http://www.hussman.net/wmc/wmc081201.htm

The chart above is based on the actual dividends at each point in time, including during the Great Depression. The blue valuation line declines on the basis of this methodology because actual dividends were cut during the Depression, and the model gives full weight to that cut.

As with earnings, however, the stream of income that is realized over time tends to be far better behaved than year-to-year fluctuations might suggest. The value of that stream is particularly well behaved. The chart below shows the price that an investor would have paid for the S&P 500 in order for the actual, realized, subsequent dividends paid on the index to deliver a long-term total return of 10% annually (the current valuation assumes a growth rate of about 6% on normalized dividends from here).

http://www.contraryinvestor.com/mo.htm

An observation from the chart above that we do believe has implications for the current market is the history of nominal dollar S&P dividends. As you look back over the period of the 1910's through 1940's, there were a number of periods where equity market dividend yield spiked very significantly. We all know that was a result of a decline in equity prices as opposed to a massive increase in company dividends. But the important issue is that post these clear and significant spikes, aggregate equity yields dropped like a rock. Was the subsequent drop in S&P dividend yield a result of massive equity market rallies? Far from it. It resulted from huge drops in nominal dollar S&P dividends themselves. Companies either went bankrupt or cut dividends very meaningfully. What we've done in the following graph is to chart nominal dollar S&P dividends over the 1900-1950 period, and the 1950 to present period. Two different time frames representing two different "eras" in investor thinking and equity price discounting regarding dividends and yield, as we explained above. In the top clip you can see the cyclical contractions in actual nominal dollar corporate dividends that occurred with some regularity during the first half of the last century. No wonder investors demanded meaningful dividend yields as they priced equities during that period. Not only were actual equity prices volatile, so too was the stream of actual cash received by investors in the form of dividends.

In like manner, it appears that stocks are reasonably priced based on the dividend yield experience of the last six decades. But is the assumption of a static or growing cash dividend stream for the S&P in aggregate also a problematic assumption? Could we be embarking on a wider sector display of dividend cuts in 2009? We believe it's a very real possibility. So, we need to incorporate this into our thinking about macro valuations and company specific investment possibilities. We need to look well beyond perceptual dividend yields of the moment in terms of individual investments. The siren song of very high single digit or low double-digit yields may turn into an air raid siren of danger before 2009 has ended. Do not blindly accept dividend yields of the moment as being sacrosanct. What has transpired so far in 2008 has already taught us that nothing is sacrosanct. Nothing. We continue to believe the macro investment theme of yield is and will be important ahead. We're just walking in a more dangerous minefield near term. When it comes to yield oriented equities, just remember that quite simply, if a yield appears too good to be true, it probably is.

Are these two views as contradictory as they seem? Any way to reconcile them?

Alot of tough questions/issues with no easy answers. One issue the contrary investor note kinda raises is to what degree does the experience of 1950-present apply going forward.

This question is interesting to me because one thing I've been reading more about recently is the Kondratieff cycle. According to that theory, right around 1950 we entered a Kondratieff spring which was the beginning of the new Kondratieff cycle, and basically right now we've entered the Kondratieff winter.

If this cycle theory is correct, then basically 1950+ really isn't the best comparison to present, but you have to go back to the last Kondratieff winter of the 30s and 40s. Based on what I've read so far, these very long cycles of credit expansion and credit contraction are a key component of the Kondratieff cycle, and maybe we've started on a 10-20 year secular contraction in credit although it does look like the Fed and government are doing everything in their power to force additional credit into the system.

In terms of buying a stock for a dividend yield, maybe it makes sense to stick to companies with stable revenues (no cyclicals, industrials, materials) that are not dependent on robust economic growth for those revenues, and also to avoid companies that are dependent on credit issuance for sales (no big ticket consumer companies).