The dangers of investing in cyclical stocks

While reading the Uttam Gawla rights issue document(http://www.sebi.gov.in/cms/sebi_data/attachdocs/1322823696255.pdf), I stumbled upon a good explanation of the sugar cycle, on which I build upon to explain the dangers of investing in cyclical stocks.


Domestic sugar industry typically follows a 4 to 6 years cycle. Profitability of sugar manufacturing units  depends largely on the stage of the cycle witnessed by the industry.Years of low production and declining sugar stocks may be followed by years of excess production that result in over-supply of sugar to the domestic markets, causing a decline in sugar prices and industry profit margins. Higher sugarcane and sugar production results in a fall in sugar prices and non-payment of dues to farmers. This compels the farmers to switch to other crops thereby causing a shortage of sugarcane, consequently an increase in sugarcane prices. And then, when limited supply causes sugar prices and industry profit margins to increase, sugar imports are usually liberalized. Import of white sugar then becomes an attractive option and which, in turn, would drop domestic prices and thereby impact sugar mills financial condition. But while the MSP of sugarcane is quick to increase with rising sugar prices, it is NOT adjusted downwards when the prices correct. Hence, the mills have it bad both ways.

So at its peak, with great earnings and profits, the valuations may seem attractive by any measure(P/E, P/BV or  dividend yield), but then when the peak slides, the same stock may look overvalued. But does data back this seemingly logical view?

Take a look at the graphs below. If the above view was to hold, then the cyclical commodity stocks should have seen intersections where high P/E accompanied a slide in the index. But that is not the case. Only once in Apr-09 did the blue P/E line cross the green index line, viz rising sharply after the index decline. Otherwise, it would be difficult to make out one chart from the other.

 
 
 So why is this happening? Is it because the various commodity cycles cancel out each other and bring the overall index to that of a reasonably broadbased level? Or is it because of survivor-ship bias etc? Since the data does not seem to back my analysis, I must now look for alternate explanations instead of discarding the analysis. Or else, do that analysis at the stock level instead of going into sector/broader perspectives.