Quantitative Strategy on Vanguard ETFs

Algorithm to manage a simulated $250,000 portfolio of Vanguard ETFs only, in a roughly 50-day competition run by the Fox School of Business (Temple University), with constraints on maximum cash, position size and a minimum number of trades. Preliminary analysis showed that the universe reduces to a few independent bets and that historical momentum has no predictive power: expected returns are therefore estimated with the CAPM, and weights optimize the upper percentile of the end-of-competition return distribution. The script also generates rebalancing orders, with a no-trade band to contain costs. Coached by Vanguard professionals during the competition.

Constrained optimization, parameter estimation, turnover management, automated rebalancing.

Python (pandas, NumPy, SciPy, yfinance), Excel, Bloomberg data (fund flows, Sharpe).

CAPM, beta, covariance matrix, principal component analysis, optimization of a quantile of the return distribution, moving-average trend filter.