Monthly Archives: August 2013

John Authers

Can CAPE guide us around the world? One reasonable complaint during the last week’s debate on cyclically adjusted price/earnings multiples is that the discussion is too US-centric. There are reasons for this. The US is still by far the world’s biggest stock market, the data are more reliable and go back further, and most of the academic players in the debate are based in the US. But it is still a reasonable complaint.

Here then are the results of the exercise in using multiples of 10-year rolling average earnings to value a range of world markets, as carried out by Mebane Faber of Cambria Investment Management, who kindly gave me his data. One huge caveat is that the data do not go as far back as for the US (although this at least means that we do not need to have arguments about whether it is possible to make comparisons with earnings from the late 19th century). The Faber data for the UK go back to 1927; none of the others go back further than 1969; and for some of the emerging markets the data only go back to the 1990s. The full details can be found on this post, and Mr Faber provided me with updated results to the end of July this year. Read more

John Authers

We all know that reported earnings are manipulated. But are they manipulated any more than they used to be, and are they manipulated so as to overstate profits, or understate them? And is the manipulation now so extreme that it is no longer relevant to compare profits over long periods of time? That is where the debate over CAPE (the cyclically adjusted price/earnings ratio) has reached. Even if it sounds technical, it is of vital importance when trying to work out whether the market is undervalued – as much of the stock broker community likes to argue – or in fact overvalue.

For those who missed them, the FT recently published my latest Long View column, defending the cyclically adjusted price/earnings ratio as calculated by Robert Shiller of Yale University and its relevance, quickly followed by a Market Insight column from Jeremy Siegel of the University of Pennsylvania’s Wharton School, and author of Stocks for the Long Run, arguing that Cape’s “overly pessimistic predictions are based on biased earnings data”. An academic conference on the subject is coming up in September.

Some navigation might be helpful. This is not just an arid academic dispute but a matter of critical interest to practising investors. As discussed last week, CAPE has been an impressive metric of value for over a century, and it sticks out from other metrics at present by signalling that stocks are badly overvalued (by 63 per cent for non-financials according to Andrew Smithers, a firm advocate of CAPEs as defined by Prof Shiller). Various different exchange-traded funds are now available that attempt to time switches between sectors based on their CAPE ratios.

However, the fact that CAPE is so bearish makes it unpopular. So does the undeniable fact that CAPE has been too bearish to be of great use to the average asset allocators over the last decade, failing to signal that stocks were cheap before two separate rallies, both of which saw stocks double over a period of four years. (You can see charts of CAPE over time in earlier LongShort posts here and here). Proponents of CAPE would counter that the measure is not for tactical asset allocation and that valuations cannot be used for timing the market.

So bulls are now trying to show either that Shiller’s CAPE was always flawed, or something has happened in the last decade or so to make the measure less useful. The academic participants know a lot of money is riding on this. Much more after the break. Read more

John Authers

One of the biggest areas of controversy over CAPEs or cyclically adjusted price/earnings multiples (see this earlier post and the huge correspondence it provoked) concerns exactly how earnings are measured, and how they can be compared over time. Neither is a unique problem for CAPE compared with other valuation metrics, but they can still change conclusions over the vital topic of whether the US stock market is now overpriced.

There is action on this front in the academic world, with Jeremy Siegel of the Wharton School proposing that an alternative measure of earnings should be used. This might change conclusions, and there will be more on that next week. For now, I would like to introduce another fascinating attempt to alter the methodology of Yale’s Robert Shiller, who has been central to popularising the CAPE over the last two decades.

Back in 2009, Alain Bokobza of Societe General released the results of his own new normalised CAPE, and his colleagues have kindly updated his data. The essential new insight was that corporate taxation has not been constant over the years. Rather, it was introduced in 1911 at only 1.5 per cent. After the First World War it rose to 15 per cent, and then, as the apparatus of the welfare state rolled out over the ensuing decades, it rose to 40 per cent. Mr Bokobza contends that this affects the multiple that investors will pay. So he recalculated the CAPE for the years before 1950, assuming a 40 per cent tax rate.

