Basics of the research
This article presents a research on the Procter & Gamble Co. (NYSE:PG). It is the third one of a series which examines some risk and reward characteristics of several well known dividend paying stocks - Johnson & Johnson (NYSE:JNJ), Coca Cola Co. (NYSE:KO), Procter & Gamble Co. , Chevron Corp. (NYSE:CVX) and PepsiCo Inc. (NYSE:PEP). Subsequent articles are built on the previous ones so readers could expect to find more information and stocks comparisons as the number of articles grows.
The main purpose of the analysis is to provide information about the historical risk, and especially the tail risk, which is present in a stock's distribution of total returns. Tail risk is the risk of getting returns which stand as far as several standard deviations from the mean. Tail risk could have a significant effect on a stock's return, mostly because of its potential magnitude. By definition, such tail returns are a rare event. The potential problem appears if the tail risk materializes in a period in which investors expect a withdrawal from the portfolio to happen. This could disturb financial strategies, and depending on the leverage used, it could have serious consequences on financial stability of the entities involved. The tail risk is represented by two characteristics of the distribution of returns - skewness and kurtosis.
Regarding the long-run risk of loss, a positive skewness is sometimes the preferred one. The reasoning behind this view is that the positive skewness incorporates a higher historical probability that a large unexpected return would be on the positive side, i.e. larger than the mean. On the other hand, a negative skewness presents a higher probability of an unexpectedly large negative return during some period. Hence, it shows an increased risk concerning the available capital at the end. For a better understanding, the reader could use the classic explanation where the negative skewness is compared to picking up nickels in front of a bulldozer. The large unexpected negative return comes with the bulldozer's move.
The kurtosis as a measure of risk indicates whether the return distribution could experience fat tails. Those are major risk descriptors which show whether there is a larger than the normal probability of having returns far from the mean, both positive and negative. For instance, a kurtosis value of zero shows there is no such higher-than-normal probability.
Readers could check the Johnson & Johnson article in which the skewness and kurtosis were explained in more detail.
The total return of a stock is a sum of its price change and the cash flow which investors get on a regular basis, basically in the form of dividends. Because dividends and share repurchases are a major part of the examined stocks' total return, special attention will be paid to risk on cash distributions.
The current research of Procter & Gamble Co. covers returns data from the last 43 years. This long period includes several different economic environments and cycles, both good and bad ones. Hence, the calculated long-run coefficients could be expected to be relatively free from sampling biases and closer to the true descriptors of the stock's population of returns.
Nevertheless, the readers should keep in mind that historical performance does not guarantee future results.
Readers should also account for the mean reversion tendency the returns generally exhibit in time. As mentioned in the first article, the performance in the more recent periods, i.e. the last 12 months or last 5 years, could be more important and useful for the moment, but it would tend to get closer to the mean characteristics of the longer time frame periods, given that no significant change in the economic environment or the company itself has happened.
The Companies Selected
All of the companies examined in the series have a long history of distributing dividends. Moreover, they have increased the dividend size in almost each of the last 40 years, with the one exception of Chevron; the company suspended dividend distribution in the period 1976 - 1984. The average yearly dividend increase for Procter & Gamble amounts to 10%.
During the last 5 years, the companies also distributed wealth to their shareholders through share repurchases. Generally, this increases the demand for stocks and has a positive effect on their price. If at the end of the fiscal year, the company has a different amount of basic shares outstanding than at the beginning, this change could also affect the size of the dividend paid and the payout ratio.
Procter & Gamble Stock's Characteristics and Analysis
Procter & Gamble Co. is an U.S. based company which provides consumer packaged goods under several well-known brands. It operates worldwide and was founded in 1837.
The company's financial year ends in June, so technically we are now in the second half of its current financial year. This is different than the other companies examined so far, but because the analysis covers a significant amount of years, we could assume the differences have a minor effect on the conclusions.
Procter & Gamble Co.
The Procter & Gamble stock experienced a compound annual growth rate (OTCPK:CAGR) of 12.2% for the whole examined period of about 43 years. For comparison, the CAGR of S&P500 for the same period is 6.9%.
We should note that the first month of 2013 which generated a return of about 11.6%, is somewhat of an outlier in the stock's price history. A PG's monthly return of between 10% and 12% could be historically expected to appear once in every 40 months. The empirical probability that such a month will be followed by a month with a similar performance is close to zero (0.06%). The historical probability of more than 99% that the January increase will not be repeated in February, could justify the use of strategies for protecting the profits generated so far, like writing calls or buying put options on the stock.
Comparison with Johnson & Johnson and Coca Cola
While P&G achieved 12.2% CAGR for the last 43 years, the CAGRs of the Johnson & Johnson and Coca Cola for the same period are 12.54% and 12.34%, respectively.
When we zoom in to the last 5 years of data, Coca Cola takes the lead with its 53% increase over the period, Johnson and Johnson is second (42%) and Procter & Gamble ranks third (31%).
During the last 12 months, JNJ started to move faster and increased with 18%, PG comes second (14%, of which 11.6% in January 2013) and KO is third with a 13% rise in the stock price. As mentioned earlier, the almost 12% return of PG looks like an outlier and the probability that the stock will continue to present similar results in the next months is low.
The following table shows the effect the last two months had on the average monthly returns of the three examined stocks for the respective 12 last months.
We see the lowest effect on KO and the biggest one on the return of PG.
Comparing the graphs of adjusted monthly prices, we see that apart from the differences in KO's price behavior which were commented in more details in the Coca Cola analysis, all the companies seem to be in a long term uptrend. PG has experienced more prolonged downward movements during the last two major crises, in 2000-2001 and 2008-2009.
