The Same Two Trades, in Every Market on Earth

Value and momentum were discovered in US stocks, studied in US stocks, and argued about in US stocks. Which raises an awkward question that quietly haunts the whole field: are they real economic phenomena, or are they just two patterns that happened to hold in one country's stock market over one particular stretch of history?

There is only one honest way to answer that. Take the ideas somewhere completely different and see if they survive. Clifford Asness, Tobias Moskowitz and Lasse Heje Pedersen did exactly that, running value and momentum across eight diverse markets and asset classes at once. The answer, and the paper's title, is that they are everywhere.

The problem: one market is not a lot of evidence

Here is the uncomfortable statistical truth. Decades of US stock data sounds like an enormous sample, and in one sense it is. But for judging whether a strategy works, you effectively have one draw from history: one country, one economy, one path the world happened to take. Run enough searches over that single history and you will find patterns whether or not they are real.

The most powerful thing you can do to a suspected pattern is take it somewhere it has never been tested. If value works in Japanese equities, in currencies, in government bonds and in commodities, then it is much harder to argue that it was cooked up by staring at US stock charts.

The key idea via analogy: the same two questions, translated into every language

To test value and momentum outside stocks, you first have to define them outside stocks. What does "cheap" mean for the Japanese yen? What does "momentum" mean for copper?

Momentum is easy: it translates directly. Has this thing been going up recently? That question makes sense for a currency, a bond, a commodity, anything with a price.

Value took more thought. The insight is that value means the price is low relative to some slow-moving, fundamental anchor. For a stock, the anchor is book value. For other assets the authors use anchors appropriate to each: for a currency, how far the exchange rate has drifted from what purchasing power would imply; for a commodity or a bond, the price relative to its level a long time ago (roughly five years back), the idea being that a very long-run price acts as a rough stand-in for fundamental worth. The specifics vary, but the spirit is constant: is this asset cheap relative to a slow anchor, and has it been going up lately?

They applied these two questions to eight places: US, UK, continental European and Japanese individual stocks, plus country stock index futures, government bonds, currencies, and commodities.

The findings, and the beautiful one

Three results, and the third is what makes the paper famous.

One: both premia are almost everywhere. Value and momentum earned consistent return premia across all eight markets and asset classes. Both effects showing up in bonds, currencies and commodities, places where the original US-equity stories about distressed firms and analyst behaviour barely apply, is powerful evidence that something deep is going on.

Two: they share a global structure. Value in Japanese stocks is correlated with value in commodities. Momentum in currencies is correlated with momentum in European stocks. Value strategies across the world move together, and momentum strategies across the world move together, more than the underlying asset classes themselves do. Whatever drives value is a global force, not a local one. The authors capture this with a three-factor global model.

Three, the elegant one: value and momentum are negatively correlated with each other. Not slightly. Meaningfully. And this holds both within an asset class and across asset classes.

This makes intuitive sense once you say it out loud. Momentum buys what has been rising, which makes those things expensive, so momentum is naturally tilted away from value. Value buys what has fallen, which is precisely what momentum is selling. They are structural opposites.

The practical consequence is enormous. A 50/50 combination of value and momentum has historically been dramatically better than either strategy alone, because when one is having a terrible year the other is often having a good one. The diversification is not an accident, it is built into the logic of the two signals. This is the intellectual foundation of a huge amount of multi-strategy quant investing.

Why it mattered

• It survived the hardest possible test. Out-of-sample evidence in genuinely different markets is the strongest thing you can offer against the data-mining accusation, and this paper delivered it at scale.
• It made "combine your factors" a first principle. The negative correlation result is why almost no serious systematic manager runs pure value or pure momentum. Running both is not diversification for its own sake, it is exploiting a structural hedge.
• It pointed at a common global driver. The correlation of value strategies across unrelated asset classes suggests some shared force, and the authors point toward funding liquidity: when the money supply to arbitrageurs tightens, value strategies everywhere suffer together. That is a testable, economically meaningful story.
• It became a template. The methodology, define your signal in a market-agnostic way and run it everywhere, is now standard practice for validating any new factor.

The honest limitations

• The value definitions are not the same thing. Book-to-market for a stock and a five-year price ratio for a commodity are conceptually cousins, not twins. The paper is comparing analogies, and reasonable people can argue that the analogies are stretched.
• The authors have a commercial interest. All three are associated with AQR, which sells exactly these strategies. That does not make the results wrong, and the paper went through top-journal peer review, but a reader should note it.
• Correlated factors are also correlated in a crisis. The paper shows value strategies worldwide move together. That is presented as evidence of a common driver, but it is also a warning: your beautiful global diversification is less real than it looks when value has a bad year everywhere at once, which is what happened for much of the 2010s.
• Everywhere is not always. Both premia have gone through long, painful droughts since publication. Working across many markets does not mean working in all periods.

The one-line takeaway

Asness, Moskowitz and Pedersen took value and momentum out of US stocks and ran them on bonds, currencies, commodities and foreign equities, found that both work almost everywhere and that they reliably zig when the other zags, which is why combining them beats either one alone.