Global

Country Momentum: Rotating Across Markets, Not Stocks

Saral Team · 1 August 2026

Every strategy in this catalogue so far asks “which stocks?” This one asks a different question: which market?

It turns out to be a more durable question than it sounds, and the reason has to do with what actually drives an index.

Countries trend for reasons companies do not

An individual stock’s price is a claim on a specific business, and the market’s estimate of that business can be revised in a single tick when earnings land. Information diffusion is fast and the mechanism for momentum in stocks — underreaction to firm-level news — is correspondingly fragile.

A country index is different. What moves it over quarters and years is a stack of slow-moving forces:

Monetary policy cycles. A central bank’s easing cycle runs for quarters, not days, and its effect on domestic asset prices compounds through the cycle.
Currency regimes. Sustained appreciation or depreciation reprices an entire market for foreign investors, which changes the flow, which reinforces the move.
Capital flow persistence. Foreign allocations shift gradually — an allocator deciding to raise emerging-market exposure does it over months through committee, not in an afternoon.
Terms of trade. For commodity exporters and importers, a commodity cycle transmits into the whole equity market over quarters.

None of these repricings happen in a tick. That is why country-level trends persist.

The evidence, and the out-of-sample point

Asness, Moskowitz and Pedersen’s “Value and Momentum Everywhere” (2013) tested both premia across eight markets and asset classes — including country equity indices — and found momentum in essentially all of them, with the returns positively correlated across classes, suggesting a common driver rather than eight coincidences.

The point that matters more for a practitioner is what happened after publication. Single-stock momentum has been heavily arbitraged: the universe is enormous, the data is cheap, and capacity is high, so decades of quant capital have compressed the premium.

Cross-asset and cross-country trend has held up better, and the reasons are structural rather than lucky. The universe is small — a few dozen liquid index futures — so capacity is limited relative to the capital that would want it. And the underlying driver is macro reallocation, which does not stop happening because someone published a paper about it.

Filter by country, not by exchange

A construction detail worth dwelling on, because it is a small decision that determines whether the strategy survives contact with a growing dataset.

The obvious way to build a world universe is to list the exchanges: XNYS, XTKS, XLON, XETR, and so on. It works, and it breaks the moment the data covers a market you did not list — the strategy silently keeps trading the old set, and nobody notices because there is no error.

The better construction asks for countries and lets the exchange-to-country mapping resolve the members:

countries: [US, JP, GB, DE, FR, HK, CN, KR, AU, CA, BR, SG, TW, IN]
security_types: [Index]

A new market becomes available and the universe picks it up as data, without an edit. This is the same principle that says a new country should require data rather than code branches, applied at the level of a single strategy’s universe.

It also means the strategy is not India-centric at all — the same pipeline, unchanged, would run on a purely European or purely Asian universe by changing a list.

Absolute trend, not just relative

The template requires a positive six-month trend before holding anything, not merely the best six-month trend.

The distinction decides drawdown behaviour. A purely relative screen ranks the universe and buys the top five — so in a global bear market, where every index is falling, it holds the five falling least, fully invested throughout. An absolute filter asks first whether the market is trending up at all, and holds cash when nothing is.

In a fourteen-market universe this matters less than it does in the two-asset crypto case, because there is usually somewhere trending. But 2008 and early 2020 were both periods when almost nothing was, and those are exactly the periods the filter is for.

What the backtest does not model

Three honest gaps, all of which understate real-world friction:

These are price indices in local currency. An Indian investor buying Japanese equity exposure earns the Nikkei’s move and the JPY/INR move. Currency can dominate over a one-year horizon, and this backtest does not model it.

An index is not directly buyable. You access it through an ETF, an index fund, or a futures contract, each with its own tracking difference, expense ratio, or roll cost. Treat the index series as the signal and expect the instrument to lag it.

Access is restricted for Indian residents. Overseas investment falls under the RBI’s Liberalised Remittance Scheme with an annual per-person limit, and international mutual fund schemes have periodically hit SEBI’s industry-wide overseas investment cap and stopped accepting fresh flows. The strategy is easy to express and not always easy to implement.

Try it

The Global Index Trend Rotation template holds the five strongest of fourteen country indices by six-month trend, requires the trend to be positive, sizes by inverse volatility, and rebalances monthly.

Worth doing: run it against the plain Nifty 50 over 2010–2025. India was one of the stronger markets over much of that window, so the rotation will often be holding it — and the periods where it rotated away are where the strategy earned its keep, or did not.