A Short History of Value Investing and its Implications

A great paper that provoked some tough reflections that are worth sitting down to sort out. Firstly, as a factor investor (in value, but also in size and quality), this paper has important implications for me. Damodaran had voiced the maxim that had made an impression to me, 'If you don't bring anything to the table, don't expect to take anything away from it.' Incidentally, I found out that Cornell had co-authored a paper with Damodaran, so it makes sense now why their thinking seems to align on this topic. In the old days, 'value investing' in the sense of sorting stocks by ratios was HARD - there was no data. Now, any idiot can systematically screen all stocks by ratios on an app on their phone. It cannot reasonably be expected to earn a risk premium from just factor investing by ratios. Sure, you can believe in Fama and French's story that value (factor!) stocks are riskier so you earn that risk premium, but Cornell doesn't address this story in this paper so I will leave that topic for another day. The bottom line is that, in light of the value factor's long stretch of underperformance, I must think hard about whether I believe investing in the value factor can earn me a premium. For now I believe that the likeliest scenario according to economic theory is that it should continue to earn a premium, just a much more moderate premium compared to its historical returns. Secondly, I had a question in 2022/ 2023 when I read Bruce Greenwald's 'Value investing'. I asked why do authors on value investing often demonstrate that value investing works by showing backtests of value stocks (which are defined by simple ratios like HML). But then, like Greenwald, they'd dedicate an entire book to arguing that it is insufficient to just screen by simple ratios, and true value investing is a rigorous fundamental analysis of the firm, its industry, and its future cash flows. Another thread of curiosity of mine can be traced to Cliff Asness's distinction between his 'quant value investing' and what Graham and Dodd folks call value investing. Over the years, 'value investing' has come to be such an incredibly loaded term that it can mean anything from Fama and French's value-factor-mimicking portfolio, to a Vanguard value factor ETF, to what an analyst at Baupost or Harris Associates or Oaktree does, or a crackpot retail investor spamming the 'value investing' label. Brad Cornell made me see the light: we only came to conceive of 'value investing' as accounting ratios like P/E - or maybe even more rigid screens like Graham's original net-nets, or price < 2/3 of tangible BV, or debt < 2x net current assets - via a historical accident. Put it this way, as the ultimate counterfactual, IF Graham was born in, say, 1960, so he was 19 when Dan Bricklin introduced Visicalc (the world's 1st electronic spreadsheet) in 1979, then value investing as the way many of us may know it (as looking for companies with low price ratios) would never have existed. Graham would have built his investing career by doing DCFs using spreadsheets, and he would never have written about ratios or screens in his books. Furthermore, Cornell led me to the light: value investing is just comparing your personal valuation of the stock to its market price. Period. A key insight in my view is that rich stocks like Apple or Nvidia can TOTALLY be considered value stocks if your personal valuation via a rigorous DCF is higher than the market price. This is why, in my view, 'true' value investing becomes an unfalsifiable philosophy (not necessarily in a bad way). If you define value investing as 'empirical quant value factor investing', then yes, its efficacy can be falsified because it becomes an empirical question - just plot the path of wealth of a value factor portfolio across time. And yes, you can definitively state whether the value factor is in demise, or not. But 'true' value investing cannot be falsified - although the value factor may be doing like shit, but the rich Mag 7 stocks that have seen such an incredible bull run could very well have been value stocks in the minds of the investors holding it, if their personal valuations were higher than their purchase prices. The only way to hypothetically falsify 'true' value investing is for God to get into the inner minds of every fundamental value investor, and do a rigorous audit over time of their strategy of buying when their personal valuation > the market price. If this audit shows that this strategy does not beat the market, then yes, you can say that you've proven value investing wrong. But since there is no way of doing this, there is no way of falsifying value investing. Thirdly, I had wondered why there was even a traditional distinction between value and growth investing. Presumably, no one would want to buy an asset for more than its value. Cornell provides me with a simple framework for determining this: the residual income equation, given by: P(t) = B(t) + [Residual income_(t+1) ] / (1+r) + [Residual income_(t+2) ] / r(1+r) + ... + Speculative value Whereby residual income is given by E[X(t+1) - rB(t)] If I understand him correctly, you can think of a growth stock as one whose value is mostly encapsulated in the speculative value term. But, as is Cornell's central insight, a stock with a large speculative value portion can still be a 'value stock'. It depends on the investor's personal valuation against the market price. Fourthly, I want to push back a bit on Cornell's conclusion that an active stockpicker MUST do a full-blown DCF in this day and age. I agree that value investing is just comparing your valuation with the market price, but then, the greatest adherents to this philosophy did not use spreadsheets! Buffett, Klarman, Howard Marks, Lynch etc did not sit down at their desks with MSFT Excel and did a DCF. My point is that the greats apparently derived their personal valuations with napkin math, not DCFs. Sure, you may argue that 'perhaps the greats were skilled enough to do a valuation in their heads without a DCF. Mere mortals like us need a DCF to discipline our thinking.' But then is there any (empirical) evidence that a spreadsheet DCF beats an intuitive fundamental analysis most of the time? If anything, spreadsheets lead to illusions of precision.