The Goldilocks P/E ratio, not too high and not too low
In this Part 1 of the series/ I discuss the Goldilocks P/E ratio, not too high and not too low. Reluctantly, I have to respectfully disagree with some of what Chuck Carnevale has written recently about P/E ratios (see here).
The first area of disagreement is that for an investor, "... a 15 P/E ratio represents an earnings yield of approximately 6.66%. A P/E ratio of 20 represents an earnings yield of 5% and a P/E ratio of 30 represents an earnings yield of 3.33%."
The second area of disagreement is Chuck's contention that applying a P/E of 15 to most companies is not overly simplistic.
Beware Buy and Sell Decisions based on A Goldilocks P/E ratio
I am concerned with the message conveyed in the above article that for most companies applying a historical average P/E ratio of 15 is appropriate and not overly simplistic. Chuck Carnevale qualifies that message throughout the article, but it is a very long article containing F.A.S.T. graphs for over 30 companies with very valuable and insightful qualifying comments accompanying the graphs.
Many readers will take on board that initial message and not read through the whole article, and so fail to learn there are many qualifications. I believe the F.A.S.T. graphs are an enormously valuable resource, providing "a picture that is worth a thousand words." But proposing a 15 P/E ratio blinkered view of those pictures is a waste of those resources.
There is so much more to be learned from the representation of historical P/E levels in those graphs. With some additional parameters and guidance, the view provided by these graphs can be greatly enhanced, with very little extra effort on the part of the viewer.
At this point, before readers might lose interest, I will provide some very real and practical examples of why buying and selling shares at the same or similar P/E ratios is too simplistic. Intuitively, if I buy shares in a company at a P/E ratio of 15.0, i.e., at a yield of 6.67%, and sell those same shares some years later at a P/E ratio of 15.0, I will achieve a yield of around 6.67%.
In practice, that is a low probability, and in fact, the actual yield could be far higher or far lower, and even negative. To demonstrate this with real data, I have utilized data from the F.A.S.T. graphs for the 8 regional bank stocks included in Chuck Carnevale's recent article linked above. The 8 banks are 1st Source Corporation (NASDAQ:SRCE), The First Financial Corp. (NASDAQ:THFF), BancFirst Corporation (NASDAQ:BANF), Bank of the Ozarks, Inc. (NASDAQ:OZRK), Bar Harbor Bankshares (NYSEMKT:BHB), Bank of Marin Bancorp (NASDAQ:BMRC), Prosperity Bancshares Inc. (NYSE:PB), and First of Long Island Corp. (NASDAQ:FLIC).
TABLE 1 below compares outcomes for the 8 banks based on buy and sell at a P/E ratio of 15.0.
TABLE 1 above, shows I definitely cannot buy and sell at a 15 P/E ratio and expect an earnings yield of approximately 6.66%. If I had bought 1st Source Corporation shares in 2005 at a P/E 15 ratio, and sold in 2009 also at a 15 P/E ratio, I would have lost an average of 9.3% per year on my initial investment. In a different time period, if I had bought 1st Source Corporation shares in 2009 at a P/E 15 ratio, and sold in 2015 also at a 15 P/E ratio, I would have gained an average of 15.8% per year on my initial investment, far above the 6.67% yield supposedly indicated by a 15 P/E ratio.
The fact is, if share price is equal to P/E multiplied by EPS, and shares are both bought and sold at a fixed P/E of 15 (or any other fixed P/E at entry and exit), the only thing that can result in an increase or decrease in share price, between purchase and sale, is growth of or diminution in EPS. In the scenarios described above, P/E is irrelevant to considerations as to the yield achievable.
The investor must make an assessment of expected growth in EPS going forward to determine if the investment is likely to achieve the expected yield. (Dividends and future dividend growth are equally as important and must also be considered. But that requires a separate discussion, which does not alter the principles discussed here).
P/E ratio at entry and exit does matter
As discussed above, if we buy and sell shares at a P/E of 15, profit on sale is limited to any increase due to EPS growth and any possible P/E impact is avoided. But if we buy and sell at different P/E ratios, there is definitely an impact. To better illustrate this, I have taken from Table 1 above, the High and Low and PE15 based share prices for years 2009 and 2015 and incorporated these into TABLE 2 below.
