One interesting aspect of looking back several years of Berkshire's performance is to note how a steadily rising book value can partially alleviate the negative effect of failing to sell at periods of high valuation.

Going back through 26 open positions in Berkshire from February 2000 to September 2006 reveals only one position from March 2003 has failed to beat the S&P 500 index fund (SPY) with dividends added back but not assumed reinvested. My calculation is that the cumulative dollar return from all positions is about 32 percent higher than if the identical dollar amount had been invested in SPY at the closing price on the same dates as the Berkshire investments.

If I take Berkshire's reported book value at the time of each open position, the range of compounding of book value/share is from 8.6% to 10.4%. Actual gain in market value is lower due to P/B compression. It is quite a remarkable performance for such a large company ... posting high single digit/low double digit growth through a period including September 11, 2001, the housing bubble and bust, and the worst recession since the Great Depression.