THE QUAKER OATS COMPANY - COST OF CAPITAL (A)

While the question of appropriate measures of the cost of capital had been under discussion within the Quaker Oats Company for several years, by mid-1973 considerable disagreement still existed among the people involved. Although it was widely accepted within the company that Quaker Oats' overall cost of capital was approximately 8%, some people believed the true cost to be much higher, particularly due to the company's recent aggressive diversification efforts. Quaker Oats' acquisitions in the toys and recreation field, the growth of its restaurant business and its chemical division, these people argued, have undoubtedly changed the nature of the company's business and the sources of its profits. Since these new areas have traditionally not been as stable as the grocery products business, they must have increased the risk associated with ownership of Quaker Oats shares. Therefore, the company's cost of capital is likely to have increased.

The Quaker Oats Company, incorporated in 1901, is a major worldwide producer and marketer of packaged brand-name food products, including cereals, mixes, table syrup, frozen foods, corn products and pet foods. The company is also a major producer of furan chemicals, which are used in the foundry industry, the refining of petroleum and in the manufacture of a wide range of products such as plywood, rubber, and plastics. Operations are conducted through the following groups:

Exhibits 1 and 2 present audited financial statements for fiscal years ended on June 30, 1972, 1973, and 1974. A breakdown of the company's sales and earnings before interest and taxes by operating group is given in Exhibit 3.

The extent and pace of Quaker Oats' recent diversification efforts is probably best evidenced by the fact that although sales of the Grocery Products Group, which has represented the company's traditional line of business, grew by 39% during the 1971-1974 period (from $449.614 in 1971 to $628.9M in 1974), its share of total company sales dropped from 64% in 1971 to 51% in 1974 (see Exhibit 3).

The debate within Quaker Oats over the company’s cost of capital is summarized in the following series of intra-company correspondence:

June 19, 1973

I feel that the debate within the company on our cost of capital is extremely important. Although it is impossible for anyone to say with certainty exactly what our cost of capital is, a relevant range is very important for decision-making. The thing, which concerns me most, is that top management feels that our cost of capital is only 8% when it is actually much higher.

A large part of my concern is centered on our acquisition evaluations. Even though the debate is over only a few percentage points or less, a few percentage points can make a tremendous difference. As an example, when we were discussing if we should buy our own stock, an evaluation was done using our acquisition model. With the cost of capital assumed to be 8%, our stock was evaluated at $37.71 per share. Using 9%, our stock was evaluated at $31.77 per share. This represents a difference in evaluation of Quaker Oats of approximately $124MM, a sizable error in evaluation. This makes clear the importance of spending more time evaluating the cost of capital for projects and acquisitions. Even though we may continua to use 10% for project evaluations, management may have a false sense of security that marginal projects are profitable because they believe the cost of capital is only 8%.

I assume that those who will be reading this have some basic understanding of the latest theories of finance which in a large part are based on work done by Modigliani and Miller, and as I know has been presented to several people in the company by Joel Stern. Even if one has no prior knowledge of the subject, I hope that the following discussion of our cost of capital is clear and easy to understand.

The true cost of capital represents the opportunity cost of capital, that is, the return that could have been obtained had the funds been invested elsewhere with the same amount of risk. There are two types of risk, business risk and financial risk. The return, which could normally be expected, from an investment in a particular industry is the return necessary to compensate the investors for the business risk. It is independent of the financing of the investment. On the other hand, financial risk is the risk arising from the method of financing. When the firm employs both debt and equity, the total risk is broken into two parts. The bondholders are accepting less than the full risk while the stockholders are accepting a larger share of the risk. The leverage associated with issuance of bonds will result in a greater fluctuation of the stockholders returns, thus greater risk. It is apparent then that there is no financial risk associated with the ownership of common stock if there is no debt outstanding. All of the risk in this case would be what we are calling business risk.

This method assumes that the cost of equity capital can be estimated from the present price/earnings ratio. Using this method, the earnings/ price ratio is taken as the measure of the cost of equity capital. This implies that a high PIE stock has a low cost of capital and implicitly should be willing to undertake projects, which have low yields. In fact, the PIE ratio is high because the investing public believes that the company can invest at higher rates of return than the average company. This method also neglects future earnings and does not look at cash flows and the timing of the cash flows.

II. The Present Price/Expected - Earnings Method

A second approach is to use the expected potential earnings or estimated average future earnings divided by the market price of the stock. This approach has two drawbacks. First, it looks at the earnings rather than the cash flows and secondly, the evaluation of a stock is dependent on many things, only one of which is earnings.

