambling in the financial markets, is like a revolving fund that never gets spent on consumption. Individual fortunes are made and lost, but except for those immediately affected, the world keeps turning much as it always has.
Does this mean that financial markets don’t matter for macroeconomic performance? That would not be the correct conclusion to draw. First of all, the existence of well-functioning, deep financial markets and of financial institutions offering a diversified set of financial instruments can matter greatly for growth and prosperity even if, once they exist, wide swings in asset prices set in these markets have little macroeconomic significance. Although even in the financially most advanced societies most investment is financed internally, access to external finance is important for enterprises. Likewise, the ability to hold a diversified portfolio of financial instruments is important to savers. Risk trading through insurance contracts and through the issuance of private contingent claims generally is a vital (and increasingly important) complement to the risk-sharing mechanisms provided through the government budget and public debt management. But much financial sector and asset market activity is individually important but collectively insignificant Sturm und Drang.
In addition to the distributional and net wealth effects of financial asset price changes, there are incentive and informational effects. If the asset price changes reflect fundamentals, resource allocation will be improved and there will be an aggregate efficiency gain. If the asset price changes reflect fear, greed, other collective mood swings and herding behaviour detached from fundamentals, resource allocation will be impaired and there will be an aggregate efficiency loss.
If either the negative net wealth effects of outside asset price declines such as the recent fall in equity prices) or the wealth redistribution effects of inside asset price changes (such as the recent increases in the price of laying off credit risk) were to cause widespread and serious default and bankruptcy (directly or indirectly), the impact on the real economy could be severe. There is no evidence of that so far.
For a number of years, starting in 2003, risk premia, especially credit risk premia of all kinds, have been ridiculously and unsustainably low (see e.g. my . “Threats to the orderly resolution of global imbalances: trade disputes, abrupt corrections of global asset market anomalies, and past, present and future monetary policy errors by the Fed”, Mimeo, European Institute, London School of Economics and Political Science, October 11, 2006 at http://www.nber.org/~wbuiter/crash.pdf). This posed a threat to macroeconomic stability, because such under-pricing of fundamental risk led to excessive leverage and to increasingly unsafe financial and real investment decisions. Excessive leverage was further encouraged through the very low risk-free long-term real interest rates in the industrial world. Some part of the low real risk-free rates can be rationalised through the some combination of the ex ante ‘saving glut ‘ hypothesis and Caballero’s ‘quality financial asset shortage’ hypothesis. However, there remains a substantial unexplained low real risk free interest rate anomaly, especially in the UK, where even today very long-term real rates are barely above 100bps. Further increases in credit risk spreads (including their contagion to instruments hitherto largely untouched) are highly likely. If this asset market anomaly correction were to coincide with a correction of the low long real risk-free rate anomaly, the asset market turbulence we have seen recently could become more of a storm. Even then, as long as the monetary authorities stand ready to provide liquidity to the markets, there should be no reason for global macroeconomic concern. Better to have the correction now than later, when further distortions would have evolved and even more excessive leverage taken on.
Posted by: Willem Buiter, London School of Economics | March 9th, 2007 at 4:04 pm | Report this commentMartin Wolf: I thank those who have commented on my column. I hope Steve’s optimism on the stability of the economy is right. Whether he is right on the equilibrium equity risk premium is another question.
If the gap between the return on equity and bonds (in real terms) is going to be lower than ever before, in equilibrium, something also has to change in the economy itself. Given the fact that real interest rates are low, by historical standards, the cost of corporate capital in the economy must then be low by historical standards, too. That, in turn, means that the return on capital, in equilibrium, must also be lower. The only way I can see that this can happen is with a higher ratio of capital to labour (and, presumably, of capital to output) in the economy. There seems to be no evidence that this is happening (it would imply a huge investment boom, which is certainly not happening) and I can’t see why it should do so.
I accept Andrew’s points. It is obviously important that the measure in question is earnings per share, not total earnings. I also thank Mr Thomson for his brief comments. My assessment of the macroeconomic picture is quite close to his. It will be interesting to see if our view or Steve Cecchetti’s of the macroeconomic prospects proves more accurate.
Posted by: FT Forum - Martin Wolf | March 9th, 2007 at 6:06 pm | Report this commentMartin Wolf: Willem has provided a lucid discussion of the (non-)relationship between financial markets and the real economy. It does raise some interesting points, however.
One is whether ultimate holders of wealth (households) are likely to make serious mistakes. Willem’s answer is that this is improbable, because changes in wealth are an insignificant determinant of consumption decisions, unless people are credit-constrained. I find this a surprising conclusion when one looks at the recent saving and borrowing behaviour of US households. But it is possible.
The second question is how far the redistributive consequences of massive changes in what Willem calls “inside assets” matter. My normal assumption is that mass bankruptcy in the financial system ultimately matters only if it threatens the solvency of the banking system. Then the banking system will need to be rescued, as it has been so many times in so many countries, by transferring a large proportion of its liabilities to the public sector. That would certainly lead to the re-regulation of finance with possibly large consequences for its future working. Is the banking system now protected? Honestly, I don’t know. Do the bankers? Really?
The third question is how far the widespread belief that the central bank can deal with any problems rests on the assumption that there is no significant inflation. I presume Willem agrees that it does: in an environment with high inflation, the central bank would have to target prices, regardless of what happened to the financial system. The consequence would be a recession that could well undermine many of the assumptions on which current risk-premia are based.
