Equity portfolio correlations across the horizons

Among all the common tasks of financial management, especially asset allocation and risk assessment rely on correlations. Both the construction of an optimal portfolio bounded by a set of constraints, and the computation of the respective standard deviation require an estimation of the covariance matrix of all the assets in the portfolio, which usually include hundreds or thousands of assets. Lately, research in empirical finance has focused on predictability of asset returns trying to find a relationship between asset allocation choices and economic variables that can explain, to a certain extent, the mean reverting behaviour of stock returns time series; but there is a general lack of studies investigating the relationship between asset allocation choices and the correlation term structure. This thesis explores the relationship between the term structure of stock return correlations and the asset allocation choices of risk averse investors, by analyzing returns on four different equity portfolios from 1964 to 2020. The first section analyzes a broad set of economic and scientific papers of highprofile researchers and economists, the second section explores the descriptive statistics of the analyzed data and the correlation term structure, and the third section’s aim is to find an optimal Markowitz asset allocation relating such choice to the correlation term structure.