U.S. Equities: Overvalued or Undervalued?
December 23, 2014
by Baijnath Ramraika, CFA® and Prashant Trivedi, CFA®
Advisor Perspectives welcomes guest contributions. The views presented here do not necessarily represent those of Advisor Perspectives.
“Be fearful when others are greedy, and greedy when others are fearful.” -Warren Buffett
“…. valuing the market has nothing to do with where it's going to go next week or next month or next year, a line of thought we never get into. The fact is that markets behave in ways, sometimes for a very long stretch, that are not linked to value. Sooner or later, though, value counts.” Warren Buffett
“The less prudence with which others conduct their affairs, the greater the prudence with which we should conduct our own affairs.” -Warren Buffett
“All intelligent investing is value investing - acquiring more than you are paying for. You must value the business in order to value the stock.” - Charlie Munger
As the U.S. equity markets continue to defy gravity, the chasm between those who suggest that U.S. equities are overvalued and those who believe they are undervalued is widening. The debate on market’s valuations is linked to the debate on U.S. corporate profit margins, a topic we discussed in one of our earlier articles.
Those who believe U.S. equities are undervalued put forward two key reasons: low interest rates and consequently low discount rates are in an environment characterized as the new normal; and that U.S. equities aren’t all that expensive on price-to-earnings basis. As we showed in our corporate profit margins paper, U.S. corporate profit margins are in the “wonderland” region and corporate earnings are susceptible to the mean reversion of profit margins.
As such, arguments in favor of U.S. equities being undervalued rest to a large extent on the assumption that corporate profit margins will stay elevated and on the continued omnipotence of central bankers.
Calculating fair value based on earnings that are far away from their sustainable level and assigning multiples to such earnings based on interest rates that have nothing to do with natural rates of interest is an exercise in futility; a folly similar to that of assigning value to shares of a company without any effort to value the business (refer to the addendum for a discussion of natural rates of interest and their significance for valuation).
Top-down valuation tools
An analyst has a variety of valuation tools at his/her disposal. In selecting one, an analyst needs to understand its applicability and limitations. All valuation models can be classified in two ways: asset-based valuation (ABV) tools and cash-flow-based valuation (CFV) tools.
In case of ABV tools, the analyst tries to assess the cost of reproducing the business. The underlying premise of such a valuation process is that the value of a business is related to the cost of reproducing the assets underlying the business. This relationship is driven by entrepreneurial actions such that where market valuations are well in excess of their reproduction costs, due to temporarily higher returns in excess of the cost of equity, entrepreneurs will establish businesses and attract capital. Once set in motion, this increased inflow (outflow) of capital drives market valuations back down (up) towards reproduction costs. ABV tools are particularly appropriate for use in businesses that have low barriers to entry.
In case of CFV processes, the analyst is trying to estimate the future cash flows that will be generated by the business. Here, a business’ value from the equity owner’s point of view is equal to the discounted value of future cash flows belonging to the equity owner. Key issues with such valuation models include sustainability of cash flows, the choice of discount rates and the derivation of appropriate growth rates. Given these issues, market participants have resorted to shortcuts, which are used as proxies for CFV processes. Price-to-earnings and price-to-sales ratios are examples of such shortcuts. CFV tools are particularly appropriate in cases where there are high barriers to entry.
Over the years, market participants have devised extensions to these tools to value overall equity markets/indices. Chief among these tools are Cyclically Adjusted Price to Earnings (CAPE) ratio, Tobin’s Q-ratio, the regression trendline, market capitalization-to-GNP1 (which is widely known as Buffett’s valuation indicator), and the trailing price-to-earnings ratio.
In the discussion that follows, we summarize each one of these valuation tools. We conclude this paper by offering our opinion on the market’s current valuation status.
Nearly three quarters of a century ago, Benjamin Graham advocated using average earnings as a way of approximating the earnings power of a company. Graham differentiated between earnings power based on past earnings and future earnings by terming them past earnings power and future earnings power. In his book, Irrational Exuberance, Robert Shiller expanded upon past earnings power by using 10-year average of inflation-adjusted earnings, a measure that is now widely known as CAPE or P/E10.
CAPE is in the CFV class of valuation tools, i.e., cash flow based valuation tools. By using a 10-year smoothing process, CAPE tackles the problem of sustainable level of cash flows related to the ebb and flow of business cycles. Figure 1 plots the CAPE ratio along with its long-term mean and standard deviation bands2 . Given the extreme nature of the dot-com technology bubble3, we have excluded CAPE ratios between 1994 and 2003 while calculating the mean and standard deviation bands. As is seen, CAPE ratio is currently two standard deviation levels above its mean.
Figure 1 4
- We have used the market cap-to-GNP (instead of GDP) ratio because GNP represents all outputs of U.S. residents. Given that the market valuations are of corporations that are resident in the U.S., using GNP as against GDP is a better proposition.
- There is a certain look-ahead bias in the way this chart has been setup. The mean and standard deviation bands have been calculated using all data available on the date of calculation and the same mean and standard deviation bands have been plotted historically. As all this data was not available in the past, the mean and standard deviation levels when calculated at any specific point in the past, using data that was available at that time, will be somewhat different. The implicit assumption here is that the ratio’s behavior doesn’t change when looked over the long-term, i.e., the bands would have been more or less the same in the past as well, if long-term data was used/available. However, this bias has no impact on the current readings / positioning of the model.
- As per our calculations, the technology bubble was a six standard deviation event based on CAPE. As of the end of the year 1993, the average CAPE ratio was 14.7 and standard deviation of CAPE was 4.6. While the CAPE peaked at 44.2x, the six sigma threshold was 42.3x.
- Data source: Robert Shiller, Online Data - Robert Shiller