Annual Data on US Stock Market
Prices, Dividends, Earnings, 1871-present
The Following is Chapter 26 (Data Appendix) from Robert J. Shiller, "Market Volatility," MIT Press, Cambridge MA, 1989, with most series updated:
26. Data Series
Tables 26.1 and 26.2. Basic Data
Notes on Data
An annual series of January values of the Standard and Poor Composite Stock Price Index starting in 1871 was used as the basis of empirical work in Chapters 1, 3, 5, 6, 8, 16 and 21. This series, Table 26.1 Series 1, was taken from Standard and Poor's Statistical Service Security Price Index Record, various issues, from Tables entitled "Monthly Stock Price Indexes Long Term."
The dividend and earnings series that correspond to this stock price series are spliced together from two sources. Starting in 1926, the nominal dividend series are dividends per share, 12 months moving total adjusted to index for the last quarter of the year. Starting in 1926, the nominal earnings series are earnings per share, adjusted to index, 4 quarter total, fourth quarter. These are from a table entitled "Earnings, Dividends and PriceEarnings Ratio Quarterly." of Standard and Poor's Statistical Service Security Price Index Record.
Standard and Poor's does not publish dividend or earnings series before 1926, however, their source for the Standard and Poor Index before 1926, a volume by Cowles , gives a dividend series corresponding to the index, series Da-1, pp 388389, which I multiplied by the ratio of the series in 1926 to adjust for change in base year.
A problem Cowles faced was absence of earnings data for many of the stocks in the Standard and Poor Composite Index. He thus presented series PEA-1 "prices of stocks for which Earnings Data are available, all stocks," a series of earnings E-1 on these stocks, and the ratio R-1 of these series, the "earnings-price ratio." We computed the Standard and Poor Composite Earnings series for years before 1926 as series R-1 (Cowles  pages 404405) times the annual average Standard and Poor Composite Index for the year. The spliced dividend and earnings series appear in Table 24.1 as series 2 and 3.
The absence of earnings data for some stocks is of some importance for the accuracy of the earnings series. One indication of the potential importance of their omission can be found by comparing the series P-1 (the Cowles index for all stocks) and the series PEA-1 (the Cowles series prices of stocks for which earnings data are available.) The ratio of P-1 to PEA-1 18711925 ranged from 0.98 to 1.27, the biggest discrepancies occuring in the earliest years of the sample. Another suggestion of the importance of the omissions is in Cowles list (, Appendix II, pp. 456475) of the companies in the index and the years for which the companies' earnings are available. Typically, lack of data on earnings comes for isolated years (as if earnings reports were occasionally missing) or for single years near the begin or end of the inclusion of the company in the index.
Wilson and Jones  have recently examined the Cowles data for accuracy. They found some apparent errors in Cowles' monthly series of cumulated returns (Cowles data implied negative dividends for some months) and produced an alternative monthly return series that attempted to correct these errors. They concluded that "the overall impact of these revisions as compared to the original Cowles Commission data is minimal." The Cowles monthly data that they criticise is not used here, returns are computed on a January to January basis assuming dividends are not reinvested during the year.
The nominal interest rate series, Table 26.1, Series 4 is the total return to investing for six months in January at the January 46 month prime commercial paper rate (six month starting January 1980) and for another six months at the July 46 month prime commercial paper rate (six month starting July 1980). (Starting 1998, 6-month commercial paper rate is replaced here by the 6-month certificate of deposit rate, secondary market.) It is computed as 100[1/((1Rjan/200)(1Rjul/200)) 1]. Data starting 1938 are from the Federal Reserve Bulletin. Data before 1938 are from Macaulay , Table 10, pp. A142A160.
The U.S. Bureau of Labor Statistics has been emphasizing the Finished Goods Producer Price Index, rather than the All Commodities Producer Price Index ever since the Producer Price Index replaced the Wholesale Price Index in 1978 (see Early ). The Finished Goods Producer Price Index is supposed to be superior since it is not affected by the double counting of intermediate and final goods that infects the All Commodities Producer Price Index. Unfortunately, the Finished Goods Producer Price Index is available only back to 1947. Any earlier series that might be spliced to it would be comparable to the All Commodities Producers Price Index. It was decided, therefore, to continue to use the All Commodities Index shown here, which the BLS makes available on a monthly basis back to 1913.
