7,2 8,4 9,2 9,8 10,3 10,8 11,3 11,8 12,2 12,7 13,2 13,6 14,1 14,6 15,2 15,9 16,6 17,5 18,6 20,0 22,2 25,4 33,0 2.304,0 MARCH 7, 2014 Stocks: Low valuation pays off We made the case here a few weeks ago that investments in stock markets with low valuations would pay off if a sufficiently long investment horizon was chosen (see Economic Situation and Strategy Good news for value investors: Long-term payoff from low valuation of February 13, 2014). However, some investors prefer to buy individual stocks. We therefore demonstrate in this second part of our series on valuation how they can use valuation ratios to improve medium-term performance. In the first section of this report, we describe our method of calculation, which should interest our technically inclined readers. This may go too far for those concerned mainly with the results, but we recommend they read the first two paragraphs on method before going on to the results in the second section. 1. Method We used monthly data from the Stoxx 660 index for our calculations, starting in January 2000. We first assumed stocks would be held for five years. Stocks bought in January 2000 were sold in January 2005, those bought in February 2000 were sold in February 2005, and so on. The last stocks considered in our calculations were bought in February 2009 and sold in February 2014. 15% 5% -5% - Stoxx 600 stocks: PE and performance p.a (buying criteria: Stoxx 600 membership; median; holding perdiod 5 years) We then compared various valuation ratios at the time of purchase with the stock s performance after five years (including reinvestment of dividends). To recognize what influence valuation had on performance, the ratios are divided up into several equal sections in ascending order. In our chart, the bar on the far left depicts the average performance of stocks that had a price-earnings ratio (P/E) of less than 7.2. The bar to the right of it, marked 8.4, represents the average performance of stocks that had a P/E of less than 8.4 but at least 7.2 at the time they were bought. It is important to note that the composition of the Stoxx 600 has changed considerably since the starting point. That is partly because companies have merged, their market capitalization has decreased in the meantime, or they have been taken off the stock exchange. Consequently, there are different companies over time in the selection universe of the Stoxx 600. As we think it would be handled on a realistic investment strategy, we therefore assumed that only companies are considered at the time of purchase that were also in the index at that time. 1 Concretely, that means for stock purchases in February 2003, for example, we only considered stocks that were listed in the Stoxx 600 at that time. If the company left the Stoxx 600 index after the purchase, the stock nevertheless remains in the portfolio until the end of the five years. If the stock is listed on the stock exchange for less than five years, its performance is taken into account up to the last available month. 2. Results Our calculations show that stocks with low valuations perform significantly better than those in the middle or upper valuation segments. That is true at least of the often-used valuation ratios that we have tested, such as the P/E, the price-book value ratio, and the dividend yield (each based on estimates for the coming 12 months). It may therefore be sensible for investors oriented to the medium term (we have assumed a holding period of five years) to give high priority to selecting stocks with low valuations. In most cases, the performance of low-valued stocks has been better not only when calculated on a classic mean average, but also on a median basis. That is good news for investors who give valuation ratios a high status. For, good (mean) average performance of the overall portfolio that we calculate for low-valued stocks can also be attributable to exceptionally good performance of individual stocks, while the majority of stocks have performed generally very weakly. However, since even relatively large investors are usually never able to buy all low-valued stocks, the prospects for investors success also depend on whether many stocks with low valuations show above-average performance. The median analysis shows that the number of stocks with very good performance is even usually higher for low-valued stocks than for high-valued