Stock market data
Stock market indicators
1. Amsterdam Stock Exchange, headline AEX index - price index, return index, dividends DY, earnings PE
(last updated 26 May 2016)
2. New York Stock Exchange, headline S&P 500 index - price index, return index, dividends DY, earnings PE
Note: There is no single convention for calculating dividend and earnings ratios. Besides the fundamental issue of trailing/backward and forecasted/forward values, institutions have different approaches to matters such as individual/aggregated values, 12-month or 4-quarter/latest quarter annualized/latest annual values, treatment of negative earnings, definition of (regular) cash dividends and various earnings concepts.
1. Euronext calculated AEX PE data unfortunately appear seriously out of step with other sources such as FT.
2. S&P quarterly EPS data can be used to fit to FT and Datastream sources, but only by lagging the trailing-four-quarter data by 2 months. (Implying quarterly earnings data reporting lag?)
Much referred to historical EPS/PE data provided in the so-called Shiller S&P500 dataset are slightly different from real world data. The interpolation of monthly EPS from quarterly EPS source data implies accurate forward looking ability by investors.
Long-run valuation ratios
Following work by Whetherilt and Weeken (2002) on the UK FTSE from 1927 and Shiller's dataset for the US S&P500 from 1871, I constructed a time series dataset for the Amsterdam Stock Exchange. The data include such popular equity valuation ratios as price-earnings, dividend yield and the Fed model.
Whetherilt and Weeken (2002) Equity valuation measures: What can they tell us? http://www.bankofengland.co.uk/publications/quarterlybulletin/qb020403.pdf
Dutch AEX price index and long-run valuation ratios (last updated 26 February 2014)
AEX – price, price-dividend, price-earnings
Long run
price-earnings
dividend yield
‘Fed-model’ EP – R
yield ratio DY/R
payout ratio DPS/EPS
Large versus small stock returns
Large versus high-dividend stock returns
US S&P 500 price index and long-run valuation ratios (last -partial- updated 26 February 2014)
Fundamental stock market model and equity risk premium
This section provides results from a reseach project on a reasonable stock market fundamental valuation model. Economists tend to focus on the stylized 1-stage dividend/earnings growth model they remember from the finance textbooks. This model simply does not work very well in practice. Multi-stage growth models provide a much better approximation. Not a new finding, but still largely ignored in the economic literature.
Following Panigirtzoglou and Scammell (2002) (and others before them) I use a 3-stage growth model and (so-called) long-term analysts' earnings forecasts. I have to do more work on the basic input variables and an empirical model for the time-varying equity risk premium.
Panigirtzoglou and Scammell (2002) Analysts' earnings forecasts and equity valuations http://www.bankofengland.co.uk/publications/quarterlybulletin/qb020106.pdf
(data last updated March 2011)
Amsterdam SE fundamental
See also my paper that resulted from this research project: Stock market fundamental valuation and the implied equity risk premium. August 2011.
Note: Recent patterns in earnings and trailing PE ratios indicate to me that I should revise the model and focus more on using (12-month) forward PE ratios. It adds to the data requirement wrt index earnings expectations, but will provide a more consistent valuation model in all periods. I am not sure how this change will affect the equity risk premium estimates.
Some troubling results on stock index dividend patterns
While doing preliminary research on so-called seasonal anomalies in stock indexes I found some disturbing results in the data. I don't think many economists are aware of these particular features of the data.
1. The nature of dividend reinvestment. Some so-called total return or reinvestment indexes are not what one would immediately expect. MSCI and Datastream indexes use a calculation that I call smoothed dividend yield gross up. They do NOT use actual daily dividend reinvestment. This may be an acceptable proxy for long-run investment returns, but it is not acceptable in empirical tests of the efficient market hypothesis. Researchers, beware.
2. Timing errors in the index data. Implied dividends calculated from stock price indexes and total return indexes -- with true dividend reinvestment -- show many instances of negative !! dividends. These appear to be related to timing problems in the price and total return index calculations (i.e. adjustments for stock splits, etc.). In turn this calls into question the reliability of stock index data for short term, i.e. daily, stock index research. Academic researchers will have to find a way to calculate improved index time series that correct the errors present in the commercial indexes.
The following file shows some features of the dividend data for the S&P500, FTSE-100 and Dutch indexes.
Stock index dividends
Puzzling implied dividends (D/P(-1), daily) calculated from AEX price index and total return index.
