home *** CD-ROM | disk | FTP | other *** search
- A Spreadsheet Using Some Historical Market Data
- By Fred Shipley, Ph.D.
- Computerized Investing, July/August 1989
-
- In this spreadsheet we use data for the Dow Jones Industrials
- to understand certain fundamental market relationships. With the
- data, in columns A through F, you can create some formulas and
- observe a number of characteristics of the Dow Jones Average.
- These long term historical characteristics provide a means of
- judging the market's current position.
-
- This spreadsheet uses some information about the Dow Jones
- Industrials that appears regularly in Barron's. Although this
- table is printed with some regularity, there is no set schedule,
- so you will have to watch for it to appear. Since most of the data
- is annual, there is no need to perform frequent updating. Data for
- the previous calendar year is usually available by May.
-
- The basic data is reproduces in columns A through F and
- examples of possible calculations are given in columns G through
- M. For comparative purposes, some historical averages and standard
- deviations are calculated near the bottom of the worksheet in rows
- 71 and 72.
-
- In the spreadsheet, the return on equity (column G) is
- current earnings per share divided by book value.
-
-
- Earnings Per Share (column D)
- ROE = --------------------
- Book Value (column C)
-
-
- The price-earnings ratio (column H) is a trailing ratio
- because it takes the index value divided by the last 12 months'
- earnings per share. For comparative purposes, we have also
- computed a normalized P/E (column I), which is the current index
- value divided by the following year's expected earnings.
-
- Closing Price (column B)
- Trailing P/E = --------------------
- Earnings Per Share (column D)
-
- Closing Price (column B)
- Normalized P/E = --------------------
- Earnings Per Share (column D)
- (Following Year)
-
-
- In some years, the calculated price-earnings ratios could be
- either unusually large or negative. For example, in 1932, the
- depths of the Depression, total earnings for the industrials were:
- $0.51.
- Using this figure would result in a trailing P/E of -117.5:
-
- 59.93
- -----_ = -117.5
- -$0.51
-
- Clearly, such a number is not meaningful, and negative P/Es
- are reported as such in the financial media. As another example,
- in 1982 earnings were $9.15 and the Dow closed at 1046.54, which
- gives a trailing P/E of 114.4. Again, this P/E is an aberration-
- -what a statistician would call an outlier.
-
- The problem for an investor is what to do about these
- numbers. If you were to calculate an average P/E, the average would
- be distorted by including these outliers. Since you are interested
- in learning something about long-run performance, determining an
- average value and the variation around that value is critical. As
- a reasonable rule of thumb, we will simply exclude for any
- calculation P/E ratios that are negative or greater than 30.
- Essentially this excludes the aberrations caused by the Depression
- (1932 and 1933) and the more recent sharp recession in 1982.
-
- To exclude these numbers, we use the database average
- function, @DAVG. This function allows us to determine an average
- of those values that satisfy the conditions, or criteria, we
- specify. The form of the function is:
-
- @DAVG(input range, offset, criterion range)
-
- where: input range is the area of the spreadsheet where
- the data is located, for trailing P/E, cells H7 to H67;
-
- offset counts the number of columns from the beginning of
- the database, starting with zero. Since we have only one
- column of data, the offset in this case is zero.
-
- criterion range is the area of the spreadsheet in which we
- specify our conditions. In this case, we have our criteria
- immediately under the data we are analyzing, in cells H69
- and H70 for trailing P/E.
-
- For our trailing price-earnings ratios, the formula is:
-
- @DAVG(H7..H67,0,H69..H70)
-
- Certain quirks of 1-2-3 must be recognized when using this
- database function. First, the input range must include the
- identifier of the data field we are analyzing. Second, the
- function does not work correctly of there is a blank cell or label
- between the field name and the first item of data. (Don't ask ne
- why; I have no idea! It's silly and often causes trouble.) This
- second reason is why we have an entry in the row for the year 1928,
- even though there is no data for that year other than the closing
- value of the Dow. Third, it makes a difference whether you use
- absolute or relative addresses in the criterion. All criteria
- should be written with relative addresses.
-
- Finally, we have to specify our criteria. There are a number
- of ways to do this. We will make a single formula, using the
- logical AND. For example, the average of P/Es that are positive
- and less than or equal to 30 in our criterion would be
- +H8>0#AND#+H8<=30. For other ways to set this up, check one of the
- many Lotus books, such as Que's "Using 1-2-3."
-
- The dividend payout ratio (column J) is the percentage of
- earnings actually paid out in dividends. It is defined as
- dividends per share divided by earnings per share.
-
- Dividend Dividends Per Share (column E)
- Payout = -------------------
- Ratio Earnings Per Share (column D)
-
- The earnings retention ratio (column K) is the percentage of
- earnings that are reinvested by the companies in the Dow Jones
- index. It is usually thought of as one minus the dividend payout
- ratio.
-
- Earnings Earnings Retained Per Share
- Retention = ---------------------------
- Ratio Earnings Per Share
-
- = 1 - Dividend Payout Ratio
-
- = 1 - Column J
-
- The dividend yield (column L) is the dividend (current cash
- flow) return as a percentage of the year-end value of the Dow Jones
- index.
-
-
-
- Dividends Per Share (column E)
- Div. Yield = -------------------
- Year-End Value (DJIA) (column B)
-
- Total return (column M) is the change in price plus dividends
- per share divided by the closing price of the previous year.
-
- (DJIA1 - DJIA0) + Dividend1
- Total Return = ---------------------------
- DJIA0
-
- Where DJIA1 is this year's price (column B)
- DJIA0 is the previous year's price (column B)
- Dividend1 is the dividend
- received during the year (column E)
-
- The change in CPI in column F is the annual percentage change
- in the Consumer Price Index. This data is taken from the Federal
- Reserve Bulletin and is reported regularly in the media.
-
- As you would expect, there is an inverse relationship between
- the rate of inflation and the price-earnings ratios. When
- inflation increases, so does investors' requited rate of return.
- This increase in the required rate of return means that investors
- are willing to pay less for a given level of earnings, since those
- earnings are inflated. The result is a lower P/E.
-
- There are obviously many other such comparisons you can make.
-
- Examining the dividend yield on a regular basis, for example, would
- have indicated that a historically low yield prevailed in August,
- 1987. Other possibilities are limited by your imagination.
-
- (c) Copyright 1989 by the
- American Association of Individual Investors