Starting with January, 1988 the BLS changed the reference base year for the producer price index from 1967 = 100 to 1982 = 100 (see "Producer Price Indexes January 1988," News, Bureau of Labor Statistics, Feburary 12, 1988). Table 24.2, Series 5 is the January Producer Price Index All Commodities with the 1982 base year starting with 1913. [Current data are available as series WPU00000000 on the Bureau of Labor Statistics web site: (http://stats.bls.gov:80/cgi-bin/surveymost?wp).]; Data from before 1913 are from the January figures from the Warren and Pearson Index , Chapter 1, Table 1, pp. 1114 divided by the ratio of the series in 1913.
Previously published papers in this volume use earlier producer price index series with 1967 = 100, Table 24.2, Series 6 and 7. These series are built up of component series, by multiplying the earlier series by the ratio of the indexes at the date of the first observation of the succeeding series.
Series 6 is the Producer Price Index for January. For 1947 to the end of the sample the series is the January Producer Price Index all commodities 1967 = 100 from the Survey of Current Business. For 1924 to 1946 the series used is January Wholesale Price Index (WPI) all commodities 1926 = 100 from the Federal Reserve Bulletin, divided by 1.933. For 1914 to 1923 the series used is January WPI all commodities 1913 = 100 from the Federal Reserve Bulletin, divided by 2.949. For 1900 to 1913 the series used is January WPI All Commodities 1890-1899 = 100 from Wholesale Prices, BLS Bulletin #149, Government Printing Office, Washington, 1914, divided by 4.1613. For 1871 to 1899 the series is the same as series 8 below.
The annual average producer price index Series 7 is, for 1947 the end of the sample, the annual average Producer Price Index all commodities 1967 = 100 from the Survey of Current Business. For 1924 to 1946 the series used is annual average WPI all commodities 1926 = 100 from the Federal Reserve Bulletin, divided by 1.9843. For 1914 to 1923 the series used is annual average WPI all commodities 1913 = 100 from the Federal Reserve Bulletin, divided by 3.038. For 1891 to 1913 the series used is annual average WPI all commodities 1913 = 100 from Wholesale Prices, BLS Bulletin #320, Government Printing Office, Washington, divided by 3.0395. For 1871 to 1890 the series used is annual average WPI 1890-99 = 100, from Appendix I of BLS Bulletin #114, divided by 4.1613.
The consumption deflator, 1972 = 100, (Table 24.2, Series 8) and real per capita consumption (Table 24.2, Series 9), both for nondurables and services, were used in Chapter 16 and 21, as well as Shiller  and Campbell and Shiller . The data sources are given in Shiller .
Campbell, John Y. and Robert J. Shiller. 1986. "The Dividend-Price Ratio and Expectations of Future Dividends and Discount Factors," Review of Financial Studies 1: 196228.
Cowles, Alfred, III and Associates. 1939. Common Stock Indexes, 2nd Edition, Bloomington, IN: Principia Press.
Early, John F. 1978. "Improving the Measurement of Producer Price Change," Monthly Labor Review 101: 715.
Macaulay, Frederic. 1938. Some Theoretical Problems Suggested by the Movements of Interest Rates, Bond Yields, and Stock Prices in the United States Since 1856, New York, National Bureau of Economic Research.
Shiller, Robert J. 1982. "Consumption, Asset Markets, and Macroeconomic Fluctuations," Carnegie-Rochester Series on Public Policy 17: 20338.
Warren, George F. and Frank A. Pearson. 1935. Gold and Prices, New York, John Wiley and Sons.
Wilson, J.W., and C.P. Jones. 1987. "A Comparison of Annual Common
Stock Returns 18711925 with 19261985," Journal of Business 60: