Factor correlations, contributions over time, volatility, and market cycles.
The baseline market return captures the quality-adjusted price appreciation unexplained by tracked factors. It represents the pure market movement - demand pressures, macroeconomic conditions, and anything not captured by HDB flat characteristics.
Low off-diagonals confirm the factors capture distinct dimensions of price variation.
Strong positive correlations indicate factors that tend to move together; negative correlations indicate opposing movements. Well-selected factors should show low mutual correlation - validating that each adds independent information to the model.
How much of the log price each factor explains in each window.
This chart shows how the relative explanatory power of each factor has shifted over time - which characteristics became more or less important in pricing the market.
Higher volatility means the market is repricing that characteristic more erratically - often a sign of thin trading or structural change. Stable factors provide more reliable hedging.
Which factors drove returns during boom, bust, and recovery? This chart stacks the cumulative factor contributions, showing the relative importance of each characteristic across different market regimes.
Rolling 3-month window used in the regression. The model requires at least 50 transactions per window; thinner periods produce wider beta confidence regions.
The model is only as good as the underlying data. Thin markets (few transactions) typically produce less stable betas and are flagged with wider confidence regions on the factor charts.