Building Macro Surprise Signals for Multi-Asset Timing and Allocation
Summary
This report develops indicators from the gap between actual Chinese economic releases and market forecasts. It groups indicators into growth, liquidity, and inflation categories, assigns expected effects on domestic equities, bonds, and commodities, and estimates release dates conservatively to reduce look-ahead bias. Individual and consecutive-period surprises are screened by whether subsequent asset moves match the hypothesized direction more than 60% of the time over selected trading horizons. The selected signals are combined into asset-level surprise indexes for timing and allocation.
The report describes long-short timing and two allocation approaches: monthly rebalancing and rebalancing when signal direction changes. It reports historical results against benchmarks, including stronger results in some periods and no clear improvement in drawdown control. Frequent turnover is a stated concern for the dynamic model. These are historical backtests using Chinese macro forecasts and representative domestic asset indexes; the reported hit-rate filters, sample selection, assumed release dates, and potential overfitting limit how confidently results transfer to live trading or other markets.
Key ideas
- Macro surprise is defined as the released value minus the market forecast, available only at the assumed release time.
- Growth, liquidity, and inflation surprises are assigned different expected effects on equities, bonds, and commodities.
- Signals are screened by the historical probability of the expected asset move over several forward horizons.
- Single-release and consecutive-release signals are combined into asset-level surprise indexes.
- Monthly and signal-triggered allocation models use the indexes, but the report notes turnover and drawdown limitations.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.