The investment industry has spent decades chasing alpha — the portion of returns supposedly independent of all systematic risk. Trillions of dollars in fees have been paid for it. Careers have been built on claiming to generate it. And yet a careful examination of the mathematics reveals something uncomfortable: in any meaningful sense, uncorrelated alpha does not exist.

This is not a nihilistic claim. It is a mathematical one. Every return stream, when decomposed with sufficient rigor, reveals embedded exposures to a small set of macro forces. What practitioners call "alpha" is, without exception, a conditional correlation to one or more of these forces — timed correctly. The skill is real. The independence is not.

Three Theories

The framework rests on three interlocking propositions. First, that mathematically absolute uncorrelated alpha is a theoretical impossibility — every portfolio carries systematic macro exposure whether the manager acknowledges it or not. Second, that any practical outperformance we observe is the result of a portfolio being correlated to one of exactly four macro drivers at the right time: the Chosen Market, its Contextual Inverse, the Real Risk-Free Rate, or Inflation. Third, that the increasing velocity of information in modern markets is destroying the efficacy of traditional slow-moving style factors, giving rise to a new paradigm centered on event-driven signals.

These three theories are not independent. The first establishes the impossibility condition. The second identifies the actual mechanism through which outperformance occurs. The third addresses the practical question of how to achieve the desired correlation in a world where old tools are decaying.

The Four Horsemen

If all returns decompose into four macro forces, what are they? We propose a parsimonious set we call the Four Horsemen of Macro Risk. Two are context-dependent — they change depending on what strategy you're running. Two are universal — they govern every strategy in every asset class.

The Chosen Market is whatever benchmark you play against. It is not fixed. For a long-only equity PM, it might be the S&P 500. For a macro trader, it might be a volatility surface. The Contextual Inverse is the natural hedge to that chosen market — the asset or return stream that is maximally negatively correlated during tail events. For the S&P 500, this is the VIX. For long-duration bonds in a rising rate environment, it is cash. The alpha opportunity, to the extent it exists, lies in the dynamic management of exposure between a market and its inverse.

The two universal forces — the Real Risk-Free Rate and Inflation — set the macro backdrop against which every market/inverse pair operates. They govern discount rates, hurdle rates, and the real value of nominal returns. No strategy escapes their influence.

Why This Matters

If this framework is correct — and the empirical evidence across equities, fixed income, commodities, and currencies is strongly supportive — then the implications for portfolio construction are profound. It means that portfolio managers should stop pretending they can generate returns independent of macro forces and instead focus on understanding which forces they are exposed to, whether those exposures are intentional, and whether they are being compensated for bearing them.

It also means that the tools for generating outperformance need to evolve. If traditional factors are decaying and the opportunity lies in timing correlation to the Four Horsemen, then the investment process must be built around event-driven signals that capture the transient dislocations created by new information arriving in markets. This is the practical engine that translates the theoretical framework into portfolio returns.

The full working paper provides formal derivations, empirical validation across multiple asset classes and time periods, and a detailed treatment of the hierarchical three-layer portfolio framework that emerges from the synthesis.

For three decades, factor investing was the closest thing quantitative finance had to a free lunch. Buy cheap stocks, sell expensive ones. Ride momentum. Harvest the size premium. Academic research provided the intellectual scaffolding, and the asset management industry built hundreds of billions of dollars in products on top of it.

That era is ending. The evidence of factor decay is now overwhelming, and the mechanisms behind it are well understood. When a profitable anomaly is published in an academic journal, capital flows in to exploit it. As more money chases the same signal, the returns compress. What was once a 5% annual premium becomes 2%, then 1%, then noise.

The Numbers

Consider the Fama-French value premium (HML). In the original sample period, it delivered statistically significant excess returns. In large-cap universes since 2007, the t-statistics have been insignificant. The premium has not disappeared entirely, but it has compressed to the point where implementation costs — transaction fees, market impact, rebalancing drag — consume most of what remains.

Momentum (UMD) tells a similar story with an additional wrinkle: it is prone to catastrophic crashes. The momentum crash of 2009 wiped out years of accumulated returns in weeks. Cross-sectional momentum exhibits elevated tail risk that becomes more dangerous as more capital crowds into the strategy.

The size premium (SMB) may never have been a real factor at all. It fails robustness checks across international markets and post-1980 subperiods. The residual returns are largely attributable to illiquidity compensation — you're not being paid for bearing size risk, you're being paid for accepting that your positions are hard to trade.