III. Dividend Valuation Method

A third approach is the evaluation of dividends. This approach would be acceptable except that earnings and dividends are only loosely related. The problem really becomes one of estimating the cash flows and the potential dividends in the future.

This method computes the return to Quaker Oats stockholders in the past and assumes that this is the return they deem necessary for the future. This is less than satisfactory because the return in the past is dependent on numerous factors, which do not necessarily project into the future.

This leaves the risk-adjusted method as the only other alternative. It is not perfect, but it is much better than any of the other alternatives. One of the primary reasons it is better is that it uses the market return to common stocks as a base. This eliminates any of the problems associated with individual stocks. The price of a company's stock is usually relatively unstable while the adjustment to the market rate of return necessary to account for a company's risk should be relatively stable over time.

The risk of owning an individual stock can be partially eliminated by owning a portfolio of stocks. This means that even though several individual stocks may be very risky, ("risky" meaning a high fluctuation or variance in price), by holding a large portfolio this risk can be diversified away. Risk may be explained by looking at a frequency distribution.

If the above frequency distributions are "standard normal," the height of the curve represents the probability of being at a particular return. In both cases the expected return is 10%, but in "A" it is much more likely that the return far exceeds 10% or that it is far less than 10% or negative. While "B" has the same expected return, it is improbable that the return will be far from 10%.

The true risk of a stock is how its return is related to the total portfolio. If all the stocks moved up and down together then this risk would be eliminated by diversification. This relationship of the movement of stocks is called the covariance. If two stocks move in the same direction, they have a positive covariance. If they move in opposite directions then they have a negative covariance, and if there is no relationship the covariance is zero.

The covariance is the risk, which cannot be eliminated by diversification. Thus, if a portfolio of stocks has a high covariance with the total market, diversification will not eliminate the risk and, therefore, investors must be compensated with a higher rate of return. This is another way of saying that the cost of capital for a firm with a high covariance with the total market is higher than a firm with a low covariance.

The statistic, which is used to describe the relationship of an individual stock to the total market, is called the "beta" (B). If an individual stock's "beta" is equal to 1, when the market moves up or down 1%, this stock also moves up or down by 1%. A "beta" of 2 would indicate a 2% movement in an individual stock when the market has a 1% change. The S&P 500 Stock Index is usually used for movement in the market.

Since the beta measures the movement in the price of individual stock in comparison to the total market, it automatically takes into consideration the amount of leverage of the individual company.

The cost of equity capital for an individual stock can be determined by using the beta of that stock, the risk free rate of return and the market rate pf return. The Treasury bill rate is usually used for risk free rate, and for the market rate of return 9,3% is often used. This is the return found by Fisher and Lone for all stocks listed on the New York Stock Exchange for the period 1928-1965. Merrill Lynch, et al, have compiled the "B" statistics for most publicly traded stocks. The cost of equity capital, Y, can be computed by using the following formula:

Y = Rf + (Rm - Rf) B

Where Rf is the risk free rate of return, Rm is the market rate of return, and B is the beta of the individual stock.

The overall cost of capital (c*) can then be computed using the weighted average cost of debt and equity capital. That is:

C* = Y*E/(D+E) + (l - t)b*D/(D+E)

Where E is the amount of equity, D is the amount of debt, Y is the cost of equity, b is the cost of debt, and t is the tax rate.

Before actually calculating the cost of capital for Quaker Oats, some additional discussion will be helpful in understanding the cost of capital and why it changes through time.

The cost of both debt and equity capital have built into them an inflation premium. For example, bondholders must be compensated for the loss in purchasing power and also receive some real rate of return. - The nominal interest rate, i, is comprised of two parts such that it is usually referred to as i = r + p where r is the real return and p is the anticipated rate of inflation over the life of the bond. Attached is a graph, which shows the nominal interest rate less the consumer price index, Note that the real rate has fluctuated within a very small band. Like bondholders, stockholders must also be compensated for the purchasing power lost through inflation. Thus, when estimating the cost of capital, care must be taken to explicitly state if this cost of capital includes or excludes the inflationary premium.

When the nominal rate of return is used for project evaluations or acquisitions, price increases should be assumed during the life of the project. If costs ate increasing at the same rate as the prices, the operating income will also be increasing at the same rate. The nominal rate of return will be above the real rate of return by approximately the rate of price increases assumed. For example, if a project indicated a DCF rate of return of 13% when prices were assumed to be increasing by 5% per year, the real rate of return would only be 8%.

