Downside-Sensitive Portfolio Optimization and Risk Overlays for Real-Estate Securities

We employ an empirical framework for real-estate securities that incorporates portfolio optimization, return distribution tail diagnostics, risk metrics, modeling of long-range dependence in return volatility, regression against benchmark indices, and option pricing; treating these as necessary layers of a risk management structure that concentrates on downside risk. Optimization compared mean-variance against downside sensitive conditional value at risk. Tail behavior was assessed via skewness, kurtosis and extreme value theory; volatility persistence was examined using ARMA--FIGARCH models. Benchmark dependence was examined via the capital asset pricing model (CAPM) employing endogenous and exogenous market proxies. Insurance instruments via European options were priced using a doubly subordinated normal inverse Gaussian pricing model capable of modeling skewed, heavy-tailed return distributions. Significant findings for the optimized portfolios include: return distributions with losses that are heavier-tailed than gains; a transition in time from moderate to high long-range dependence in conditional volatility; smaller values of CAPM ``alpha'' and ``beta'' for minimum-risk portfolios compared to tangent portfolios; and significant implied volatility values.