ones. Generally, investors who practice a style that bets on low valuations can use both estimates and realized earnings to value stocks. In the case of realized earnings, the P/E is then calculated with actual earnings of the last 12 months, and the price-book value ratio with the book value of the 1 It would lead to incorrect results if one based the back-calculation solely on the companies in the Stoxx 600 today. That is connected with what is called the survivorship bias, i.e., that only those companies remain in the index that have survived and have had some business success, at least as measured by market capitalization. Investors have at least certain reservations about the business models of companies that are valued very low. If such companies are still in the index years later, then it is very likely that the concerns about their business models were exaggerated. However, this means that especially those once low-valued companies still listed in the index today exhibit aboveaverage performance. The performance of stocks with low valuations would thus be hugely overestimated if the composition of the index today were used as the starting point for the calculation. Moreover, this approach would not be realistic for investors who had followed our valuation strategy. After all, the information that a company was still in the Stoxx 600 investment universe in January 2010 was not available to investors in 2000. 1
0,20 1,17 1,68 2,15 2,60 3,08 3,64 4,35 5,53 1809,45 5,3 7,3 8,6 9,5 10,3 11,0 11,6 12,2 12,9 13,5 14,2 14,9 15,6 16,5 17,4 18,5 19,8 21,3 23,4 26,5 33,0 54,5 651,9 152.614.900,0 last 12 months. But does our conclusion also stand if one determines the valuation of the stock in terms of realized earnings instead of estimates? We also tested that. Our calculations show that stocks with realized book values and earnings likewise generally show better performance in the medium term than stocks that are valued high on this basis. Overall, however, the connection between low valuation and above-average performance appears to be somewhat more pronounced when estimates are used than in the case of realized earnings. Moreover, P/Es show a stronger connection than price-book value ratios. 12% 8% 6% 4% 2% -2% Stoxx 600 stocks: Trail 12m PE and performance p.a (buying criteria: Stoxx 600 membership; median; holding period 5 years) A peculiarity of the results presented so far is that the number of especially low-valued stocks in the overall market is not always at the same level. Instead, the valuations change even for whole indexes they are generally higher in times of economic strength and lower in crises. Up to this point, our results also show to a certain extent that investments sometimes (including in crisis phases) entail better performance than investments at other times (e.g., at the peak of a boom). Investors can certainly use this conclusion. Individual investors can seek heavier exposure in stocks during economic weakness. Flexible investors can increase the stock ratio of their portfolios when low-valued stocks are plentiful on the overall market. But are our results also useful to investors who must always be largely or even 10 invested in stocks? We tested this question, and the answer is generally affirmative. Even investors who are always invested in stocks achieve better performance in the medium term when they bet primarily on low-valued stocks. To show this, we have changed our method of calculation somewhat and have divided stocks each month into equal classes. The stocks that are valued low according to the relevant valuation ratio are in Class 1, the somewhat higher-valued stocks are in Class 2, and so on. The advantage of this analysis is that there is a constant number of stocks that fall into Class 1 and are accordingly valued very low even in times of economic strength. 