The implications of this for project evaluations and acquisitions are obvious. First, if a discount rate, which included the inflationary premium, is used, it should be assumed that prices are increasing by the same amount as the premium. Second, if the discount rate does not include the inflationary premium, the cost of capital is much lower but no price increases should be assumed. Thirdly, as inflation and inflationary expectations change, the computation of the nominal cost of capital will be changing. Thus, it is dangerous to use one fixed rate for the cost of capital without periodic review.

For domestic acquisitions, there is general agreement that the seller's cost of capital is the appropriate rate of discount to be used for evaluation. For foreign acquisitions the exact same principles apply.

Because a small domestic acquisition has very little effect on the parent company and can be financed at our cost of capital, it is tempting to say that our cost of capital is the appropriate rate for evaluation. However, when one realizes that if this policy were continually followed, several small acquisitions would start to change the company's financial position and the risk, which the investment community associates with our stock. That is to say, a series of acquisitions could change the parent company's cost of capital. Likewise, one small foreign acquisition will have little effect on our financial structure and could be evaluated at our cost of capital. However, several such acquisitions would change our own cost of capital and the value of our stock. If one carried to an extreme the misguided logic of using our cost of capital for foreign acquisitions, almost all such acquisitions would look attractive. We would find that the foreign part of our business would be larger than our domestic, and our company's cost of capital would become the foreign cost of capital.

Most foreign countries have a capital scarce situation compared to the U.S. The cost of capital in these countries should be determined not by our cost of capital, but by rates prevailing in that particular country. The rate of return would tend to equalize if it were not for higher risk in foreign countries and the barriers to capital and labor mobility. The associated risks are the uncertainty of future tax rates, the possibility of nationalization, and a less stable political and economic climate. The appropriate discount rate for foreign projects and acquisitions is something which needs more study, but from very rough estimates which have been done, I believe that the cost of capital for foreign projects and acquisitions is considerably higher than in the U.S. Actually how much higher depends on the particular country. In Latin America I feel that the real cost of capital is approximately four percent above that in the U.S.

Calculations of Quaker's cost of capital have been made using three different approaches: (1) the "standard" cost of capital calculation which might be done by just plugging numbers into a formula without taking into consideration the current market conditions, (2) the cost of capital considering the current market conditions, and (3) the cost of capital in real terms.

Merrill Lynch calculates the "B" for Quaker's stock and most other listed stocks on a monthly basis. When our cost of capital was previously calculated, a "B" of .79 was used based on the December 1971 issue of "Security Risk Evaluation" published by Merrill Lynch. This "B" was based the previous five years using monthly observations. Using the previous 60 monthly observations for "B", and assuming the riskless rate of return is 6.0% and using the 9.3% that Fisher and Lone found to be the market rate of return, Quaker's equity cost of capital is:

Y = Rf + (Rm - Rf ) B

Quaker's current bond yield is approximately 7.5%, thus our after-tax nominal cost of capital using 8.6% as the cost of equity is:

Taking into consideration the current market considerations, there are two factors which have changed: (1) Quaker's "B", and (2) the market rate of return.

Quaker's "beta" has changed considerably in the last few years. Our acquisitions in the toys and recreation field, the growth of our Chemical Division, and our restaurant business have considerably changed the nature of our business and the sources of our profits. These areas are not as stable as the grocery products business and have increased the risk associated with ownership of our stocks. This is the same thing as saying that our "beta" and, thus, our cost of capital, have increased.

For evidence of this, several methods were used. First, an update of Merrill Lynch's "Security Risk Evaluation" was obtained from the Treasurer's department. The April 1973 issue showed our beta as .87. This reflected the price movement of our stock from March 1968 to March 1973. Even though the B had increased from the previously calculated .79, a large part of the recent calculation was from a period, which did not properly reflect the recent changes, which have taken place in the company. As a result, I calculated our B for increasingly shorter time periods to see what the result would be. Using data through May 31, 1973, if the time period is shortened to four years, the beta is .98, three years .95, two years 1.75 and one year 2.0. The high beta for the most recent year reflects the fact that since December 1972, the price of Quaker stock has dropped 22.5% while the S&P 500 has declined 11%.