Similar to both JNJ and KO, Procter & Gamble has split its stock 6 times in the examined period - in May 1970, February 1983, November 1989, June 1992, September 1997, and June 2004. With the exception of the first split, all the other prices at which the stock experienced a split were above $100. This makes a difference between the current price and a potential split price of at least $25. This 33% potential increase might seem like a nice deal at first, but the time horizon of such increase is uncertain. The current PG price also seems to be a bit high to present an attractive risk/reward ratio, especially when compared with the one of KO which is above 6.
Regarding the risk adjusted price return for the whole examined period, the PG stock shows the best result from the three companies, with a Sharpe ratio of 0.20 against 0.19 for both KO and JNJ. PG's risk-adjusted return is also better than the S&P500's one (a Sharpe ratio of 0.15) for the same period.
Analysis of Risk on Dividends and Cash Distributions
Because the total return consists of a price and a dividend return we should also examine the risks on the dividend portion of the total return.
Procter & Gamble is constantly increasing the size of its dividend, similar to the other two examined companies. The average annual growth rate for the last 40 years is 10%. It is interesting that during the current worldwide financial crisis, the company has increased its dividend at a higher-than-average rate during the 2007-2009 period. For the next three years till now, the rate of increase is below the average one.
The projected dividend for 2013 (see the table above) seems realistic, without putting too much pressure on the company's financial results. It implies a dividend growth rate of about 9.8%. Having in mind the number of basic shares outstanding the company has as of 31 December 2012, this would translate into a net income for the P&G's fiscal 2013 of about $12.3B using the last 5 years average payout ratio of 55.1%. Such a net income is above the management's recent implied projections of $11.2B but having in mind the results from the first two quarters of the fiscal 2013 (net earnings of $6.75B), it does not seem very improbable.
An interesting thing to note is that during the last five years, with the exception of 2012, the company has spent more money on repurchasing its own shares than on distributing dividends, as the table below shows.
This policy of distributing cash through share repurchases has generally contributed to the price stability, has increased the total return of the stock and might have presented some tax advantages for the shareholders. The possible tax effects are explained in this Forbes article. Its focus is Berkshire Hathaway, but the main principle is the same.
The continuation of such a policy and the tax advantages it provides to shareholders, could be hard to follow however, if the company faces a free cash flow or operational constraints. Even without such operational constraints, the share repurchases might take a lesser part in the future cash distribution, because the management generally is more reluctant to cut the dividend than to cut repurchases. As the dividend is projected to grow, the share repurchases must decline, both as a percentage of dividends and as percentage of the whole cash distributed, because otherwise the distributed sum might grow to unsustainable levels.
If we compare the three examined companies, we see from Table 3 that for the last 5 years, Coca Cola and Johnson & Johnson have distributed on average about 80% of their net income to shareholders, while Procter & Gamble's average distribution ratio goes to 113%. This generally means the company gives back to shareholders more money than it earns. In order to proceed with such a strategy Procter & Gamble has to come up with the money from somewhere else than its main operations, either by issuing debt or by divesting of non-strategic assets. The asset base of PG has declined by 8.3% (from $144B to $132B) for the last 5 years, so it seems the company has chosen the second way to arrive at its ratio of cash distribution to shareholders. Regarding sustainability, this is the preferred way over taking more debt. A potential risk here is that should the company's management decide that there are better options to invest the company's money, the price supporting effect of share repurchases, and hence the stock price, could suffer.
Regarding the effect of share repurchases on the stock price in a shorter term, such effect could be diminishing till the end of the company's fiscal year. Recently Procter & Gamble announced that its shares repurchases outlook for 2013 is in the $5 to $6 billion range. For the first two quarters of its fiscal 2013, the company has already repurchased stock for about $4B. This leaves a maximum of $2B to be used in the next two remaining quarters. This is about 50% of the amount used for the first half of the fiscal year. Investor should also keep in mind that the higher the stock price goes, the less supportive power such purchases would have.
Another risk characteristics
(click to enlarge)
Examining the whole period of 43 years we see a returns distribution which has a negative skewness and a significant kurtosis. The negative skewness (-0.34) speaks of a higher-than-normal historical risk of an unexpected negative monthly return which would stand far from the mean value. The kurtosis (3.36) is higher than both the one of Coca Cola (1.75) and of Johnson & Johnson (0.39). It shows a distribution with fatter tails, both compared to the normal one and to the distributions of the other two companies. The fatter tails present an increased probability of having extremely large returns on both sides of the mean, generally. The combination with a negative skewness suggests more of those returns could be lower than the mean. The most notable reason for these descriptors of the distribution is the negative return of -35% in March 2000.
The closer time frames change the situation significantly, especially regarding the kurtosis and the existence of fatter tails. With the kurtosis close to zero, we could expect no higher than the normal probability of returns far from the mean. The skewness gets increasingly positive during the last 12 months due most probably to the monthly increase during January 2013.
The decreasing correlation and beta values of the PG stock for the last 12 months also deserve attention. They are lower than both the KO (correlation = 0.29, beta = 0.36) and JNJ (correlation = 0.66, beta = 0.62) values. This increases the stock's appeal as a defensive investment. Due to the low correlation with S&P500, PG might be used to improve the risk characteristics of a portfolio which follows the general market.
Because of the mean reversion tendency which the characteristics of the returns generally follow, we could expect that the more recent descriptors would converge to the longer term one. If the long-run coefficients are indeed closer to the true population descriptors, we should expect the recent risk characteristics of PG to gradually revert to the longer-run values. This would mean the future might again presents a higher-than-normal risk of having returns far away from the mean value because of fatter tails.