We assert that stock selection based on a single metric is an overly simplistic approach to investing—it is factor investing. The value factor, as introduced by [Nobel Laureate and professor of finance, Eugene] Fama and [professor of finance, Kenneth] French, is one of the two most prominent approaches to value investing. The other approach made famous by [the 'father of value investing', economist, academic and investor] Benjamin Graham takes a more extensive view of a company and requires not only quantitative analysis but also fundamental research as well as the critical element of human judgment. In other words, value is not a factor; it is a philosophy.

This paper comprises a reproduction of an interview with Rich Pzena on this topic.

From your perspective, what's the difference between the value factor and value investing as a philosophy?

Put simply, value investing means that you're trying to buy something for less than it's worth. Most people would say, 'That's what every investor does'. This may be true, but most investors don't really do detailed analyses to support their belief of what a company is worth.

They may buy simply because the stock's going up, because they think the incremental news flows will improve, or because they love the concept. This is buying based on characteristics, which is one way to invest. For instance, I could buy a company's stock because I think it will grow by 20%, or it has a high return on capital or a good management team, or because the products are good. But that doesn't mean it is a good value. Buying based on characteristics has historically been a bad way to invest.

What are some alternatives?

Well, there is buying based on momentum, which is a totally valid way to invest that's been proven academically. This approach assumes that the things that have done well will continue to do well.

And there's buying based on valuation, which is merely the fundamental calculation of its present value. If future cash flows were already known, then there would be no question that you could calculate a company's value and decide whether it's cheap or overpriced. So, if value investing means buying companies that are cheap based on their long-term cash flows, then to say that 'Value is dead', is absurd. That's like saying 'Arithmetic is dead'.

If valuations are mere mathematical calculations, then why do stocks get cheap?

Stocks generally get cheap for behavioural reasons, because something is going wrong currently, and investors don't believe that the long-term investment case remains intact. Or, there can be fear that something may go wrong in the future.

So how can you distinguish if a company's stock represents value?

There are some typical characteristics of companies that become cheap. When a business hits a problem, the market doesn't know if the issue is permanent or temporary. So, the stock gets sold off because there's discomfort in uncertainty.

As value investors, we accept that we may not know the full extent of the problem. But we can research a company, and the industry in which it operates, to understand what the business issues are that we believe will skew the potential range of outcomes in our favour.

The issues to consider vary by industry or company. Is there a long history of an established market position with high barriers to entry? Perhaps management expanded the business too rapidly, or hired too many people, or made an unreasonable acquisition. Or maybe a regulatory decision went against the company. It could just be the economic cycle. These are all examples of the things that value investors consider collectively.

The practice of value investing explores the range of things that could happen if a company succeeds at restoring profitability to its former lustre, the upside case, with the possible downside if the company fails. And what's the most likely outcome. If there's a big divergence between these two outcomes, then there's potential value. How can one replicate that with a factor-based model?

What do you say to people who believe that value investing is inherently risky?

Well, the original Fama-French study on price to book (P/B) worked off the inherent assumption that markets are efficiently priced. For low P/B stocks to outperform high P/B stocks, there's an implication that those with lower valuations have some risk that hasn't been identified. If risk is defined as stock price volatility, we would have to accept that value stocks may appear risky.

But what if we defined risk as the chance of losing money on an investment? By paying a low price for a stock that already reflects uncertainty, an investor inherently reduces risk. The reason people shy away from companies that have declined in value is they don't like uncertainty. I don't view that as a real risk. I'd argue that the best investors gain comfort in uncertainty.

Then isn't this the argument for factor investing and buying stocks at low valuations?

Anybody in the world can create a low P/B or a low price-to-earnings (P/E) portfolio. It doesn't take any skill. You can take a sophisticated quantitative model and combine multiple factors, which may even be better than a single factor. Maybe they are, but it's certainly possible to find a value stock that's cheap and still trades at a high P/B or high P/E, or both.

As an investor, you simply cannot look at numbers in a vacuum. I mean, P/B is flawed for obvious reasons; the book value is based on historical costs that don't reflect the reality of the company's true assets. Price to earnings is flawed because earnings can be highly cyclical. Just because a company's earning something today doesn't mean it will earn that in the future. Similarly, price to sales is just a snapshot of the company's sales and doesn't tell you anything about the margin structure of the business or its future sales. So, none of these things are value investing. They're factor investing.

In the end, just because you can buy a buggy-whip business [a business that fails to keep up with the times] at a low P/B or at a low P/E, it's still probably a bad idea. And you could have bought Google at a high P/B when it was clearly cheap.

When was that?

About five years ago. Everything was going mobile, and Google was a purely desktop search engine with a fledgling mobile presence. Apple was flexing its muscles and knocked Google Maps off the iPhone platform. Most people believed that either Apple or the carrier (AT&T or Verizon) would control all the advertising revenue. Google had a lot of cash and, net of its cash, traded down to just over 12x current earnings for a business whose share of the market was something like 85% and growing—in a world that was continuing to move toward digital.²

Yes, mobile was becoming increasingly important, but Google's valuation implied it would capture none of the mobile business or that the toll placed on it by Apple would eat away at Google's profits. And even if this assessment was 100% correct, an analysis of Google's share price versus its long-term ability to generate earnings still would have concluded that the stock was cheap.

So, what you're saying is that value investing is more nuanced than looking at a snapshot of a factor. A company like Google (Alphabet today), with its proven ability to generate long-term earnings, could be considered clearly cheap when priced at 12x earnings. Meanwhile, a business priced below book value may still be overpriced. Is that correct?

Of course. A factor model is merely a snapshot of a factor. It cannot replace human judgment or common sense. A good example would be buying a commodity company in the wake of the China-driven super cycle of the 2000s.

Consider what happened with iron-ore miners, who scrambled to add capacity when strong Chinese demand led to skyrocketing iron-ore prices. Throughout the 100-year history of the data, the real price of iron ore had been relatively stable, about [US] $50 per [US] ton. The cost to miners, too, had remained remarkably consistent at about $40 per ton. They typically pocketed a $10 profit margin. However, with the iron-ore price jumping to $180 a ton and the cost remaining at $40, miners' profit margins jumped to $140 per ton.

So, what did they do? They produced as much as they possibly could by expanding mines and getting more capacity online before the price fell.

In a situation like that, it's human nature to throw away all long-term investing principles. Although it usually costs $100 a ton to build a mine, it's suddenly going to cost $200 a ton because everybody's trying to develop at the same time competing for the same engineering and labour and everything else. But do they care about these added costs if they're going to make $180 a ton and $140 profit margin?

In the middle of that kind of a cycle, greed and optimism kick in, and people assume a new paradigm has materialised.

Yes, and the valuations get bid up as a result.

Eventually, the iron ore prices came down, and most of these stocks traded at less than book value. But remember that the miners were putting new assets on the balance sheet at an inflated $200 per ton versus a history of closer to $100 a ton. The book value of the new assets was twice the long-term economic value of the assets. To someone deciding to buy the stocks based on accounting book value, it looked like they were trading at a significant discount. But if you step back and ask if book value represents economic value, you arrive at a very different investment decision.