In this post, I plan on evaluating whether Du Pont (DD) is a good opportunity to create alpha [α], through beta [β]. When we look at the make-up or components of α, it is clear that the market return, and risk free rates are uncontrolled by investors. It is equally clear that investors cannot influence β. The question is therefore, what does influence β? There is more detail on β below in this post, but in my view the key factors that influence β are business mix, financial leverage, and leadership quality. If a change in any of these factors can be reasonably expected, the β will change. When it does change, the perception of risk associated with the security versus the market also changes. And that provides a massive opportunity to generate abnormal returns or α.
Ultimately creation of α for a stock is all about stock selection. And it is about returns that the stock selected will generate. Du Pont operates in several segments including Agriculture, Electronics & Communications, Industrial Biosciences, Nutrition & Health, Performance Chemicals, Performance Coatings, Performance Materials, Safety & Protection and Pharmaceuticals. The Performance Materials, Electronics & Communications, and Safety & Protection business segments have been performing very well. The Performance Chemicals business is expected to be spun off by 2015. This segment adds much volatility to earnings at Du Pont. An exit would lower earnings volatility, and the reduced volatility and change in business mix is expected to lead to a contraction in the β for Du Pont. In addition, once this difficult segment is divested, the Agriculture, Nutrition & Health, and Industrial Biosciences segments would benefit from more management focus and better capital allocation. The Agriculture segment is the one I like best. There is a powerful global secular tailwind for this sector - simply put, incomes are rising, there is hunger in the world, and food must be produced to feed the world. This anticipated change in business mix and expected lowering of earnings volatility make Du Pont a very interesting β convergence opportunity. I would strongly recommend that you read Value Line's more comprehensive overview here.
At first glance the return potential is not terribly exciting. Value Line projects a target of $90 to $75 for Du Pont by 2016-18. That indicates an annualized price appreciation of 6.8% to 11.7% over the 3 to 5 years through 2016 to 2018: you can download their pdf report here. Add to that a dividend yield of about 3% and the total annualized return potential is between 9.8% and 14.7%, or an average of 12.25%. This 12.25% return potential is very much in-line with the risk adjusted return investors should demand using the present Reuters β of 1.65. However, the total return could be far higher should the β contract as earnings volatility reduces with a changed business mix. As β declines, so do the risk adjusted returns expected by investors. If it occurs, much of the decline of the risk adjusted return expectation will come from superior than market price appreciation generating an abnormal return, or α.
What is α?
The underlying principle is the recognition that investor returns and risk are directly related. Returns reward risk takers. An investor's return and return expectation is based on risk associated with an investment. α can be positive or negative, though when we seek α, be assured that we are seeking a positive α! Α is created when the return from an investment is higher than the risk adjusted return which should have been earned from the investment. Seeking α is the art of stock selection, where efforts are made to identify stocks which will generate higher actual returns, compared with the risk adjusted returns which should have been earned after considering the risk of the investment relative to the market.
The risk adjusted returns are often computed using β, risk free rates, and market return expectations. Since β is a measure of risk of an investment relative to the market, the Capital Asset Pricing Model computes the risk adjusted return for a specific asset as the risk free rate plus β multiplied by the difference between market return expectation and the risk free rate.
The risk adjusted returns which should have been earned after considering the risk of the investment relative to the market can be expressed mathematically as:
[Rf + β * (Rm-Rf)], where Rf is the risk free rate, Rm is the market return, and β is a measure of risk of an investment relative to the market.
Mathematically α can be expressed as:
σ = Ri - [Rf + β * (Rm-Rf)], where Ri is the return generated by the investment, Rf is the risk free rate, Rm is the market return, and β is a measure of risk of an investment relative to the market.
If you held every stock listed, Ri would be equal to Rm and β would be 1. Thus there would never be positive α. α for a stock can only be created through stock selection skills and α for a portfolio can only be created through stock selection and capital allocation skills. Creating α repeatedly and consistently over the very long-term is believed to be near impossible by believers in the efficient market hypothesis. But for those of us who believe markets are both efficient and inefficient, α is important. In my view, markets are in a constant state of flux. Markets are chaotic. And markets are in a constant state of disequilibrium in the journey towards equilibrium. Inefficient markets tend to efficiency. And what we must do is follow the path to efficiency, and the journey to equilibrium. Thus as α rises, so does my conviction that before long movements towards equilibrium will snatch away the α created. And so an escalating α can be used to re-balance portfolios to remove money to an entirely different asset class.
What is the risk free rate?
The risk free rate is the theoretical return on an investment with zero risk over a specified period of time. I tend to use the ten-year US government securities as a measure of the risk free rate. At present the rate is 2.77%. However, over the coming several years, I expect it to rise towards its ten-year median levels of 3.69% and so I will use 3.5% as an estimated risk free rate for the coming five year investment horizon.
