Institutional Flow Analyzer

It varies by source: ETF flows are observable daily (end-of-day), ICI publishes domestic fund flows weekly, the Treasury publishes monthly TIC foreign-flow data, and the SEC publishes quarterly 13F holdings. There is no single daily foreign-vs-domestic print as in some markets.

Flow data is one useful input but shouldn't be the sole basis for trading. Much of it is lagged and backward-looking and has limitations. Use it together with technical analysis, fundamentals, and other indicators.

Domestic funds (especially via 401(k)/index/target-date plans) have steady recurring inflows that must be deployed. They also have longer horizons and may treat foreign selling as a buying opportunity. Their mandate differs from foreign investors.

Generally, net flow above $1-2 billion is considered significant. Extreme flows would be above $3-5 billion. Significance also depends on market context and whether the pattern is sustained.

Daily flows have short-term impact. Sustained flows (5+ days same direction) have medium-term impact. Cumulative flows over weeks/months show longer-term trends. The impact depends on magnitude and persistence.

Futures positioning (e.g., CFTC COT) includes hedging and speculation, not just directional views. Look at OI changes with price (long/short buildup). Cash/ETF flows are purer investment flow. Combine both for a complete picture.

Track: Fed policy and yields, the U.S. Dollar Index (DXY), VIX for risk sentiment, global fund-flow data (ICI/EPFR), and global market direction. These influence foreign demand for U.S. assets.

Track month-over-month changes in institutional sector allocation (sector-ETF flows and 13F). Overweight sectors where allocation is increasing. This helps identify sector rotation before it's obvious in prices.

Extreme selling may signal opportunity when: it reaches historical extremes (2+ std dev), steady domestic inflows are absorbing it, technical support is holding, and there's no fundamental reason for continued selling. Still wait for some stabilization.

Correlation is significant but varies by market regime. In trending markets, flows often follow prices (momentum). At extremes, flows may lead reversals. Use correlation as one input, not a guaranteed predictor.

Construct a normalized flow signal (z-score over a rolling window), apply smoothing (MA), define thresholds for bullish/bearish, backtest using walk-forward methodology, and combine with other factors. Monitor signal decay and regime dependency.

Combine multiple proxies: U.S. Dollar Index moves, order-flow imbalance at institutional sizes, ETF creation/redemption, ADR premiums (foreign ADRs), and overnight ES/NQ futures direction. Validate against end-of-day data to calibrate. Accept estimation error.

Use Granger causality tests or VAR models. Test the flow->return direction. Account for regime dependency. Be aware the relationship can be bidirectional and may change over time. Walk-forward validation is essential.

Calculate the U.S. beta to global flows, track relative allocation, and measure U.S.-specific alpha. Use global fund flows as a leading indicator. Adjust exposure based on global sentiment and the relative attractiveness of U.S. assets.

Limitations: much data is lagged (EOD/weekly/quarterly), the relationship varies by regime, flows can persist longer than expected, other factors matter, institutional flows aren't always 'smart money', and intraday estimation has error. Use as one component of a multi-factor approach.