Chapter 9 Regression, Multivariate, and Nonlinear Models Flashcards
Coefficient
The values of alpha and Beta
Independent Variable
One-factor regression (Rmarket-Rriskfree) is the independent, the variable, used to predict expected return, and is frequently labeled X.
Error Term
Error terms (eit) or residuals measure the amount that the regression line fails to exactly match individual data points after choosing alpha and beta.
Dependent Variable
Return of individual securities (Ri) and is frequently called Y, because the regression seeks to predict this value based on another variable.
Intercept
The intercept (alpha) is an estimate of the excess return for the level of riskiness of the stock.
Simple linear regression
Ordinary least squares or simple linear regression analysis provides a way to estimate (alpha) and (beta).
Slope
Measure the amount of undiversifiable risk in the returns of a particular stock.
Non-normality of returns
Outliers occur frequently in return data, where extreme values are observed more frequently (leptokurtosis or fat tails) than would be expected if the errors were normally distributed.
Autocorrelation or serial correlation
Apply the standard Pearson’s correlation coefficient to sequential data.
Present when there is a correlation between error terms and the same error terms lagged.
Durbin-Watson measures autocorrelation.
Heteroskedasticity
Data set that does not have constant variance over the range.
Nonstationary
Characteristic means that the regression parameters are differs for subsets of data.
Beta of a stock differs over time, the regression is non stationary.
Goodness of fit
Rsquared statistic, which measures the extent that the regression matches the dependent variable using the independent variable.
R-squared
Measure of how well a regression results fit the data.
T statistic
equals the regression parameter (alpah or beta) devided by the standard error of that parameter.
