For at least two decades, the "value premium" – the hallowed expectation that value stocks beat growth stocks – has been a mainstay of financial theory. But the value premium has failed to show up in recent years and disappointed investors are pulling billions of dollars from once irreproachable value managers.

So have we seen the last of the value premium?

The argument between growth and value investors, one of the oldest in finance, can be reduced to one question: What is a better bet, high growth stocks or low valuation stocks?

No, you can’t have both, but thanks for asking.

In the dark decades before the widespread availability of market data and the computing power to process it, that debate raged without a victor. But by the mid-1990s a consensus emerged: value trumps growth. Not always, but most of the time, and value’s win percentage climbs as the investment period grows.

If you had looked two decades ago at data compiled by finance researchers Eugene Fama and Ken French, you would have seen that the cheapest 30 percent of U.S. stocks by price-to- book had returned 13.7 percent annually to investors from July 1926 to December 1995, whereas the most expensive 30 percent of U.S. stocks had returned 9.6 percent over the same period. You would have also seen that the cheapest stocks had beaten the most expensive stocks 91 percent of the time over rolling 10-year periods.

You might then have wondered whether value’s best days were in the history books. Surely investors would begin to chase previously unloved value stocks and thereby diminish the valuation gap between value and growth – which, of course, had been the very source of value’s edge over growth.

Well, something funny happened on the way to value’s coronation – investors went on an epic growth binge, shunning value stocks in favor of fast growing Internet and technology companies. The S&P 500 Growth Index beat the S&P 500 Value Index for four consecutive years from 1996 to 1999, a margin of 12.9 percent annually in favor of growth during the period.

Investor’s abandonment of value in favor of growth from 1996 to 1999 only served to widen the valuation gap between value and growth.

One way to measure this valuation gap is to calculate the spread in earnings yield between value and growth. The average spread in earnings yield between S&P 500 Value and S&P 500 Growth has been 1.4 percent since 1998, with a standard deviation of 0.7 percent (calculated using five-year trailing average earnings from inception of the data). In 1999, that spread had climbed to 3 percent, implying that value was exceedingly cheap relative to growth at the time.