In the research phase of the trading strategy, model scores are converted into portfolios and then examined to assess how these portfolios perform over time. This process is referred to as backtesting a strategy. The backtest should mirror as closely as possible the actual investing environment incorporating both the investment's objectives and the trading environment.
When it comes to mimicking the trading environment in backtests, special attention needs to be given to transaction costs and liquidity considerations. The inclusion of transaction costs is important because they may have a major impact on the total return. Realistic market impact and trading costs estimates affect what securities are chosen during portfolio construction. Liquidity is another attribute that needs to be evaluated. The investable universe of stocks should be limited to stocks where there is enough liquidity to be able to get in and out of positions.
Portfolio managers may use a number of constraints during portfolio construction. Frequently these constraints are derived from the portfolio policy of the firm, risk management policy, or investor objectives. Common constraints include upper and lower bounds for each stock, industry, or risk factor—as well as holding size limits, trading size limits, turnover, and the number of assets long or short.
To ensure the portfolio construction process is robust we use sensitivity analysis to evaluate our results. In sensitivity analysis we vary the ...