What is the market rate of return?
Over the very long term, the markets have tended to deliver returns (including dividends) of 9%, 6.5% in real terms and 2.5% inflation. With the markets trading at 1,790 recently, I expect long-term real returns to be between 5.5% and 6.5% with an upside of 1.2%. This 1.2% upside to long-term real returns comes about as a result of the S&P 500 being leveraged to higher global real growth rates instead of U.S. growth rates. Overall, an expected value of long-term return of 6.6% is not unreasonable. I expect inflation to run at between 1.7% and 2.5% and thus an expected value of total nominal return of 8.7% is not unreasonable.
What is β?
β is an imprecise measure of risk of a stock relative to the market. It represents systematic risk that cannot be diversified away. The β co-efficient is calculated from a regression analysis of the relationship between daily/weekly/monthly percentage changes in the price of a stock and daily/weekly/monthly percentage changes in the Index over a period of over two years. The obvious problems associated with β are:
- It works with historic data, which makes it backward looking. Thus it is no more than indicative of future β.
- As a result of price volatility, β has a high standard error, and as a result it is less reliable as a predictive tool of where β might be in the future.
- β can and does change over time: it displays a tendency to converge towards 1.
When we work with regression analysis to calculate β, we know that the resulting β has a high standard error and reflects the financial leverage and business mix over a period of time instead of the point in time that you invest. This problem can be addressed by calculating a bottom-up β by calculating industry-wide regression β's for the various businesses in which the company operates, and then adjusting the industry-wide β for the business mix and financial leverage of the company being evaluated. If we calculate a bottom-up β to address these concerns, we are exposed to β errors caused by changes in financial leverage and the business mix at company and industry level over the investment horizon. In addition, we add a new risk of incorrect analyst judgment, because the exercise of judgment is required when calculating a bottom-up β. Finally, a bottom up β cannot quantify the impact of qualitative facts on β, which get picked up by the β calculated using regression analysis - for instance high confidence in management, leadership and ownership quality will have an impact and lower β. On balance, in the case of Du Pont, in my view regression analysis is the better tool, with the five-year regression being most appropriate for long horizon investors.
Investors use different methods to compute β. Thus β quality is important. For instance, a time horizon close to the duration of a typical economic cycle is better than a shorter time horizon. History suggests that on average an economic cycle lasts sixty-six month, thus of the various durations used for regressions, I feel the five year duration is most appropriate. Another important consideration is which of the daily, weekly or monthly price change is best to use in the regression analysis. In my view, daily data adds too much noise, and a month is probably too long, thus a weekly price change is probably the most appropriate. Then we have the issue of β with a high standard error - this reduces the usefulness of β as a reliable indicator of risk. The standard β is of lower quality than a β with adjustment made in consideration of the β's tendency to converge towards 1.
Bloomberg reports the raw β and a standard error for β. Thus at a 67% confidence level the β will fall in a range of raw β plus or minus one standard error. They also provide an adjusted β, which is 67% of the raw β plus 33% of 1: this is intended as an estimate of the future β as it gives 33% weightage to the securities' tendency to move towards 1 and 67% weightage to the historic raw β. Value Line says that they calculate the β co-efficient from a regression analysis of the relationship between weekly percentage changes in the price of a stock, and weekly percentage changes in the NYSE Composite Index over a period of five years. In the case of shorter price histories, a shorter time period is used, but two years is the minimum. Value Line then adjusts these β's to account for their long-term tendency to converge toward 1.00. Reuters says their β is calculated based on trailing five-year prices, on a monthly basis, relative to the S&P 500. Which is best?
Reuters reports β for Du Pont at 1.65, Value Line at 1.20, the two and five year regressions are 1.13 and 1.60 respectively. Of these, I feel 1.20 from Value Line reflects the most appropriate forward β, but it too does not reflect the alteration to β as a result of the anticipated change in the business mix - the exit of the volatile and cyclical Performance Chemicals will impact β, as will the increase in significance of the agricultural segment as a proportion of total business. Adjusting the β of 1.20 to reflect the change in business mix gives me an expected forward β of 1.08. However, my 1.08 projection is a target for five-years after a successful spin-off of the Performance Chemicals. A more likely β projection for 2018 is 1.3. I derive this number from a projection of a raw five-year regression β of 1.44 in 2018, adjusted for its tendency to converge towards 1 over the next five years.