These High and Low share prices represent the highest and lowest P/E ratios for each of the years in question. In TABLE 2, I show the comparative shareholder returns for each of the 8 banks from buying at the highest P/E ratio in 2009 and selling at the lowest P/E ratio in 2015. Similar calculations are provided for buying at the lowest P/E ratio in 2009 and selling at the highest P/E ratio in 2015, and for buying at a 15 P/E ratio in 2009 and selling at the highest P/E ratio in 2015.
The results in TABLE 2 demonstrate the magnitude of difference in returns from a higher or lower P/E ratio at entry. To achieve either of these results would necessarily involve a tremendous amount of bad or good luck. On the other hand, the third calculation, buying at a P/E of 15 in 2009 and selling at, not necessarily the highest PE in 2015, but at a higher than 15 P/E seems much more achievable.
This maybe is why Chuck Carnevale suggests screening with a 15 P/E ratio. It is virtually impossible to achieve ideal timing on entry and exit from a shareholding. If one buys at a historically high P/E ratio for a particular company it might be very difficult or impossible to exit at that P/E level, resulting in an adverse impact on returns due to a lower P/E ratio at exit.
Conversely, it is very difficult to time entry at the lowest P/E level. Using a 15 P/E ratio as a guide to market entry might not achieve the lowest entry P/E ratio ultimately available. But it certainly gives a very good opportunity to selectively exit at a higher P/E ratio, with a resulting additional gain over and above any gains from EPS growth.
It can be seen from TABLE 2 above, buying at the lowest P/E ratio and selling at the highest P/E ratio can provide outstanding gains. The converse is true for buying high and selling low. But for most of us, we will never achieve perfect market timing, or anywhere near it. Chuck Carnevale's suggestion of looking for market entry around a 15 P/E ratio gives a particular advantage.
It gives a basis for market entry at a time where we do not know whether PE ratios might go up or down soon after we buy. But with a 15.0 P/E ratio locked in at purchase of shares, it is likely we will see those shares subsequently priced both below and above a 15 P/E ratio. So we have the ability to exit opportunistically, at a higher P/E ratio than at entry, giving rise to an enhancement in returns, compared to what might otherwise be achievable. But that is not the whole story.
Buying at a Low P/E and selling at a High P/E does not guarantee a gain; nor does buying at a High P/E and selling at a Low P/E necessarily mean a loss will be incurred
Just as I have done above to disprove the relationship between P/E and yield for an investor, I will provide some historical examples to verify that buying at a high P/E and selling at a low P/E does not necessarily mean a loss will be incurred, and vice versa. Various Authors have been suggesting Johnson & Johnson (NYSE:JNJ) is fully valued or overvalued based on the current P/E ratio versus its long-term average. I will say to those Authors,
"Oils Ain't Oils" and "PEs Ain't PEs."
To demonstrate this, I have collated high and low share price and EPS data for various Dividend Aristocrats for the past 20 years to see what the impact of P/E ratios has had for investors. The first of these is one of my favorites, Johnson & Johnson. The current P/E ratio for JNJ is over 26 while its average P/E ratio from 1997 to now is around 18.7.
I have divided my analysis into three overlapping periods to better illustrate my point that P/Es are not necessarily directly comparable. This is particularly so for different periods in time, making the use of long-term averages unhelpful.
Review of Johnson & Johnson P/E ratio history 1997 to 2016
TABLE 3.1 - Johnson & Johnson P/E ratio history 1997 to 2005
From TABLE 3.1 we can see that in the period 1997 to 2005 JNJ P/E ratios ranged from a high of 44.9 in 1998 to a low of 17.8 in 2005. Despite this massive drop in P/E over 7 years, an investor, buying at the high 44.9 P/E and selling 7 years later at the low 17.8 P/E, would have realized a gain on sale of the shares equivalent to a return of 3.6% per year. The return on investment including dividends would be 5.6%.
Furthermore, desperately unlucky investors who managed to buy only at the Highest P/Es in the 7 years 1997 to 2003 would, for 5 of the 7 years, have been able to exit with a gain on sale (excluding dividends) even if they sold their shares at the very lowest P/E in the year of sale. Of particular note is buying at a P/E of 34.2 in 2000 and selling at a P/E of 17.8 in 2005 resulted in a 4.2% average yearly return, whereas buying at a much lower P/E of 25.8 in 2003 and also selling at the above PE of 17.8 only gave a 2.6% average yearly return.