One year or even two years is too short a period to accurately measure a company's beta. However, the above does yield some evidence that our beta and our cost of capital have increased in recent years. What is our "B" today? As can be seen, a case for almost any number could be made. A better perspective can be obtained by looking at our individual sources of profit. If the International sector is deleted, our planned F-74 profit is 49% from Grocery Products, 10% from Industrial and Institutional, and 41% from Toys and Recreational. Since 90% of Industrial and Institutional's profit will come from Chemicals, the "B" of the chemical industry would be a good approximation for that part of our business. The average"B" of sixteen chemical companies is 1.2. The average "B" of four toy manufacturers (Ideal, Mattel, Milton Bradley and Tonka) is 1.4. If we assume grocery products are a low risk, stable industry with a "B" of .8, we can calculate a weighted average "B" for Quaker Oats.

Quaker B = (Grocery Products%) * B + (Toys&Recreation%) * B + (Industrial&Institutional%) * B

Quaker B = .49 (.8) + .41 (1.4) + .10 (1.2)

Quaker B = 1.1

I feel that with the recent radical changes in the nature of our business and on the basis of all calculations made that a "B" of 1.1 for Quaker is a good estimate.

The second changed factor is the market rate of return. The Fisher-Lorie study was done for the years 1928-1965. During this period the average rate of inflation (CPI) was only 1.7%. Fisher and Lorie found the market rate of return to be 9.3, which means that the real rate of return over this period was 1.7% less or 7.6%. The recent inflation rates have resulted in an inflationary expectation far in excess of 1.7%. The average rate of increase in the consumer price index has been 4.1% since 1965. The current level of long term interest rates would indicate that investors are estimating a rate of inflation in the future of approximately 4 to 4-1/2%. If we assume a 4 rate of inflation for the future, then the nominal market rate of return will be equal to the real rate of 7.6% plus 4% or 11.6%. Using the new "B" of 1.1 and the new market rate of return, Quaker's cost of capital is:

Y = 6.0 + (11.6 - 6.0) 1.1 = 6.0 + 6.2 = 12.2%

After-tax Overall Cost of Capital

 

• B = Beta
• Rm = tbe market rate of return
• Rf = the riskiest rate is assumed 6%
• Best estimate is 9.5%

Finally, I would like to discuss the necessity for reviewing the cost of capital over time. As indicated above, the cost of capital of a company will change as the company changes. Very few companies have changed as radically as Quaker Oats has in such a short tine. However, as mentioned above, the cost of capital for the company is unimportant. The important thing is to estimate the cost of capital for individual projects and acquisitions. Real rates of return change very little over time. Thus, -the importance of reviewing the cost of capital is to be certain that the rate of inflation assumed corresponds to the rate of inflation assumed when the cost of capital was last calculated.

Date: July 3, 1973 .

To: Mr. Thomas J. Curley, Jr.

From: Peter B. Fritzsche & Corporate controller

Subject: Cost of Capital

Copies to: Mr. W F. Guinee, Senior Vice Pres

Mr. R. A. Bowen, Vice President

Mr. R D. Flenison, Vice President, Corporate Planning& Analysis

Mr. E. J Garneau, Manager, Project Evaluations

Mr. T. D. Henrion, Financial Analyst

We have reviewed Tom Henrion's a memorandum on the cost of capital and have the following comments.

We are in agreement with the proposition that a company's cost of capital should be reviewed from time to time to determine if it is increasing or decreasing and what effect this might have on our project evaluations.

In addition, we agree with the idea that for acquisitions in particular, it is the sellers' cost of capital (assuming Quaker's debt to equity ratio) that should be used as the minimum cut off rate, which in turn determines the maximum purchase price we could afford to pay. Finally, we are also in agreement with the proposition that foreign acquisitions tend to have higher costs of capital primarily because of higher costs of debt as well as higher market rates of return.

In terms of Quaker’s current cost of capital, we disagree with the assumptions made by Tom, although we agree with the methodology used. Specifically, our disagreements are as follows:

1.It's inappropriate to use a weighted average beta based on the corporate mix of business. The market will take into account the different businesses you're in and it's the long term, actual market beta that should be used, not a weighted average of the different industries in which you operate. Only a long-term beta should be used, because a short-term beta can be affected by unnatural and unrealistic factors. In Quaker's case, the appropriate beta to use would be the most recent four or five years average. The past two years, Quaker's beta (and that of many others) has been affected by the unrealistic and unnatural affects of the price freeze while at the same time the S&P 500 appears to have been much more stable. In Quaker's case the current beta that should be used should be either .87 or .98, not the weighted average of 1.1.

