Quant Flashcards
Formula for t-stat
t = [r √(n-2)] / [√(1-r^2)]
Confidence Intervals
Predicted Y +/- (critical t-value)*(standard error)
R^2 =
RSS/SST or (SST - SSE) / SST
RSS/SST =
(SST - SSE) / SST or R^2
SST =
RSS + SSE
MSR =
RSS / k ; (k = # of independent variables)
MSE =
SSE / (n-k-1) ; (k = # of independent variables)
SEE =
√(MSE) = Standard Error of Estimate
F =
MSR/ MSE or (RSS/k) / (SSE/(n-k-1))
Conditional Heteroskedasticity… its effect…
Residual variance related to level of independent variables… Too man Type 1 errors
Type 1 errors
The incorrect rejection of a true null hypothesis (a “false positive”)
Type 2 errors
The failure to reject a false null hypothesis (a “false negative”)
Serial Correlation… its effect…
Residuals are correlated… Type 1 errors (for positive correlation)
Multicollinearity.. its effect…
Two or more independent variable are correlated… Too many Type 2 errors
6 Misspsecifications of regression models
1) Omitting a variable
2) Transforming a variable
3) Incorrectly pooling data
4) Using a lagged dependent variable as an independent variable
5) Forecasting the past
6) Measuring independent variables w/ error
3 necessary conditions for covariance stationarity
1) Constant and finite expected value
2) Constant and finite variance
3) Constant and finite covariance w/ leading or lagged variables
Tests for covariance stationarity
1) Scatter plot
2) AR model and test correlations
3) Dickey-Fuller test (unit root)
If the two series each have a _______, regression results will be consistent, provided that the two series are ______.
Unit Root; Cointegrated
If neither of two time series in a regression analysis has a _____, you can safely use a _______ to test the relationship between them
Unit Root; Linear Regression
If ARCH exists, use ______ or other methods that correct for ______ to correctly estimate the _____ of the parameters in the time series model.
generalized least squares; heteroskedasticity; standard error