The high growth rates for EPS and dividends meant that an investor's 'current yield on cost' and 'share cost price to current earnings' quickly improved over time. TABLE 3.1 indicates, all in all, it was quite a good time to be investing in JNJ. But any investor who used a P/E of between 15 and 20 as a guide to fair value, above which they would not invest, would have completely ruled out JNJ as an investment between 1997 and 2003.
P/E ratios reflect the relationship between share prices and EPS, they are not a driver of share prices and/or EPS. To understand P/E, first understand the drivers of EPS and share price.
To understand the relevance of any level of P/E, we need to understand the potential future drivers of share price and EPS. It is fairly obvious from TABLE 3.1 above that expectations during the years 1997 to 2005 of continuing high EPS and Dividend growth were driving share prices higher, ahead of the actual growth in EPS and Dividends, resulting in high P/E ratios.
TABLE 3.2 - Johnson & Johnson P/E ratio history 2005 to 2016
Comments TABLE 3.2 - 2005 to 2012
Unlike in the years 1997 to 2003, TABLE 3.2 reveals that an investment in JNJ shares at times of high P/Es in the years 2005 to 2012 would prevent exit at the lowest P/Es for up to 8 years, otherwise a loss on sale of the shares would be incurred. This was despite average P/E of 16.1 for this 8-year period, with a high of 20.89 and a low of 10.51. Low EPS yearly growth of 2.0% (1997-2005 16.8%) together with increased EPS volatility, and reduced dividend growth rate of 9.5% (1997-2005 14.8%), are key reasons for the far lower P/Es in this period.
So, the investor who only invested in JNJ when the P/E was around its long-term average would not have invested during the 1997 to 2003 period of very high P/Es when positive results would have been achieved. But they would have invested in the 2005 to 2012 period when P/Es had fallen, with possibly significant losses resulting.
Of particular note is even investing at the very lowest P/Es and exiting at the very highest P/Es would not achieve the levels of returns in the 2005 to 2012 period that were possible in the 1997 to 2005 period. I have brought forward part of TABLE 3.1 and appended it to the foot of TABLE 3.2 to more readily illustrate this. It can be seen that an investment in 1998, at a P/E of 31.7, followed by an exit 7 years later in 2005, at a much lower P/E of 20.9, would have resulted in an average yearly return of 13.8%.
But an investment in 2005, at a P/E of 17.8, followed by an exit 7 years later in 2012, at a higher P/E of 18.8, would have resulted in an average yearly return of just 5.8%, 8 percentage points lower than for the 1997 to 2005 period. Even an entry at a P/E of 11.4 and exit at a P/E of 18.8 in the 2005 to 2012 period cannot match the returns available in the ultra high P/E period of 1997 to 2005.
So I repeat, "Oils Ain't Oils," and "P/Es Ain't P/Es." But do note the effect on dividend yields of achieving a low P/E entry. At an 11.4 P/E entry in 2008 and exit at 18.8 in 2012, the contribution of dividends to total return is 3.7% (12.4% minus 8.7%). This 3.7% yield is above the average dividend yields shown in the table and is totally due to the low P/E at entry, and nothing to do with the P/E at exit. So for an investor purely interested in the dividend yield, a low P/E at entry assumes a high level of importance.
Comments TABLE 3.2 - 2012 to 2016
We can also see for the 2013 to 2016 period, entering at a high P/E level from 2013 onwards, would prevent exiting at a low P/E level, without incurring a loss. Average yearly dividend growth percentage has continued to decline. But average EPS growth rate has picked up after the lows of the 2005 to 2012 period. This would be one of the drivers of the increasing P/E levels in this period. It is not possible to make a judgment on current 2016 P/E levels for JNJ, without making some assessment of many factors that will or could impact on future EPS and share prices.
I trust I have by now conveyed a sense among readers that past historical P/Es, by themselves in isolation, are of little use in decisions whether or not to invest in JNJ. JNJ is just one company and its own particular experiences will have influenced historical P/E levels. SA Contributor, Lasrman, commented on a recent Chuck Carnevale article, " When I look at your fast graphs it seemed like the "bond-like" securities such as PG and KO tracked closer to the interest rate trend." I believe the above analysis for JNJ could be useful for confirming if there is a common trend in PEs for any or all periods 1997 to 2016. On this basis, I have completed similar analyses to JNJ for Procter & Gamble (NYSE:PG), Coca-Cola (NYSE:KO), and also McCormick (NYSE:MKC).