To give a quick example; If you pay $100 for $10 of earnings untaxed, you have paid a p/e of 10. If it is taxed at 40 per cent, you have paid a multiple a little above 16 on post-tax earnings. This comparison may more accurately, according to Mr Bokobza, help build a norm for what investors are prepared to pay for the earnings they buy. Without making such an adjustment, Mr Bokobza contends, we are effectively comparing post-tax earnings post-war with what can almost be called pre-tax earnings pre-war. This is a contentious point of view; throughout the period, taxation was a given for investors. Many other factors were also changing. But it makes a case that pre-war multiples may not be directly comparable to post-war ones, and suggests an intuitive fix. (And indeed the comparability of earnings is the subject of growing academic debate, and is very relevant when trying to get a handle on long-run valuations). A look at how this changes the historical picture comes after the break. Read more

James Mackintosh

Christmas has come early for US traders – Christmas trading levels, that is. Volumes in the benchmark S&P 500 index are down to levels only seen in the last five years in Christmas week. Perhaps investors are just determined to enjoy their time at the beach after having to deal with a crisis every summer for the past three years.

Whatever the cause, the volumes are miserable. The thick line here shows the five-day moving average, smoothing out the daily swings in the thin line. The five-day average has just dropped below the low point of last August and is now the lowest apart from Christmases since Bloomberg’s data series started in 2008. Read more

John Authers

Cyclically-adjusted price/earnings multiples (CAPEs), as made famous by Yale’s Professor Robert Shiller, are growing inconvenient for the brokerage community.

Last week, BofA Merrill Lynch’s Savita Subramanian pointed out that of 15 popular measures of equity valuation, CAPE (which compares share prices to a 10-year moving average of real earnings) was the only one that made stocks look expensive. The list of valuations suggesting US stocks are either cheap or at fair value includes:

  • trailing p/e
  • forward consensus p/e
  • trailing normalised p/e
  • price/book
  • enterprise value/ebitda
  • forward PEG (p/e ratio divided by growth)
  • trailing PEG
  • price/operating cash flow
  • price/free cash flow
  • enterprise value/sales
  • market-based equity risk premium
  • normalised equity risk premium
  • S&P 500 in WTI oil terms
  • and S&P 500 in gold terms

CAPE is thus beginning to stick out like a sore thumb. As it has been showing that stocks are expensive throughout most of the current rally, there is now a widening attempt to discredit or ignore it. Merrill’s own complaint is typical:

The Shiller P/E, which is based on inflation-adjusted earnings over the past 10 years, currently suggests that stocks are overvalued. However, this metric
assumes that the normalized (cyclically-adjusted) EPS for the S&P 500 is today less than $70—well below even our recessionary scenario for EPS. The
methodology assumes that the last 10 years is a representative sample, but the most recent profits recession was the worst we have seen and was exacerbated by a high leverage ratio which has since been dramatically reduced. Assuming that this scenario is going to repeat itself is, we think, overly pessimistic

But is it? Earnings volatility has certainly been extreme over the last decade, and arguably unprecedented. But if anything that suggests that the measure – which grew famous from efforts to predict the bursting of the dotcom bubble in 2000 and to show that the 2003-07 bull market was a “fools rally” – is more, not less, useful. That is the contention of Prof Shiller himself, and it is a reasonable one. More on this, with charts, and some comments from Prof Shiller, after the break. Read more

John Authers

Let’s try to drill into the global picture for earnings, following on from Monday’s column. The picture is undeniably unexciting, but we need to take two divisions into account. First, financials have followed a different logic from the rest of the corporate sector over the last five years or so, for obvious reasons. Second, US companies have profited far more than companies elsewhere in the world, for reasons that are far less clear.

Let’s start by looking at the global picture, with and without the financials. Earnings per share for the MSCI indices (with thanks to Andrew Lapthorne of SocGen for providing this data) look like this: Read more