Value Line Estimates
As far as estimates are concerned, I find the track record of Value Line is nothing short of exceptional. A Bloomberg Terminal is likely the most popular data service. If you are comfortable with looking at a variety of other analyst expectations to arrive at your own conclusions, then Bloomberg or Reuters Eikon is what you are looking for. But if you are looking for a single point of high quality data, vetted by a single organization, then Value Line is the place to go. They project growth rates of 8% through 2010-12 to 2016-18. With 2013 coming in at $3.88, an earnings growth of 8% can be expected to lead to earnings of $5.70 in five years. With cyclical volatility, this target can arrive at any time during the next 3 to 5 years.
Value Line also projects share-count declining from about 925 million in 2013, to 880 million by 2016-18. This indicates that buybacks net of dilutive employee compensation will be 1% annualized. Thus while growth for shareholders earnings is at 8%, growth for the business earnings is at 7%. Given real global growth potential of 4.2% and global inflation at 3.8%, I feel growth rates of 7% to 8% for Du Pont are conservative estimates of both current and terminal growth rates.
With an industry average return on incremental equity of 20%, to sustain growth at 7%, 35% of earnings will need to be re-invested, with the remaining 65% being available to shareholders. However, efficient allocation of capital across all sectors and industries in the market suggests that it is likely that in the long-term return on equity will revert towards the 13.5% level we have seen over the long-term for the S&P 500. Thus Du Pont will need to re-invest 51.85% of earnings to generate growth of 7%. The remaining 48.15% is the notional payout. Some of the notional payout will be returned via dividends, some via buybacks, and perhaps some through re-investment which will deliver returns via higher than 7% growth. What needs to be valued is the notional payout.
What is this notional payout worth?
Fast forward to 2018 for the vision of the soothsayer!
Value Line estimates a range of $75 to $90 for Du Pont between now and 2018. In my view, the price is more likely to be $90 to $107 during this time frame. With these targets, the five-year return potential falls between 8% and 11% annualized. Add to this the dividend yield of 3%, and Du Pont is an attractive investment prospect. However, if your quest is to generate α, then perhaps Du Pont is not for you - using my 1.3 β estimate, we see a negative α of 2.15%. A target entry price of $55 would generate tiny positive α, while a target entry price of $50.50 will generate α of over 2%.
How far-fetched is a potential price target of $107?
I have presented an alternative valuation for Du Pont below. This model uses historic PE, and cyclically adjusted PE's to project value ranges for 2014 and 2018. I have used earnings and average annual PE data from Value Line, except for 2013, where I have used the $3.88 recently reported earnings instead of Value Line's estimate from prior to the release. For 2014-2018, I have used my own estimates which are largely consistent with Value Line estimates. The annual average price noted below is derived from the earnings and annual average PE ratios from Value Line. The earnings growth rate is a computed one year earnings growth rate.
The 6 year median EPS is exactly what it is described as: it reflects how earnings have moved over a six-year period, which I see as a reasonable estimate of the duration of a typical economic cycle. The 6 year median EPS is a useful tool when looking at companies like Du Pont where earnings are heavily influenced by cyclicality. EPS 6 Growth is a calculation displaying how 6 year median earnings have grown over time. The CAPE is a cyclically adjusted PE which is computed by dividing the annual average price by 6 year median earnings.
As you can see a price target range of $90 to $107 reflects a distinctly bullish view, using this methodology as a guide to valuation. History never repeats, but it often rhymes. If history does rhyme, the target price range of $90 to $107 may be too bullish. A price target range of $80 to $90 might be more appropriate.
I have looked at and liked Du Pont, primarily because of the anticipated spin off of the Performance Chemicals business and the secular tailwinds provided by the agriculture industry. I do not however own Du Pont. I also have no intention of initiating a long position in Du Pont, and as a long only investor, I do not short stocks.
I have sufficiently high conviction that a $107 as a five-year price target is reasonable. Thus I would like to own Du Pont. However, the capital I manage is not large enough to allocate at industry level. My allocation is focused at the sector level, and to get coverage to the basic materials sector, I prefer more diversified companies like BHP Billiton (BBL) and Rio Tinto (RIO). Where possible I do try to get exposure to industries with secular tailwinds, through stock selection: for example my allocation to industrials includes Deere & Co (DE).
Let me conclude by saying that this post is not a recommendation of any sort. Nor is it research. Nor can it be considered due diligence. It is merely an idea, or an investment thought-piece: if you like, it is a penny for my thoughts! I'd be delighted if you enjoyed it and it got you thinking, but if you buy, or if you sell, be sure to do your homework, research and due diligence first. If you did not like the post, or the thought of Du Pont representing what might be an interesting buy pick that is fine too. There are two sides to every trade. We need a seller and buyer for the two parties to walk away satisfied: if you are a seller of Du Pont, be glad that there are buyers available.