Measuring Strategy-Decay Risk: Minimum Regime Performance and the Durability of Systematic Investing

This paper introduces Minimum Regime Performance (MRP) as a novel quantitative framework for assessing the durability of systematic investment strategies by measuring their lowest risk-adjusted returns across distinct market regimes, thereby addressing the often-overlooked risk of strategy decay caused by changing market conditions.

The Gist

The Big Idea: The "Worst-Case" Test for Money-Making Machines

Imagine you have a car. You want to know how good it is. Usually, you look at its average speed over a long road trip. If the average is 60 mph, you think, "Great car!"

But what if that car only goes 60 mph on smooth highways, but slows down to 5 mph on gravel roads, and stalls completely on snowy hills? If you only look at the average, you miss the fact that the car is actually very fragile. It works great in some conditions but fails miserably in others.

This paper is about a new way to measure investment strategies (the "cars") not by their average speed, but by their worst performance in different weather conditions.

The Problem: The "Alpha Decay" Trap

In the world of investing, people use computer programs (strategies) to find patterns in the stock market to make money. This is called "Systematic Investing."

For a long time, investors looked at the Sharpe Ratio. Think of this as the "Average Speed" of the car. A high Sharpe Ratio means the strategy made good money compared to the risk it took.

The Catch: Markets change. Sometimes it's sunny (bull market), sometimes it's raining (bear market), and sometimes there's a blizzard (crisis).

• A strategy might work perfectly in the sun but crash in the rain.
• Over time, as more people discover a "good" strategy, they all start using it. This crowds the market, and the strategy stops working. This is called Alpha Decay (the money-making power slowly rots away).

Traditional tools (like Volatility or Drawdown) tell you how much money you might lose on a bad day. They don't tell you if the engine itself is broken or if the strategy just doesn't work in the current weather.

The Solution: Minimum Regime Performance (MRP)

The authors, Nolan Alexander and Frank Fabozzi, invented a new metric called Minimum Regime Performance (MRP).

The Analogy: The "Survival Test"
Imagine you are hiring a chef.

• The Old Way (Sharpe Ratio): You ask, "What is your average rating over the last 10 years?" If the answer is 4.5 stars, you hire them.
• The New Way (MRP): You ask, "What is the lowest rating you ever got?"
  • If the chef gets 5 stars for 9 years but 0 stars (serves raw chicken) during one specific month, their MRP is 0.
  • The MRP tells you: "This is the worst the chef has ever performed. If the market turns into that specific bad month, this is how bad your results will be."

How it Works:

1. They take a long history of a strategy's performance.
2. They chop that history into different "Regimes" (different market weather: e.g., high inflation, low interest rates, tech bubbles).
3. They calculate the performance for each piece of history.
4. They pick the lowest number. That is the MRP.

It answers the question: "How low can this strategy go when things go wrong?"

The Big Discovery: The "Durability Trade-Off"

When they tested this on many famous investment strategies, they found a surprising pattern, which they call the "Decay-Risk Frontier."

Think of it like buying a car:

• Fast but Fragile: Some cars are incredibly fast on the highway (High Average Sharpe) but fall apart on gravel (Very Low MRP). These are strategies that look amazing on paper but are dangerous because they rely on specific conditions that might disappear.
• Slow but Tough: Some cars are a bit slower on the highway but handle gravel, snow, and mud perfectly (Lower Average Sharpe, but High MRP). These are "boring" strategies that keep working even when the market changes.

The Lesson: Just because a strategy has a high average return doesn't mean it's safe. In fact, the strategies with the highest average returns often have the lowest durability. They are the most likely to "decay" or stop working when the market shifts.

Why This Matters to You

If you are an investor (or someone managing money for others), this paper suggests you should stop obsessing over "How much money did we make on average?" and start asking: "How long will this strategy keep working?"

1. It's a Warning System: If a strategy's MRP starts dropping, it's like a "Check Engine" light. It means the strategy is becoming fragile and might stop working soon, even if it's still making money today.
2. Better Portfolio Building: Instead of just picking the fastest cars, you should build a fleet that includes some "slow but tough" cars. This ensures that no matter what the market weather is, your portfolio keeps moving.
3. New Kind of Risk: We used to worry about the market crashing (Market Risk). Now, we have to worry about our tools breaking (Strategy Decay Risk). MRP helps us measure that.

Summary in One Sentence

Don't just look at how fast a money-making strategy has been in the past; check how slow it gets in its worst moments, because that "worst moment" is the only thing that tells you if the strategy will survive when the market changes.

Technical Summary

1. Problem Statement: The Vulnerability of Systematic Strategies

The paper addresses a critical, often overlooked risk in systematic investing: strategy-decay risk (or alpha decay).

• The Core Issue: As systematic strategies (factor models, machine learning) proliferate, the underlying market relationships they exploit often weaken, become crowded, or disappear due to market adaptation and data mining.
• Limitation of Traditional Metrics: Standard risk measures like volatility, Value at Risk (VaR), and drawdown quantify the variability of outcomes but fail to capture the structural degradation of the investment process itself. They assume stationarity, whereas modern markets are non-stationary.
• The Gap: Investors lack a quantitative metric to distinguish between a strategy that is merely volatile and one whose fundamental efficacy is eroding across different market regimes.