14% 12% 8% 6% 4% 2% Stoxx 600 stocks: PE and performance p.a (buying criteria: Stoxx 600 membership; average; holding period 5 years) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Valuation class (lower values represent lower valuations) As our chart shows, the average differences in returns between stocks of high and low valuation are by far no longer as large for someone who invests more in a whole market that is valued low. Nevertheless, significantly better performance may still be achieved on average if investors bet on low-valued stocks. In the case of investors 10 exposed in stocks, however, this result depends much more heavily on individual stocks that exhibit very strong performance. On the other hand, the differences are not very pronounced in the median, and given a holding period of five years, the number of low-valued stocks with good performance is only negligibly above the number of highvalued stocks with good performance. Here, investors 10 exposed in stocks are at a disadvantage relative to those who can invest more or less heavily in stocks over long periods of time, as needed. Both strategies that of the flexible investors and that of the 10 investors are successful even at holding periods shorter than five years. A holding period of two years, or even one, is enough for investing successfully. This applies to all ratios. A shorter holding period actually seems more suitable for 10 investors. In contrast, holding periods longer than five years no longer work, at least not with all ratios. In particular, in the case of the price-book value ratio, an advantage for low-priced stocks seems no longer clearly to result on an investment of significantly more than five years. The chart below illustrates the performance of stocks classified according to dividend yields expected at the time of purchase. The holding period is two years. As in the case of the other valuation ratios, a high dividend yield at the time of purchase (that corresponds to a low valuation) leads to above-average performance. 25% 2 15% Stoxx 600 stocks: dividend yield and performance p.a (buying criteria: Stoxx 600 membership; average; holding period 2 years) 5% -5% - -15% -2 2
9,3 11,1 12,6 14,1 15,7 17,9 20,8 25,4 33,0 49,0 1125,2 6,91 8,03 8,81 9,42 9,96 10,45 10,90 11,35 11,81 12,27 12,74 13,21 13,67 14,17 14,72 15,32 16,01 16,86 17,91 19,29 21,23 24,48 32,11 9108,06 9,9 11,3 12,4 13,2 14,1 14,9 15,9 17,2 18,9 22,2 359,0 More stocks are hidden behind the individual bars than on the other strategies. The background is that there is nothing to sensibly distinguish among stocks with a dividend yield of zero, so the size of the individual groups must be chose in such a way that all stocks with a dividend yield of zero fall into one group. The extremely high value of the bar on the far right is attributable to two factors: financial stocks during the financial crisis of which analysts nevertheless expected that those companies would pay a dividend (this assessment proved incorrect as a rule) and companies that were subsequently taken off the stock exchange. However, some important points remain open for the interested investors who want to follow a corresponding strategy. First is the question whether the quite clear connection between low valuation and above-average performance might be attributable to just a few years in which the strategy worked far above average. If it were only an outlier in financial market history, then an investment strategy according to valuation would presumably not be very promising for the future. However, we can dispel some of the concerns in this regard. Admittedly, the approach presented here of investing in low-valued stocks would not have continuously led to investment success. Nevertheless, an analysis of investment successes broken down by investment years permits the conclusion that the connection between low valuation and good performance appears to be basically stable. We have calculated this on the basis of the P/E ratios (estimates for the coming 12 months) and reproduced the investment successes of a strategy based on valuation for each year. Given a holding period of five years, as we assumed at the beginning, one would have registered greater investment success on average with low-valued stocks in the case of stocks purchased in 2000 to 2004. Above all, however, the impression is confirmed that the connection between performance and valuation is largely stable across the entire valuation spectrum. 2 Stoxx 600 stocks: PE and performance p.a, buying period 2000 (buying criteria: Stoxx 600 membership; average; holding period 5 years) holding period of five years, the strategy would thus have worked in seven years and not worked in three. 