Every allocator knows the experience. A hedge fund manager with a stellar track record suddenly falters. Performance dips, peers outperform, and questions surface in the Investment Committee. Do you stay patient and give the manager room to recover, or do you cut losses and reallocate?

It is one of the hardest decisions in investing. Too much patience risks compounding avoidable losses. Too little patience risks exiting just before a rebound. Worse, redemption decisions are often reactive, emotional, or driven by narrative rather than process.

At Resonanz, we believe redemption choices deserve the same discipline as initial allocations. This post introduces a structured framework—grounded in statistical rigor, qualitative signals, and portfolio-level context—to guide allocators through the inevitable moment when a “good” manager stumbles.

Distinguishing Noise from Breakdown

The first task is to determine whether a manager’s underperformance is statistical noise or evidence of a structural breakdown in skill. In our earlier article, Why Track Records Matter and What They Actually Tell You, we argued that a track record is more than vanity—it gives you bounds on plausible future outcomes and helps calibrate expectation ranges. That framework underpins the diagnostics below.

Key question: Is underperformance within normal bounds, or has the investment edge deteriorated?

  • Factor alpha regression with robust errors. Run a multi-factor model (e.g. Fung–Hsieh, equity/credit/vol factors) on the manager’s net returns. Use Newey-West / HAC standard errors to guard against serial correlation and heteroskedasticity. If alpha is significantly negative statistically over your chosen window, that’s a red flag.

  • False discovery control across your manager universe. When evaluating many funds, some will show “alpha” by chance. Use a False Discovery Rate (FDR) filter to classify managers as having likely genuine alpha vs likely luck.

  • Bootstrap / Bayesian inference for short records. For managers with fewer than, say, 5 years of data, bootstrap the return series (block resampling) or apply Bayesian shrinkage to estimate the probability their true alpha > 0.

  • Structural break detection. Use tests like Bai–Perron to detect regime shifts in alpha or factor exposures. A newly broken regime where alpha vanishes or volatility jumps is a warning sign.

rolling_alpha

Source: Resonanz Capital

These tools help you avoid the classic trap: mistaking bad luck or a weak cycle for permanent decline—or conversely, mistaking a true breakdown for mere noise.

Qualitative Diagnostics Worth Watching

Even a manager with statistically sound methods can falter due to non-quantifiable deterioration. Use these qualitative checks as triggers or sanity checks alongside quantitative signals.

  • Thesis integrity: Has the original investment edge changed? If a value manager is forced into momentum trades, that's drift.

  • Team & key personnel risk: Did a star PM leave? Has the analyst pipeline thinned?

  • Style drift / mandate creep: Is the manager quietly stretching limits—adding leverage, relaxing sectors, chasing yield?

  • Operational & governance health: Are risk controls still robust? Are audits, compliance, and reporting intact?

  • Communication tone: Does the manager overly blame market conditions instead of acknowledging mistakes? Transparency is a leading indicator.

If multiple qualitative red flags align with quantitative weakness, consider trimming or exiting.

Portfolio-Level Context: Contribution & Opportunity Cost

Redemption is not just about absolute performance—it’s about relative opportunity within your portfolio. That means assessing contribution to return, diversification value, liquidity implications, and opportunity cost.

In our prior piece, Putting a Price on Flexibility: Reassessing Hedge Fund Lockups, we explored how illiquidity and fund lockups embed a hidden cost in portfolio construction. This insight is especially relevant when judging whether to stay or exit:

  • Marginal contribution to risk & return: Compute how much incremental Sharpe or drawdown protection this manager adds (or detracts) to your portfolio.

  • Alpha vs AUM relationship: As a fund grows large, alpha often decays due to capacity constraints. Monitor historical alpha vs capital.

  • Correlation & crowding: If the manager’s exposures correlate increasingly with your other holdings or crowded trades, its diversification benefit may erode.

  • Drawdown tolerance & liquidity fit: A manager with lockups, illiquidity, or drawdowns that exceed your tolerance should face stricter exit thresholds.

portfolio_contribution

Source: Resonanz Capital

A manager with modest underperformance but strong contribution may deserve more leeway; a manager with similar losses but negative marginal contribution should face quicker review.

A Structured Decision Protocol

To avoid ad hoc decisions, set up a “traffic-light” ruleset in your Investment Committee (IC) charter:

Signal Bucket Quant Trigger Qualitative Triggers Action
Green α t-stat > critical threshold, PSR/DSR significant, no break No red flags in governance or drift Maintain or modestly increase
Amber / Watch α borderline; PSR moderate; weak break test One or two qualitative flags Scale down, monitor closely, require remediation plan
Red / Exit α not significant, statistically negative, break test confirms regime shift Multiple red flags; capacity loss Partial or full redemption, reallocate capital

Key implementation notes:

  • Use rolling windows (e.g. 36–60 months) and overlapping windows to smooth volatile outcomes.

  • For newer managers, set minimum track record thresholds before they move into “Green” status.

  • Document every decision with quant justification and qualitative rationale in your IC materials to maintain governance transparency.

  • Use staged redemptions when possible: reduce size rather than full exit immediately—gives optionality to reverse course if performance recovers.

Applying the Framework: Two Hypothetical Cases

Case A – Quant Equity Manager

  • Over past 24 months: negative net alpha, but historically strong.

  • Structural break test detects a regime shift mid-period.

  • Qualitative review: key quant researcher departed; turnover has spiked.

Decision: Move to Amber / Watch. Partial reduction, require revised research plan, maintain oversight.

Case B – Event-Driven / Distressed Manager

  • Underperformance in recent cycles; alpha statistically insignificant in multiple windows.

  • Correlations to your existing credit sleeve rising; negative marginal contribution.

  • Manager communication evading accountability.

Decision: Exit. The evidence points to sustained decay rather than transient misstep.

These are stylized, but the point is clear: redemption decisions should emerge from consistent signals, not emotion or inertia.

Conclusion

The hardest decision allocators face is not which manager to hire, but when to let go. When good managers stumble, the choice is not about blind loyalty or knee-jerk exits.

Deciding when to stay or go with a manager is seldom binary. The most defensible approach blends statistical inference, qualitative judgment, and portfolio-level evaluation. By building a structured, auditable protocol, you minimize emotional biases and reactive mistakes.

At Resonanz Capital, our philosophy is that risk is deterministic: We cannot predict outcomes, but we can design processes that make our decisions more robust. When managers stumble, having a disciplined framework lets you act with conviction rather than regret.

Subscribe to our Newsletter