- How to Forecast Stock Sector Performance
- How to Calculate Market Price Acceleration
- How to Calculate Supply Demand Indicators for the Stock Market
- Advantages and Disadvantages of a Diversified Portfolio
- How to Calculate Linear Regression Lines and Slopes for Stock Prices
- What Are the Advantages of a Simple Moving Average Over an Exponential Moving Average?
Portfolio sector contribution is a simple concept. It refers to the specific role played by each part of an investor's holdings to the whole. In most portfolios, certain sectors play a greater role than others. Some sectors lose money. Most of the time, the contribution calculation is simple, and can be done just by glancing at the money each section of your holdings is bringing in. For large portfolios, a more complex regression analysis must be done.
Enter the data into SPSS, STATA, or some other regression software. In most cases, it needs to be done manually. In firms and governments, that data is already in electronic format, so it is just a matter of taking each sector of investment — commodities, currency, stocks, bonds — and entering their financial performance over time. The time period is up to you, but the longer the period, the more accurate the results.
Clarify both your independent, control and dependent variables. The dependent variable is your profit over time. The independent variables are the various sectors of your holdings. The control variables are those things that might distort the results, such as inflation, interest rates, investor pessimism, wars, and oil price hikes. These control variables will give a more accurate determination of each sector's contribution because these distorting elements will be removed.
Interpret the R-square. This is a fairly simple number that will give you an indication of how good the model is. A score of 1, is perfection. It means that the data perfectly match your model. That never happens. One means the model is random, meaningless and nonsense. There is no connection between your dependent and independent variables. The R-square will fall somewhere in between. The closer to 1, the better your model. Some experts, such as Neil Henry at Virginia Commonwealth University, hold that anything hovering around .8 is very good. If you have a bad R-square, then the model is useless. This normally happens only when your variables are ill defined, or your control variables are poorly functioning.
Interpret the P score. In this particular model, the P score is the most important element. Each of your independent and control variables will receive one. Assuming your R-squared is good, then the next step is the P score. This is a measure of actual contribution. In mathematical terms, the "P" stands for probability. It refers to the probability that the variable in question, by itself, is totally explaining the dependent variable. In this case, it is a score that estimates how much each sector is contributing to the entire portfolio.
Use your data. If P shows that commodities are doing 40 percent of the work, and currency holdings 5 percent, then either currency holdings should be sold of in favor of more commodities, or different currencies should be bought. If your control variables are doing most of the work — which will rarely be the case — then that suggests that your investments are poorly thought out. You are making money because of accidents like low interest rates, rather than the overall rationality of your portfolio. If there is one sector that is doing nearly all of the work, several things could be happening, such as a recent spike in those values. In this case, this is a misleading reading, and you must enter data over a longer time period to even this out. It can also be that it is the only rational part of your portfolio, and you should reconsider how to invest.