3,5% 3, 2,5% 2, 1,5% 1, 0,5% 0, -0,5% -1, Stoxx 600 stocks: PE and performance p.a, buying period 2006 (buying criteria: Stoxx 600 membership; average; holding period 5 years) If the holding period is shortened to two years, as would be sensible in any case for 10 investors, the number of years with below-average performance then decreases further and falls to two years versus eleven years in which the strategy would have worked as expected. If this corresponds to the facts, where is the catch in such a strategy? One disadvantage of low-valued stocks is certainly that they exhibit much greater volatility and make for considerably larger fluctuations in portfolios. Ultimately, in the case of extremely low-valued stocks, investors are betting at least to a certain extent that the business trend of the company in question will take a turn for the better. That necessarily involves greater risks, which are reflected systematically in higher volatility in nearly all our analyses. 35% 3 25% 2 15% 5% -5% - -15% Stoxx 600 stocks: PE, performance and volatility p.a (buying criteria: Stoxx 600 membership; average; holding period 2 years) 45% 4 35% 3 25% 2 18% 16% 14% 12% 8% 6% 4% 2% However, low-valued stocks led to below-average performance in the case of investments from 2005 through 2007. That is probably due in part to the financial crisis, which led to considerable price losses in some sectors, and those have still not been made up in all cases. Whoever bought stocks in 2005, would have sold them in 2009, in some cases at their lowest levels. On the other hand, this strategy worked again for stocks bought in 2008 and 2009. With a performance p.a. volatility p.a. (r.h.s.) This result also agrees with the study of the stock indexes that we carried out a few weeks ago. The difference in volatility (risk) can be quite considerable depending on the length of the holding period. Investors that mainly seek exposure preferably in very low-valued stocks must have strong nerves and should be clearly aware that this is absolutely not a conservative investment strategy. It is therefore advisable for many investors to buy stocks from the middle valuation segment, where risks seem much lower, but according to our calculations, better performance than in the case of high-valued stocks may nevertheless be expected in the medium term. This is easy to see in the chart above. Already at P/E ratios of just over 10, volatility is much lower than in the case of stocks that are valued very low. Nevertheless, it has also been possible to achieve above-average performance with stocks from the moderate- 3
ly low valuation segment. How willing investors are to buy mainly stocks with very low valuations depends on many factors, including risk tolerance, portfolio design, and investment horizon. Ultimately, it is likewise an important consideration how many stocks at least must be bought and put into a portfolio. For our analyses, we have evaluated the performance of very many stock purchases (more than 45,000, with a holding period of five years). Among them are, of course, also individual stocks that have performed very poorly. Investors who implement a valuation-based strategy with very few stocks would be affected more than average by the extremely poor results of a few stocks. To provide a point of reference as to how many stocks at least should be held, we have simulated how dependent risks of loss are on portfolio size. As to be expected, the risk of a negative return here is much higher for small portfolios than for large ones. With increasing diversification, however, the risk decreases less and less. Ultimately, of course, it is not possible to state universally how diversified a portfolio must be. Nevertheless, it appears that respectable diversification effects can already be achieved with portfolios consisting of 15 to 20 stocks. 4
Jan. 00 Jan. 02 Jan. 04 Jan. 06 Jan. 08 Jan. 10 Jan. 12 Jan. 14. Weekly outlook for March 10-14, 2014 Sept. Oct. Nov. Dec. Jan. Feb. Release DE: Exports, m/m 1.4% 0.3% 0.7% -1. 