The final question is perhaps the most obvious. If vast amounts of ingenuity are being devoted to essentially zero-sum activities, what is the gain for humanity?
Posted by: FT Forum - Martin Wolf | March 13th, 2007 at 10:08 am | Report this commentBy invitation - Andrew Tilton, Goldman Sachs: Martin Wolf asserts that the US equity market is overvalued. His main point is that, although the price-earnings ratio of the S&P 500 is currently near its long-run average of 15, the denominator — the ‘E’ — is unsustainably high. Relative to a “cyclically adjusted earnings” measure espoused by Professor Robert Shiller of Yale University (a ten-year inflation-adjusted average), the P/E of the market would be 26.5. That P/E would be hard to justify even in the current
low-interest rate environment.
Certainly, profits have shown phenomenal and unsustainable growth in recent years. At the level of the national economy,
before-tax corporate profits, adjusted for inventory valuation and depreciation distortions, have more than doubled over the past
five years. The profit share of GDP was nearly 12.5 per cent in the third quarter of 2006, the highest share since 1950.
We sympathize with the idea that the current profit cycle is mature, growth in profits is likely to slow further and quite possibly
turn negative, and that profit margins as a share of GDP will decline. In fact, our forecast of 1.5 per cent year-over-year growth in 2007
in the national income measure of pretax profits implies a slight drop in the profit share of GDP this year. (For the more widely
followed measure of S&P 500 GAAP earnings, bottom-up consensus estimates are for a 6.5 per cent increase in 2007 and an 11 per cent
increase in 2008.)
If margins returned to their average of the past decade, the scenario implicit in Wolf’s “cyclically adjusted earnings” calculation,
this would mean a drop of about three percentage points to 9.5 per cent of GDP. In terms of nominal dollars, this would mean at least a
15per cent - 20 per cent drop in profits, assuming the margin contraction occurred over 2-3 years of modest top-line growth. Profits would
need to fall further to maintain the long-term average at 9.5 per cent (since some period of time below this level would be necessary to
offset the time spent above it), and/or if the reduction in the profit share accompanied a sharp drop in top-line growth. Clearly,
such a scenario would be unlikely to occur without significant damage to the equity market.
However, it is by no means clear that profit margins need to fall so far. A look at the post-World War II history of profits suggests
Posted by: FT Economist Forum | March 14th, 2007 at 6:54 pm | Report this commenttwo distinct phases. In the first, a period of generally high growth and low inflation from the mid-1940s until the mid-1960s, the
profit share averaged roughly 11 per cent of GDP, and never fell much below 9 per cent. In the second, a period of generally weaker activity
and significant inflation from the late 1960s to the early 2000s, the profit share averaged a bit more than 8 per cent of GDP, and never
touched 11 per cent.
Andrew Smithers: Andrew Tilton questions whether the level of profits (or margins) needs to fall dramatically and implies that the stock market would not be seriously overvalued if they do not.
Economic theory suggests that profit margins in mature economies should be mean reverting and this is supported by the US data. Profit margins, defined as profits before depreciation, interest and tax, are currently 32% of output, compared with an average of 29%. As margins are strongly mean reverting around this level, it must be assumed to be their equilibrium level and thus the level to be used when calculating value.
An alternative way of calculating the cyclically adjusted PE is to calculate what earnings would be if margins were average. The overvaluation of the US stock market using this approach is virtually identical to that shown by the method pioneered by Professor Shiller.
Profit margins need to fall by 10% to be at their average level. Whether or not this would be a dramatic fall, it should be noted that those of equal or much greater size have regularly occurred in the past over periods of five years or less.
While it is probable that profit margins will return to their average, it is of course possible that they won’t. It does not, however, follow that this unlikely event would justify the current level of share prices. Even if margins were to stay improbably high, the market would remain just as overvalued as if they did not. The cost and return on equity must be in long-term equilibrium. If the return on equity is at a higher average level in the future than it has been in the past, then the future equilibrium PE must fall in proportion to the rise in profitability.
It is helpful in this context to consider the q ratio, which measures the level of the stock market without any reference to current or future earnings. Over the long-term and in equilibrium the cost of capital, including equity capital, must equal its return. Whatever the future return on net worth will prove to be, that must also be the rate at which future profits should be discounted. The present value of the corporate sector must be its current net worth.
Two general points should be emphasised. First, it must be illogical to use the past as a guide to fair value, for example with regard to PEs, while refusing to use the past as a guide to the future with regard to other ratios such as profit margins. Second, value does not depend on the likely level of profits in a year or three’s time, it is and must be an equilibrium concept.
Posted by: Andrew Smithers | March 19th, 2007 at 4:47 pm | Report this commentMartin Wolf: I do not want to add much to this already lengthy discussion. But let me make one further comment on what Andrew Tilton of Goldman Sachs has written. The ten-year period over which the most recent cyclically adjusted earnings per share was calculated was 1997-2006. This is indeed a period of relatively rapid growth and low inflation, similar to the mid-1940s to the mid-1960s. I do not see why one should regard the average real earnings of this already benign period as considerably below today’s sustainable equilibrium. If so, the current cyclically-adjusted P/E is indeed very high, as I argued in my column.
Posted by: FT Forum - Martin Wolf | March 20th, 2007 at 6:07 pm | Report this comment