15 seconds to alpha: Higher frequency risk pricing for commercial real estate securities

This is the first paper in the literature to estimate risk decompositions of default, interest rate, liquidity and excess liquidity at intraday frequencies for securitized commercial real estate securities. In intervals of fifteen seconds for 572 days during the Covid pandemic we reveal stark patterns in the price formation of risks for CMBX. We then exploit links between the projected CMBX risk decomposition signals and the related REIT sector in twenty-four long-short daily trading strategies. In Year 1 of the Covid-era, 88% of our risk signalled automated trading strategies produced significant alphas, with 90% of those strategies also generating strong positive abnormal returns. These patterns of revealed mispricing of risk, and trading skill, persist across the entire sample period. In portfolio aggregations across all signals, our eigenvalue weightings perform well in- and out-of-sample compared with standard Markowitz, and equal weighted, portfolio allocations formed from the risk signalled portfolios.