Review of Coca-Cola P/E ratio history 1997 to 2016
TABLE 4.1 - Coca-Cola P/E ratio history 1997 to 2005
As alluded to in the note in the body of TABLE 4.1, if you bought KO shares during this period at a year High P/E there would have been very little opportunity to exit those shares without making a loss on sale. But if you did buy at a particular P/E, and you could find a time to exit at that same P/E, it would not matter if that PE was 15.0 or 48.7, the gain on sale would yield 4.6% on cost. It is no coincidence that 4.6% is the same as the average yearly EPS growth rate for that period.
The only thing that will change that percentage gain is if the P/E at exit is higher or lower than the P/E at entry. If the P/E at exit is higher, the percentage gain will be greater, and vice versa. But, as noted above for JNJ, the P/E level (at entry only), will make a difference to the dividend yield on cost. In TABLE 4.1 above, dividends contribute 3 percentage points to the overall yield of 7.6% for the 15.0 P/E case but only 1 percentage point for the 48.7 PE case, because the cost per share is roughly 3 times that of the 15.0 P/E case.
That just reflects the obvious: the amount paid per share is a dollar amount per share, and the dividend received per share is a dollar amount per share so the less paid per share the higher the dividend yield. Note that the dividend yields shown in all tables are based on the average of the High and Low share price for each year.
While JNJ had the earnings growth rates to justify to some extent the high PE ratios in the 1997 to 2005 period, it was not the case for KO. While JNJ grew earnings by an average of 16.8% per year for this period, KO only managed 4.6%. In fact, Chuck Carnevale was moved to write this article, Would Ben Graham Give Warren Buffett an 'F' for Holding Coca-Cola?, questioning why Warren Buffett held KO shares during the period 1992 to 2010, when the company appeared very much overvalued based on high share prices compared to earnings.
TABLE 4.2 - Coca-Cola P/E ratio history 2005 to 2016
This article seeks to stress the importance of future EPS growth rates over P/E levels when making a decision on investment in shares. TABLE 4.2 indicates with P/E levels for KO going ever higher, and patchy and negative EPS growth rates, the question of where to from here for KO is of high importance.
Review of Procter & Gamble P/E ratio history 1997 to 2016
TABLE 5.1 - Procter & Gamble P/E ratio history 1997 to 2005
We saw above with JNJ, the high 16.2% per year EPS growth rates in the 1997 to 2005 period allowed plenty of opportunity to exit an investment without loss even when buying at a year High P/E at entry and selling at a year Low P/E at exit. For PG, with a lower average growth rate of 10.5% for this period, the opportunities to exit were still there, but at lower percentage gains than for JNJ.
The opportunities for PG investors to make above average gains by buying low and selling high between 1997 and 2005 were also there, but once again, the lower 1997 to 2005 average EPS for PG versus JNJ meant, with one exception, the returns were lower for PG than for JNJ. The exception is for the PG 2001 to 2005 period. As shown at the foot of TABLE 5.1, an investment in 2001 with an exit in 2005 could have given a return of 22.4% per year.
This was brought about by EPS average growth of 25% per year for the period 2001 to 2005, offsetting negative growth of 2.3% in the 1997 to 2001 period. TABLE 5.1 shows PG share price fell from a High of $59.73 in 2000 to a Low of $40.73 in 2001. This proved to be a great buying opportunity, as PG quickly recovered from that period of declining EPS.
TABLE 5.2 - Procter & Gamble P/E ratio history 2005 to 2016
Two things stand out from the above TABLE 5.2. The first is that decline in EPS growth over time is an issue for PG investors, current and prospective. As we have seen from the discussions above, EPS growth is vital to the level of return achievable.
The second is that continued dividend growth might not be possible unless it is more closely matched by EPS growth. The type of investigation and analysis of fundamentals at the company level, that Chuck Carnevale always encourages, should be undertaken for this company and all other companies included in this article
Review of McCormick P/E ratio history 1997 to 2016
TABLE 6.1 - McCormick P/E ratio history 1997 to 2005
It can be seen from TABLE 6.1 above that MKC did not enjoy (or suffer?) the high P/E ratios of JNJ, KO and PG in the 1997 to 2005 period. But if an investor were to have waited for the P/E ratio to come all the way down to 15.0, the only year an investment could be made was the year 2000. In that year, the P/E came down to 11.9, but if the focus is on 15.0, this P/E might have been considered too low and investment avoided.