2. The most significant change in Quaker's cost of capital in Tom's analysis, however, is caused by the supposed increase in market rate of return from 9.3% to 11.6% to take into account the effects of inflation. The 9.3% rate is the market rate of return for the years 1928-1965. While the average rate of inflation during this period was only 1.7%, it did include several periods where the inflation rate was close to 4%, i.e., just after World War II and the Korean War. As such, the market rate of return of 9.3% includes periods of high inflation as well as periods of low inflation or actual deflation. The Fisher-Lone figure should be updated to include the latest period, but the long-term rate of return would not be significantly different from the 9.3%. As a matter of fact the market rate of return in the past three years, which have been periods of very high inflation, has been very poor in terms of stock prices. Such was not the case in prior periods of high inflation. In addition, the effects of inflation in the calculation of the cost of equity is taken care of in the short term risk free rate of return which is currently slightly in excess of 6%.

3. Based on our own calculations, the cost of capital for Quaker currently is 7.5%. The mathematics is as follows:

a) Cost of Equity. Y = 6.3 + (9.3 - 6.3) .98 = 6.3 + 2.9 = 9.2%

b) After-Tax Cost of Capital. c* = .68 (9.2%) + .32 (.5) (7.5%) = 6.3 + 1.2 = 7.5%

The real difference in our numbers comes from the fact that Tom is adjusting the cost of capital primarily through changes in market rates of return based on the current inflation rate, while our approach would be to have the current rate of inflation be one small piece of the long-term average market rates of return. As you can see the different assumptions have a very significant effect on our cost of capital.

Peter Fritzsche, Vice President

Business Development

Date: July 13, 1973

To: File

From: T. D. Henrion

Subject: Response to Peter Fritzsche on the Cost of Capital

There seems to be a growing agreement on the different issues associated with the cost of capital. This makes it necessary to concentrate only on the areas of contention. With that in mind, I address the issues, which Peter Fritzsche raised with respect to my cost of capital memorandum.

Using a weighted average beta is quite appropriate for estimating the cost of capital, especially considering the recent changes in the company's structure due to acquisitions. It is becoming more apparent among academicians and the investment community that using the beta calculated for an individual stock can be misleading. Most beta calculations have very high standard errors (that is, they are very unreliable) and can be affected by "unnatural and unrealistic factors" in the short run which will have enough impact on the beta to make it unreliable. Thus, there has been a tendency to look at betas from a "risk class" point of view.

I would, however, be in general agreement that using a period as long as five years should eliminate most of the distortions associated with random unnatural" movements in the price of a stock. On the other hand, one should never plug numbers into a formula and assume the results are reasonable. For evaluating the cost of capital for Quaker some adjustment must be made to take into consideration the fact that a large part of our business has changed.

On the basis of the estimates I ran for our beta, in a year or two the Merrill Lynch service will be estimating our beta at something higher than 1.1, which I estimated. At that time I will probably be arguing that the Merrill Lynch figure is too high. However, it will depend on the market conditions at that time, not on just what a formula says it is. As an example, the wage-price controls have had an injurious effect upon us and have probably had a tendency to raise our beta. If the danger of additional price controls is no longer present in a year or two, some adjustment should be made to our calculated beta.

The second area of contention is how to handle inflation. The point I was trying to make is that the cost of capital must be adjusted for inflation if we have estimates, which include inflation. The adjustment should have nothing to do with historical data, but should be adjusted in conjunction with the assumptions being made about future price increases used in the project and acquisition evaluations.

I believe Peter feels past rates of inflation are good predictors of future rates. While I don't feel this is true, the important point is to include inflation (regardless of the specific expected rate used) in the cost of capital and the project projections. If this is handled in a consistent fashion, the specific rate of inflation used doesn't matter. Any consistent treatment will result in the same answers.

The 6% used as the risk free rate and the 7.5% used for our borrowing rate both have inflationary expectations built into them. I am simply contending that the same adjustment needs to be made for the market rate of return.

Another argument is that the addition of Fisher-Price, Marx and, Needlecraft have reduced the company beta through diversification. From an investor's point of view this does not reduce the beta because this is diversification he could get himself through buying several similar stocks. Beta is the risk, which cannot be diversified away. These acquisitions reduce our beta only if the acquired company's beta is lower than our beta.

In summary, let me again emphasize that the company's cost of capital is a moot question; the specific rate of inflation to use is also a secondary consideration. What is important is the cost of capital we should use for project evaluations and acquisitions. In order to do this, two adjustments must be made. First, it is necessary to make an adjustment to the beta to take account of the market assigned risk to the type of business. Secondly, the cost of capital has to be adjusted to correspond to the inflation assumptions made in the projected sales.