1.2% March 11 DE: Exports, y/y 0.8% 0.7% 3.8% 2.2% 2.3% March 11 DE: Imports, m/m -1.9% 3. -1.2% -1.4% 0.8% March 11 DE: Imports, y/y -2. -1.5% 1.3% 1. -1. March 11 DE: Trade balance, in EUR bn 18.5 16.5 18.0 18.1 18.7 March 11 DE: Consumer prices, m/m 0. -0.2% 0.2% 0.4% -0.6% 0.5% March 14 DE: Consumer prices, y/y 1.4% 1.2% 1.3% 1.4% 1.3% 1.2% March 14 EUR18: Industrial production, m/m -0.2% -0.7% 1.6% -0.7% 0.5% March 12 EUR18: Industrial production, y/y 0.2% 0.4% 2.6% 1.2% 1.9% March 12 MMWB estimates in red Chart of the week: Global upswing gaining strength 65 Global Purchasing Manager Indexes 60 55 50 45 40 35 Services Manufacturing Despite all the political bad news, the purchasing manager indexes for the manufacturing and services sectors released in the past few days show that the economic environment continues to improve in most countries. In the manufacturing sector, the average that we calculate for more than 30 countries reached 53.0 points in February, the highest level since June 2011. The average for the services sector based on the data that we analyzed from 14 countries was 53.4 points, also the best since summer 2011. It is good to see that the crisis in the emerging markets, which gripped the capital markets in January, has shown hardly any negative effects. For example, the services sector has recovered lately in all four BRIC countries, while the picture in the industrial sector has been mixed with two improvements and two worse results compared with the preceding month. The upswing also continues in the euro zone, led by Germany, which will be the region s economic engine again this year. The global purchasing manager indexes would have been even better if the harsh winter in the United States had not heavily burdened businesses there. It is thus possible that many economic forecasts will prove too conservative and will have to be revised upward during the year. However, whether the conflict between Russia and Ukraine has a negative impact in the months ahead remains to be seen. 5
As of Change versus 03.07.14 28.02.2014 06.02.2014 06.12.2013 31.12.2013 Stock markets 16:53-1 week -1 month -3 month YTD Dow Jones 16422 0,6% 5,1% 2,5% -0,9% S&P 500 1877 0,9% 5,8% 4, 1,6% Nasdaq 4352 1, 7,3% 7,1% 4,2% DAX 9543-1,5% 3,1% 4, -0,1% MDAX 16876-0,1% 3,3% 5,8% 1,8% TecDAX 1293 0,5% 5,9% 14,2% 10,8% EuroStoxx 50 3145-0,1% 4,4% 5,5% 1,1% Stoxx 50 2945-0,8% 3,7% 4,1% 0,9% SMI (Swiss Market Index) 8484 0,1% 3,2% 5,2% 3,4% Nikkei 225 15135 2, 6,9% -1,1% -7,1% BOVESPA 47093 0, -1,4% -7,6% -8,6% RTS 1165-8,1% -12,5% -16,3% -19,3% BSE 30 21514 1,9% 5,9% 2,5% 1,6% China Shanghai Composite 2060 0,2% 1,3% -7,9% -2,7% MSCI Welt (in ) 1686 1,1% 3,9% 3,9% 1,8% MSCI Emerging Markets (in ) 970 0,9% 2,5% -3,8% -2,9% Bond markets Bund-Future 144,47 9 114 447 530 Bobl-Future 126,82 12 45 243 239 Schatz-Future 110,56 2 2 20 26 3 months Euribor 0,29 0 0 4 0 3 months $ Libor 0,24 0 0-1 -1 10 year US Treasuries 2,74 8 4-14 -27 10 year Bunds 1,65 2-4 -19-29 10 year JGB 0,60 1-2 -7-14 US Treas 10Y Performance 519,22-0,6% 0,1% 2,2% 3,1% Bund 10Y Performance 513,28-0,2% 0,5% 2,9% 3,7% REX Performance Index 450,16-0,1% 0,2% 1,6% 2,2% US mortgage rate 4,37 0 14-9 -11 IBOXX AA, 1,79 2-5 -21-27 IBOXX BBB, 2,49 3-15 -32-33 ML US High Yield 6,11 1-27 -27-27 JPM EMBI+, Index 665 0, 2,7% 3,1% 2, Convertible Bonds, Exane 25 6265 0, 2,1% 4,1% 2,3% Commodities CRB Index 567,40 3,1% 8,5% 11,7% 11,7% MG Base Metal Index 323,59 1,6% 3, 1,9% -1,7% Crude oil Brent 108,47-0,4% 2,2% -2,8% -2,8% Gold 1345,82 1,5% 7,1% 8,9% 11,4% Freight rates Baltic Dry Index 1480 17,6% 35,5% -32, -35, Currencies EUR/ USD 1,37-0,5% 1,9% 0,6% -0,3% EUR/ GBP 0,83 0,5% -0,6% -1,1% -0,5% EUR/ JPY 141,22 0,4% 3,1% 1,1% -2,4% EUR/ CHF 1,22 0,3% -0,3% -0,3% -0,7% USD/ JPY 103,07 1,2% 0,9% -95,8% -2,1% Carsten Klude +49 40 3282-2572 cklude@mmwarburg.com Dr. Christian Jasperneite +49 40 3282-2439 cjasperneite@mmwarburg.com Matthias Thiel +49 40 3282-2401 mthiel@mmwarburg.com Martin Hasse +49 40 3282-2411 mhasse@mmwarburg.com Darian Heede +49 40 3282-2452 dheede@mmwarburg.com This document does not constitute and shall not be construed as an offer or an invitation to make an offer. It may only be used as guidance and to illustrate potential business activities. No claim is made as to the exhaustiveness of the information contained in this document, and it is therefore non-binding. The opinions expressed here may change at any time without notice. Any statements made about prices or interest rates or any other indications that are given relate exclusively to the time that the document was produced and do not contain any statements about future trends or, in particular, about future profits or losses. In addition, this document does not constitute and shall not be construed as advice or a recommendation. Before concluding any transactions presented in this document, you should always obtain client- and product-specific advice. 6