A policy of investing only within a P/E range of 14 to 16 would have been very rewarding in the case of MKC. But it was also possible to invest in JNJ at a far higher P/E of 24.1 and still achieve a 16.1% average yearly return (see TABLE 3.1). This 16.1% compares favorably to the returns shown above, but a rule restricting investments to P/Es of 15.0, or even to a range of 14 to 16, would have ruled out JNJ as an investment at that time.
TABLE 6.2 - McCormick P/E ratio history 2005 to 2016
TABLE 6.2 above reveals that a P/E rule of 15.0 (or range of 14 to 16) could keep an investor from investing in MKC for long periods. However, the table shows there have been numerous opportunities to enter and exit MKC at say an 18 to 20 P/E range or even a 20 to 24 P/E range. We know from the various discussions above, so long as the exit is at a similar or higher P/E than entry, the gain will be similar to or greater than the EPS growth for the period.
For the 2005 to 2012 period, there were plenty of opportunities to enter and exit at a 19 P/E. There was good EPS growth throughout the period so there would always be a gain on exit plus dividend income.
Summary and Conclusions:
High P/Es and Dividend Growth Companies
I trust I have convinced readers that use of a P/E of 15, or historical P/E averages, is generally inappropriate as a guide to investing in shares of individual companies. I hope I have shown one of the more important elements in deciding on an investment is the expected future growth rate of EPS. If the number of companies I have included to illustrate these points seems like overkill, I have to say I have just scratched the surface of this subject.
P/E is just a function of dividing current share price by current EPS. To understand shifts in P/E, we must understand what causes shifts in, what drives both EPS and share prices. I need to use the data gathered for the various companies as part of analyzing what drives share prices and EPS. Obviously, EPS is one element that drives share prices, but many other factors also influence share prices. And many external factors can affect EPS, independent of how a company and its management perform. Consideration of the potential impact of black swan events will form part of this work.
By definition, it is not possible to predict a black swan event. But it is possible to consider what events, brought on by a black swan event, might result in a drastic direct or consequential impact on the share market and on individual companies. Good economic news in the form of higher productivity, employment and wages growth could result in an increase in interest rates and inflation.
I do believe an event that causes interest rates to increase significantly will have an adverse impact on share prices, and thus P/E ratios. For those companies that are highly geared there will be a double whammy in the form of higher interest rates depressing EPS, with a flow on to share prices, together with higher interest rates depressing the price of shares generally.
Low interest rates have continued for a long time now, and I do not see how this will change any time soon. And that is why I believe any change will likely arise from a black swan event. It will happen, but one could be a very long time out of the market waiting for it to happen. Or it could happen in the very short term. We need to prepare by understanding the relationship between interest rates and P/E levels for companies we are considering investing into.
I have seen comments that past high P/E ratios have occurred in times of high interest rates and thus there is not a strong connection. In further parts to this article, I will provide additional analysis and comment on the matter of current high, but not historically high, P/E ratios. I have only scratched the surface of this subject in this article.
For now, I will leave readers with the following to consider -
An extract from my article, "Brexit: Current Day British Citizens' Boston Tea Party"
The ERM was of course another example of an EU invention that could not work. George Soros, and other speculators shorting the GBP, were just taking advantage of what they could see was an unworkable system.
The current monetary and interest rate situation is another invention that is clearly not working, despite the best efforts of governments. If George Soros could crash the British pound, could a group of George Soros type speculators bring this current unworkable invention crashing down, taking the matter out of the hands of governments?
Disclaimer: The opinions in this document are for informational and educational purposes only and should not be construed as a recommendation to buy or sell the stocks mentioned or to solicit transactions or clients. Past performance of the companies discussed may not continue and the companies may not achieve the earnings growth as predicted. The information in this document is believed to be accurate, but under no circumstances should a person act upon the information contained within. I do not recommend that anyone act upon any investment information without first consulting an investment advisor as to the suitability of such investments for their specific situation.
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.