2. Methodology: Minimum Regime Performance (MRP)

The authors introduce Minimum Regime Performance (MRP) as a diagnostic framework to quantify the durability of a strategy.

A. Conceptual Definition

MRP is defined as the lowest realized risk-adjusted return (typically the Sharpe Ratio) across distinct historical market regimes. It represents the "worst-case" structural environment in which a strategy has operated.

• Analogy: Just as Expected Shortfall (ES) measures the tail risk of returns, MRP measures the "tail risk of model efficacy."
• Philosophy: It shifts the question from "What is the average performance?" to "How low can performance fall when the environment changes?"

B. Mathematical Formulation

The framework involves segmenting a return series $x$ of length $T$ into $s+1$ distinct regimes based on structural breaks.

1. Regime Splitting: The return series is split at time points $t_1, t_2, \dots, t_s$.
2. Constraint: Each regime must have a minimum length $d$ to ensure statistical validity.
3. Optimization: The algorithm searches for the set of splits $\mathcal{T}$ that minimizes the performance metric across the resulting regimes.
   • Let $S(r_i)$ be the Sharpe ratio of regime $i$.
   • For a specific split set $\mathcal{T}$, the minimum performance is $m_s(x, \mathcal{T}) = \min(S(r_1), S(r_2), \dots, S(r_{s+1}))$.
   • The MRP is the global minimum across all valid split configurations:
     $$MRP_s(x) = \min_{\mathcal{T}} m_s(x, \mathcal{T})$$
4. Interpretation: A lower MRP indicates higher fragility; a higher MRP indicates greater robustness. The metric is intentionally conservative (biased downward) to reflect a worst-case robustness view.

C. Empirical Implementation

• Data: The authors applied the framework to a universe of 13 established cross-sectional factor strategies (e.g., Value, Momentum, Quality, Investment) from the Jensen, Kelly, and Pedersen (2023) dataset, covering U.S. equities from 1980–2023.
• Regime Identification: Regimes were segmented based on structural changes in volatility, liquidity, and monetary policy, rather than arbitrary time windows.
• Parameters: Sensitivity analyses were conducted on look-back lengths (10–40 years) and minimum regime duration ($d=1$ to $5$ years).

3. Key Contributions

1. Novel Risk Metric: MRP operationalizes "robustness" into a single, interpretable statistic comparable to the Sharpe ratio.
2. The Decay-Risk Frontier: The paper identifies a trade-off frontier between efficiency (full-sample Sharpe ratio) and durability (MRP). This mirrors the classic mean-variance frontier but focuses on temporal stability.
3. Orthogonal Risk Dimension: The authors demonstrate that MRP is weakly correlated with traditional risk metrics (volatility, max drawdown), proving it captures a distinct dimension of risk: the persistence of the alpha-generating process.
4. Governance Tool: MRP provides a mechanism for continuous monitoring. A declining MRP trajectory serves as an early warning signal of model fatigue before significant losses occur.

4. Key Results

• Substantial Dispersion: There is a significant gap between full-sample Sharpe ratios and MRPs. For many factors, the MRP is substantially lower (often negative) compared to the average Sharpe ratio.
• Fragility of High-Performers: Strategies with high historical Sharpe ratios are not necessarily robust.
  • Example: Debt Issuance (Sharpe 1.14) and Investment (Sharpe 0.45) showed severe decay, with MRPs of -0.05 and -1.01, respectively.
  • Example: Size and Low Leverage also exhibited deep negative MRPs, indicating they fail completely in specific regimes.
• Resilient Factors: Quality was the only factor to maintain a positive MRP (0.06) despite a high full-sample Sharpe (0.65), indicating superior durability. Accruals also showed relative resilience.
• Robustness: The ranking of fragile vs. resilient factors remained consistent across different look-back periods and minimum sample sizes ($d$), confirming the metric is not a statistical artifact.
• The Frontier: Exhibit 4 visualizes the "decay-risk frontier," showing that investors must choose between higher average efficiency and higher durability. Many popular factors lie in the "efficient but fragile" quadrant (high Sharpe, low/negative MRP).

5. Significance and Implications

• Portfolio Construction: MRP allows allocators to optimize for resilience. Portfolios overweighting strategies with high MRPs are likely to experience smaller drawdowns during regime transitions. It suggests a heuristic for weighting strategies based on the ratio of MRP to full-sample Sharpe.
• Risk Management Evolution: The paper argues for expanding the risk taxonomy to include strategy-decay risk. Just as liquidity and counterparty risk are monitored, the validity of the investment process itself must be tracked over time.
• Practical Application:
  • Monitoring: Investment committees can use MRP as a real-time dashboard metric. A falling MRP triggers a review of the strategy's logic.
  • Due Diligence: It helps distinguish between strategies that are "lucky" (performance concentrated in one regime) and those that are truly robust.
• Theoretical Contribution: It bridges the gap between static risk models and the reality of adaptive market dynamics (Lo's Adaptive Markets Hypothesis), acknowledging that models can become obsolete as markets evolve.

In conclusion, Alexander and Fabozzi provide a rigorous framework for measuring the longevity of alpha. By focusing on the weakest link in a strategy's historical performance, MRP transforms the abstract concept of robustness into a manageable, quantitative risk factor essential for the next generation of systematic investing.