🔑 Hidden Causal Lever
Target Date Funds Rewired Equity Autocorrelation Post-2010
US equities exhibited trend-following (positive autocorrelation) pre-2010. Post-2010, they shifted toward mean-reversion (negative autocorrelation). The causal mechanism is target date fund growth: from under $10B (early 2000s) to ~$3T by 2020. These funds systematically sell equities after rallies and buy after drops, creating a structural mean-reverting counterforce. Any regime model calibrated on pre-2010 data will systematically over-allocate to trend signals that now destroy alpha.
What most people do
Calibrate regime models on the full available history (including pre-2010 data) and treat trend-following as a reliable equity signal.
What the best do
Explicitly test for autocorrelation regime shifts and recalibrate the model to the post-2010 structural period, recognizing that $3T of systematically mean-reverting capital has permanently altered equity autocorrelation.
Why it's an edge: Explains decade-long underperformance of equity-based trend systems without invoking "the strategy is broken" — it was broken by a specific, identifiable structural change.
How to exploit: Measure rolling 6-month autocorrelation of equity index returns in your strategy universe. If autocorrelation has been consistently negative since 2010, replace equity trend filters with vol-targeting or carry mechanisms. Trend filters still apply to commodities, rates, and currencies where TDF money doesn't flow.
Corey Hoffstein, "Liquidity Cascades," Investment Magazine, 2021-07-10
⚡ Conventional Wisdom Is Wrong
Regime Classification Without A Decision Purpose Is Useless
Most regime classification systems are built as intellectual exercises — identifying which regime exists without defining what the strategy does differently in each state. Without explicit, pre-defined strategy changes tied to each regime label, a regime classifier has zero operational value. The regime model only earns its complexity when it changes position sizing, strategy allocation, or exposure in a measurable, predefined way.
What most people do
Build multi-signal regime classifiers as a research project. Run the classifier in parallel with an unmodified strategy, watching for "confirmation." Never formalize the link between classification and action.
What the best do
Define the decision first: "In a risk-off cascade regime, equity exposure drops to 20%. In a trending regime, trend allocation doubles. In ambiguous, everything halves." Then build the classifier to detect those states.
Why it's an edge: Forces regime research to be decision-linked rather than academically interesting. Regime systems that change behavior generate P&L; regime systems that generate dashboards generate nothing.
How to exploit: Before building or purchasing any regime classifier, write the complete decision table: for each regime state, what specific, numerical change is made to strategy parameters? If you cannot fill in the table, stop building the classifier.
Rodrigo Gordillo & Corey Hoffstein, Return Stacking podcast, 2021-11-15
🔑 Hidden Causal Lever
Endogenous Cascade Risk Is Invisible To Macro Regime Models
The worst market events (March 2020, August 2015, February 2018, August 2007) are triggered by endogenous systematic participant de-leveraging, not by macro deterioration. A macro regime model (growth/inflation quadrant, moving average, macro factor) is structurally blind to this mechanism because it models the world, not the market's internal plumbing. The signal that actually matters — simultaneous maximum leverage across CTAs, target-vol funds, and risk-parity — requires a positioning model, not a macro model.
What most people do
Build regime classifiers from price signals and macro data. When these classifiers fail during the worst events, conclude that better macro signals are needed.
What the best do
Add a separate positioning-based overlay that monitors estimated leverage of systematic participants. When all are at maximum simultaneously, reduce exposure pre-emptively — independent of the macro model's output.
Why it's an edge: Correctly identifies the true early-warning signal for the events that cause the most damage. Macro models will always lag on endogenous cascades; positioning models can be genuinely leading.
How to exploit: Monitor CTA net-exposure proxies (e.g., published by major futures brokers) and vol-targeting fund leverage estimates (implied from VIX and systematic fund AUM). Build a composite "systematic participant leverage" indicator. When composite hits >85th percentile of historical values, reduce net equity exposure by 30-50% regardless of macro regime.
Corey Hoffstein, "Liquidity Cascades," Investment Magazine, 2021-07-10