There are two hugely opposing forces in the valuation of the S&P 500.
The first force is that current earnings (89.6 est) are at multiples (1.51 times) of Robert Shiller's real 10 year average earnings (59.15) that have never been sustained for more than a few years and commonly only for less than a year and this can be taken as an indicator of risk to the market or overvaluation. Earnings divided by E10 turning down from above 1.2 times has been associated with market tops as can be seen from the chart below. This measure would suggest considering selling on negative price or momentum signals as the measure turns down.
The historical pattern since 1960:
We can look at percentiles of results of this ratio to give us the percentage of results of this calculation that are higher and lower than the current result. From this we can see that the current result is in the 97th percentile. That is 97% of results show the earnings of a period as being a lower multiple of the previous 10 years average real earnings. Average earnings can rise over time, but historically earnings also fall to cause a reversion to the mean, but maybe this time is different (but why should it be?). We are in an extreme situation regarding earnings compared to the historical averages. Note that this is true not only in relation to the period since 1960 but for the whole period since 1881.
2. Bond Yield Relativity
The second force is that the 10 Year bond PE (49 times) (the inverse of the 10 year bond yield of 2.04%) has rarely been at such a large multiple (3.483 times) of the S&P500 PE (14.07 times). Simplistically this reflects the current extreme risk aversion of investors. Over time this ratio could be expected to revert to towards the mean by one of three methods: bond yields will rise, earnings will fall or share prices will rise. Naturally a combination of all 3 is also possible.
This measure would suggest a once in a lifetime buying opportunity, unless of course you think earnings will fall significantly as analysed above or interest rates rise significantly. The chart below could be as much of a warning to holders of long term bonds once the current crisis in Europe passes.
The historical pattern since 1960 (Note how extreme the current situation is):
We can also look at this in percentile terms, which highlights how extreme is the current situation for the period since 1960:
Note that during much of the period before 1960 bonds were very much preferred over equities (when judged by the relative PE's), perhaps because of the crash of 29 to 33 and the patriotism during the two World Wars and the rebuilding after the Second World War. Looking in the context of the whole period since 1871, the current degree of preference for bonds is still obvious but much less extreme.
So all you need to know to make a decision is to know what will move first, earnings fall, interest rates rise, or stock prices rise in a new earnings paradigm to return yields to a more common relativity with bonds.
What is clear is that some long term relationships are at extreme positions, so volatility and high risk/potential reward will remain in the system.