Why This Is Structural, Not Cyclical

Some argue that factor underperformance is cyclical — that value will come back, that momentum always recovers. This misses the point. The decay is structural, not cyclical. It is driven by three forces that are accelerating, not mean-reverting.

First, publication itself destroys the edge. Once a factor is documented in an academic paper, thousands of portfolio managers and quantitative researchers can implement it. The information advantage evaporates. Research suggests that anomaly returns attenuate by roughly a third in the years following publication, with the decline attributable to informed capital reallocation.

Second, the proliferation of factor products has created crowding at industrial scale. When hundreds of ETFs and smart-beta products are all rebalancing into the same names on the same schedule, the entry and exit prices become unfavorable. The factor premium is transferred from the investors to the market makers.

Third, the increasing speed of information processing in markets means that any edge based on slowly-updating characteristics — book-to-market ratios, trailing momentum windows, market capitalization — is structurally disadvantaged against strategies that process information in real time.

What Comes Next

If static, cross-sectional factors are decaying, where does the opportunity lie? We believe the answer is in event-driven signals — patterns that are triggered by specific, identifiable market events rather than persistent stock characteristics. An event-driven factor is not "this stock is cheap" (a static attribute). It is "this stock just experienced a specific type of information event, and historically, similar events have been followed by a quantifiable price response over a specific time horizon."

Event-driven signals have structural advantages over traditional factors. They are harder to crowd because they are path-dependent and regime-contingent — two implementations of the same event signal in different macro environments will produce different results. They are harder to arbitrage away through publication because they require a continuously maintained library of event-pattern associations, not a simple sorting rule. And they operate at the timescale at which modern markets actually process information: hours and days, not quarters and years.

The factor zoo is collapsing. What replaces it will look very different from what came before.

Ask a portfolio manager what drives their returns and you'll get a story about stock picking, sector bets, or timing. Ask them what their macro exposures are and you'll often get a blank look. This disconnect — between the narrative of idiosyncratic skill and the reality of systematic exposure — is the central problem in institutional investing.

The Four Horsemen framework is an attempt to resolve it. The claim is simple: every return, in every asset class, decomposes into loadings on exactly four macro forces. Two are context-dependent (they change based on what strategy you're running). Two are universal (they affect everything). The decomposition is exhaustive.

The Context-Dependent Pair

The first horseman is the Chosen Market — the benchmark against which performance is evaluated. This is not necessarily a standard index. For a macro trader, the "chosen market" might be the level of implied volatility. For a credit portfolio, it might be the IG spread curve. The chosen market is whatever return-generating process the portfolio is designed to interact with.

The second horseman is the Contextual Inverse — the asset or return stream that is maximally negatively correlated with the Chosen Market during periods of stress. This is the natural hedge. For an equity portfolio benchmarked to the S&P 500, the contextual inverse is the VIX. When the market sells off, the VIX spikes. A portfolio that is intelligently allocated between the market and its inverse captures the essence of what we call "alpha" — though as we've argued elsewhere, it is more accurately described as a conditional correlation timed correctly.

The key insight is that the market/inverse pair is not fixed. It is defined by the portfolio's objective. Change the objective and you change the pair. A Treasury bond portfolio has a completely different contextual inverse than an equity portfolio. This context-dependence is what makes the framework generalizable across asset classes.

The Universal Pair

The third horseman is the Real Risk-Free Rate — the inflation-adjusted price of capital over time. This is the discount rate applied to every future cash flow in every asset class. When real rates rise, the present value of all long-duration assets falls, regardless of what the underlying business or economy is doing. Real rates modulate the equilibrium relationship between every market/inverse pair by changing the opportunity cost of capital.

The fourth horseman is Inflation — the stochastic erosion rate on nominal returns. Inflation functions as a universal hurdle: the minimum return threshold any strategy must exceed to generate positive real wealth. An equity portfolio returning 8% nominally in a 7% inflation environment has generated almost no real value. Inflation affects every asset class, every strategy, every time horizon.

Why Four and Not More?

There is an extensive academic literature proposing five, six, or dozens of factors. Why do we claim four is sufficient? Because the Four Horsemen operate at the most fundamental level of the return-generating process. Other factors — value, momentum, quality, size — are not independent forces. They are conditional expressions of exposure to the Four Horsemen, filtered through specific market structures and investor behaviors. A "value" premium, for example, tends to perform well when real rates are stable and inflation is moderate — it is a conditional loading on Horsemen III and IV, not an independent force.

