The latest in a long line of evergreen critiques of the hedge fund industry is that "alternative" strategies are starting to look suspiciously like plain-vanilla equities. The data certainly supports the claim: correlations have crept steadily higher over the past few years, with many active strategies now moving in lockstep with the broad market.
But pointing out that correlation is rising is only the first step of a larger discussion. The more pressing issue is that we are still trying to measure this shift using a metric that simply wasn't designed for this level of market complexity.
Most institutional discussions live and die by a single data point: the five-year monthly correlation to the S&P 500. It’s a clean, spreadsheet-friendly figure. It’s also incredibly blunt.
The problem with a long-window average is that it treats a decade of "Goldilocks" growth and a week of sheer market panic as equivalent data points. It smooths over the exact moments when you actually need a strategy to be uncorrelated.
A strategy can carry a modest 0.3 correlation for years, yet behave exactly like a levered index the second volatility spikes. If a fund's equity-like behavior is episodic, aka flaring up only during stress, a long-term will actively lie to you.
Correlation isn't a fixed property of a fund, it’s a function of market regime. Instead of asking "what is the correlation?", we could also be asking:
- What is the correlation when the S&P is down 5%?
- How does it move during a liquidity squeeze?
- What happens when the VIX breaks 30?
Down-market correlation (or "tail correlation") is far more informative than a full-period average. Many strategies are uncorrelated in calm seas but become highly synchronized the moment the exits get crowded. That asymmetry is what breaks portfolios. If you’re paying for diversification, you need to measure it in the environments where it’s actually supposed to show up.
When hedge funds start "looking like equities," it’s often because they are fueled by the same macro tailwinds: persistent momentum, growth leadership, cheap liquidity.
We need to distinguish between structural exposure and regime-driven convergence.
Structural: The manager is intentionally running high beta or factor tilts (like Mega-cap Tech) to chase returns.
Regime-driven: The manager’s process is sound, but the entire market is being pulled by a single gravity well (e.g., a sudden spike in real yields).
One is a design choice; the other is a market reality. Attribution analysis, including looking at factor-level overlaps, tells you which one you're dealing with. A single correlation coefficient cannot.
Rising correlations aren't a moral failure of the hedge fund industry. In a market where capital crowds into whatever is working, overlap is inevitable.
The mistake isn't that correlations are rising, it’s more-so the assumption that a backward-looking, smoothed-out number can predict how a portfolio will hold up tomorrow. Correlation analysis should be a diagnostic process, not just a line item on a quarterly report.
