The Vault

Trend Is Structural, Not an Inefficiency: Why It Cannot Be Arbitraged Away

“If this channel still exists, it was never an inefficiency.”

If markets truly eliminated persistent opportunities, trend following should not exist.

It has been documented for more than a century, traded at institutional scale for decades, studied exhaustively by academics, and deployed by some of the largest pools of capital in the world. If markets behaved the way equilibrium theory insists they do, trend should have been competed into irrelevance long ago.

It has not been.

That is not an embarrassment for trend following. It is an embarrassment for the inefficiency frame.

Trend is not an inefficiency.
It is a structural outcome of how markets function.

And that is precisely why it persists.

The Equilibrium Reflex

The dominant reflex in finance is to treat persistence as error. Prices are assumed to absorb information rapidly, competition is assumed to erase excess returns, and anything that survives must be a temporary anomaly waiting to be corrected.

Within this worldview, trend following is tolerated only conditionally. If it works, the explanation must be behavioural bias, slow learning, or limits to arbitrage that will eventually be overcome. The assumption is always the same. Markets are converging toward equilibrium. Structure is noise.

The problem is that real markets do not behave this way.

Markets are not closed systems solving for a single fair price. They are open, adaptive systems populated by heterogeneous participants with different objectives, constraints, horizons, and risk limits. There is no single optimisation problem being solved, and no stable equilibrium being approached.

In such systems, persistence does not signal inefficiency. It signals structure.

How Trends Actually Form

Trends emerge because capital does not move as a single, synchronised mass. Decisions arrive at different times, for different reasons, under different constraints. Trades vary in size, urgency, and impact. Price changes then feed back into behaviour, altering risk budgets, portfolio constraints, hedging needs, and mandate compliance.

When these forces align, price does not instantly jump to a new level. It migrates.

That migration through time is the trend.

This is not a mistake. It is how an adaptive system adjusts.

A Concrete Example: Bonds in 2022

Consider the global bond market in 2022.

Inflation surged, policy rates repriced, and bond prices fell sharply. But this was not a one-day adjustment to new information. It was a drawn-out cascade.

As yields rose, duration-heavy portfolios breached risk limits. Volatility-targeted strategies mechanically reduced exposure. Balanced funds rebalanced away from bonds. Pension funds faced collateral calls. Central banks shifted from buyers to sellers. Each step reinforced the next, not because anyone mispriced bonds, but because constraints forced action over time.

There was no single moment where bonds became “cheap” and snapped back. There was a trend.

No amount of arbitrage capital could short-circuit that process without absorbing the same balance sheet stress and volatility that caused it. The price path was not an error. It was the footprint of constraint-driven adaptation.

What the Evidence Confirms

Academic research on momentum confirms what market structure already implies. A comprehensive review of cross-sectional equity momentum shows that winner–loser effects have persisted across more than 150 years of data, across global equity markets, across thousands of portfolio constructions, and across decades of institutional adoption.

Most importantly, the evidence shows no meaningful post-publication decay. If arbitrage were capable of eliminating trend, it should have happened already.

This does not prove trend exists. Markets prove that every day. What the research does is remove the last refuge of the inefficiency argument.

But that same research also highlights an important boundary.

It studies cross-sectional momentum within equities. It shows persistence inside a stock universe. Trend following, as practiced by futures and macro investors, operates on a broader plane.

Trend Is Not an Equity Quirk

If trend were merely a behavioural artefact of equity investors, it should weaken or disappear outside stocks.

It does not.

Long-horizon research on global factor premiums documents persistent trend effects across futures, government bonds, corporate bonds, currencies, and commodities, spanning more than two centuries. These patterns appear before modern portfolio theory, before electronic trading, and before institutional asset management existed.

They survive changes in market structure, trading technology, regulation, investor composition, leverage regimes, and risk management practices.

That is not what inefficiencies look like.

When the same directional behaviour appears across assets, geographies, and centuries, the explanation cannot be local. It must be structural.

Why Arbitrage Cannot Remove Trend

Arbitrage requires three conditions. A clearly defined mispricing. The ability to hedge risk. And the ability to deploy capital without altering the system itself.

Trend violates all three.

There is no reference price that defines when a trend is wrong. There is no hedge that removes path dependence. And capital deployed into trending markets changes liquidity, volatility, and future flows.

More importantly, trend does not arise from errors. It arises from constraints.

Pensions rebalance on schedules. Volatility targetters de-risk mechanically. Trend-following systems scale exposure with price. Discretionary managers delay exits. Hedgers respond to price, not valuation.

These are not behavioural flaws. They are design features.

As long as capital is deployed under constraint and responds to price through time, trends must exist.

Capital Does Not Eliminate Trend

If trend were an inefficiency, greater participation would compress it.

Instead, we observe adaptation.

Trends fragment across horizons. They express differently across asset classes. They become noisier at the margin. But they do not disappear.

This is exactly what one would expect from a structural phenomenon embedded in an adaptive system. Trend followers do not create trends. They align with them. Their presence redistributes exposure through time rather than eliminating directional movement.

Trend as Emergent Structure

In complex systems, order does not arise from optimisation. It arises from interaction.

Rivers form channels as water responds to terrain. Traffic forms waves as drivers react to one another. Ant colonies form trails without maps or planners. None of these are inefficiencies. They are emergent structures shaped by flow, constraint, and feedback.

Markets are no different.

Trend is the channel carved by capital responding to changing conditions over time. It is the geometry of adaptation made visible in price.

Eliminating trend would require eliminating heterogeneity, delay, and feedback from markets themselves. At that point, markets cease to be markets. They become static pricing engines.

The Real Question

The real question is not whether trend will be arbitraged away.

The real question is why persistent structure continues to be treated as error simply because it does not fit equilibrium thinking.

Trend following does not work because markets are broken.

It works because markets are alive.

And as long as markets remain open, adaptive, and constrained by real-world capital, trend will remain a structural feature of price.

Not an inefficiency.

The shape that capital traces as it adapts.