The parsimony of the framework is a feature, not a limitation. Four forces are enough to explain the macro structure of returns across equities, fixed income, commodities, and currencies. Everything else is either a derivative of these four forces or noise.

Practical Implications

If you accept this decomposition, portfolio construction changes fundamentally. The first question is no longer "what stocks should I buy?" but rather "what are my exposures to each of the Four Horsemen, and are those exposures intentional?" The framework provides a common language for evaluating portfolios across asset classes and strategies — a treasury portfolio and an equity portfolio can be analyzed in the same terms.

For the investment team at 42, the Four Horsemen serve as the macro conditioning layer for all signal evaluation. Before any event-driven signal is acted upon, it is evaluated in the context of the current macro regime defined by these four forces. A signal that works in a low-rate, low-inflation environment may not work in the opposite regime. The framework prevents the team from applying signals outside of the conditions under which they have historically been valid.

Traditional quantitative investing is built on a simple idea: sort stocks by some characteristic — cheapness, momentum, quality — buy the top decile, sell the bottom, and harvest the spread. For decades, this worked. It doesn't anymore, at least not reliably. The question is what replaces it.

We believe the answer lies in a fundamentally different approach to signal generation — one rooted in market microstructure theory and the study of how new information is absorbed into asset prices through the trading process.

What Is an Event-Driven Factor?

An event-driven factor is a quantifiable, repeatable pattern in an asset's price behavior that is triggered by a specific, identifiable "experience." This is a fundamental departure from traditional factors. A traditional factor like value is a static, cross-sectional attribute: a company is cheap relative to its peers. An event-driven factor is dynamic and time-series in nature: a company experiences a specific type of information event, and there is a quantifiable price response that follows.

The critical distinction is that the alpha opportunity generated by an event is not a permanent risk premium. It is a temporary dislocation that exists only until the new information is fully absorbed and priced by the market. This duration can range from hours to weeks, reflecting the speed at which modern markets process information.

Why Events Are Harder to Arbitrage

Traditional factors decay after publication because they are simple sorting rules that anyone can implement. Buy stocks with low P/E ratios. Buy stocks with high 12-month trailing returns. Once published, thousands of participants can replicate the strategy, and the returns compress.

Event-driven signals are structurally more resistant to this dynamic for three reasons. First, they are path-dependent: the same "type" of event produces different results depending on what happened before it and what the current market conditions are. You cannot just sort stocks by a single characteristic. Second, they are regime-contingent: the Four Horsemen macro framework conditions which event patterns are active and which are dormant. An earnings surprise in a low-volatility, low-rate environment produces a different price response than the same surprise in a high-volatility, tightening environment. Third, they require continuous maintenance of a large-scale event-pattern library — not a formula, but an evolving knowledge graph of associations between information events and market responses across asset classes.

Cross-Asset Relationships

Perhaps the most powerful aspect of the event-driven paradigm is that it naturally extends across asset classes. Traditional factor models are typically confined to a single universe — an equity factor model knows nothing about credit markets, and a fixed income model knows nothing about equities. Events, by their nature, propagate across boundaries.

Consider a real example: water level data from the Panama Canal affects transit capacity, which affects maritime freight rates, which affects the profitability of specific shipping companies, which affects equity prices and CDS spreads in the energy and commodity transportation sectors. A traditional factor model would never surface this relationship. An event-driven system that maintains cross-asset associations can.

These multi-hop causal relationships — where an information event in one domain creates a predictable price response in a seemingly unrelated domain — represent a category of signal that is essentially invisible to conventional quantitative approaches. They require both the data infrastructure to capture events across domains and the analytical framework to evaluate whether the historical pattern is likely to repeat under current conditions.

The Role of Discretion

A natural question is whether event-driven signals should be traded automatically. Our view is that they should not — at least not entirely. The signal library surfaces high-conviction opportunities and provides the PM team with conviction scores, reasoning traces, and historical context. But the final decision on whether and how to act belongs to the investment team.

This is not a concession. It is a design choice. Event-driven signals operate in a space where context matters enormously. A signal that is historically valid may be inappropriate given information the model does not have — a pending regulatory decision, a private conversation with management, an emerging geopolitical situation. Human judgment adds a layer of contextual evaluation that no model, no matter how sophisticated, can fully replicate.

The hybrid model — AI-generated signals with discretionary execution — combines the breadth and speed of machine learning with the judgment and accountability of experienced portfolio managers. We believe this is where the industry is heading, and it is the architecture 42 Capital is built on.