LASSO and Quadratic Optimization for Portfolio Return Replication

Summary

This research summary presents a framework for replicating the returns of an index, fund, or individual stock with a selected set of holdings. It first applies LASSO regression to choose securities that capture the target portfolio’s characteristics, then uses a quadratic optimization model to determine their weights. Potential applications include tracking a broad index with a small portfolio and infrequent rebalancing, approximating a fund by holding its underlying assets, and substituting for securities that cannot be held.

The reported backtests use portfolios of at most 30 holdings with monthly rebalancing. The summary reports monthly average tracking errors for index, fund, and stock replication, along with the standard deviations of those errors, and says long-run tracking was relatively stable while short-term tracking remained an area for improvement. It also claims the LASSO selections captured portfolio style and industry features and identified fund holdings. These findings are reported in an abstract rather than detailed methodology; the excerpt does not provide sample periods, costs, constraints, or enough evidence to assess out-of-sample robustness.

Key ideas

• LASSO regression selects securities for a portfolio intended to replicate a target’s returns.
• A quadratic optimization step assigns weights to the selected securities.
• The described applications include index tracking, fund replication, and replacement holdings when direct ownership is restricted.
• The reported backtests use no more than 30 holdings and monthly rebalancing.
• The summary reports stable long-term tracking but notes that short-term tracking can improve.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.