Stock portfolio design and backtest overfitting

Bailey, David H. and Borwein, Jonathan M. and de Pardo, Lopez M. (2016) Stock portfolio design and backtest overfitting. (Submitted)

[img] PDF
Download (521Kb)


    In mathematical finance, backtest overfitting connotes the usage of his- torical market data to develop an investment strategy, where too many variations of the strategy are tried, relative to the amount of data avail- able. Backtest overfitting is now thought to be a primary reason why investment models and strategies that look good on paper often disap- point in practice. In this study, we focus on overfitting in the context of designing an investment portfolio or stock fund. We demonstrate a computer program that, given any desired performance profile, designs a portfolio consisting of common securities, such as the constituents of the S&P 500 index, that achieves the desired profile based on in-sample back- test data. Unfortunately, the program also shows that these portfolios typically perform erratically on more recent, out-of-sample data, which is symptomatic of statistical overfitting. Less erratic results can be obtained by restricting the portfolio to positive-weight components, but then the results are quite unlike the target profile on both in-sample and out-of- sample data. One implication of these results is that so-called smart beta funds, which are designed in-sample to deliver a desirable performance profile, are likely to disappoint out-of-sample.

    Item Type: Article
    Subjects: UNSPECIFIED
    Faculty: UNSPECIFIED
    Depositing User: Mrs Naghmana Tehseen
    Date Deposited: 28 Mar 2016 11:19
    Last Modified: 28 Mar 2016 11:19

    Actions (login required)

    View Item