Term Of The Day – Regression Analysis
A long standing debate has been occurring between myself and actuaries over whether or not that Regression Analysis is the best predictor for setting future reserves. Actuaries will usually say Regression does not give enough credence to the recent development of claims. I say that history repeats itself even if over a long period of time. Most long term business cycles can extend up to 20 years.
Regression (also called the Sum of Least Squares Method) uses a known independent variable such as an employer’s E-Mod to forecast a dependent variable such as future E-mods. I use Regression very often when predicting any insurance variable over the long term. Most spreadsheet software has regression analysis built into it such as Microsoft Excel(R).
If x is the independent variable, y is the dependent variable, m is the slope of the regression, and b is the y intercept – the basic equation is y=mx+b. There are many other types of regression such as stepwise, multiple, etc.
Excel(r) provides a number of regression analysis packages including my favorite the regression graphs for prediction. You can quickly create a great regression analysis chart for presentations.
©J&L Risk Management Inc Copyright Notice
