SP500 Logistic Regression (Essay Sample)

Student’s Name
Professor’s Name
Course Number
Date
SP500 Logistic Regression
In this paper we will use logistic regression to predict the direction of the S&P500 index whether it will move up or down based on the cut off probability of 0.5. We will use four independent variables of Lag1, Lag2, Volume and Liquidity.
SPSS statistics software is used for the analysis.
Part 1: Fit Logistic Regression
The table below shows the fitted logistic model with our independent variables and the dependent variable of Direction:
Variables in the Equation


B

S.E.

Wald

df

Sig.

Exp(B)

95% C.I.for EXP(B)








Lower

Upper

Step 1a

Lag1

-.075

.050

2.225

1

.136

.928

.841

1.024


Lag2

-.046

.050

.838

1

.360

.955

.866

1.054


Volume

-.049

.228

.045

1

.831

.953

.609

1.489


Liquidity

.180

.164

1.201

1

.273

1.197

.868

1.653


Constant

.051

.286

.031

1

.860

1.052



a. Variable(s) entered on step 1: Lag1, Lag2, Volume, Liquidity.

From the model’s coefficients, the logistic regression equation is represented as follows:
LogitDirection=0.051-0.075Lag1-0.046Lag2-0.049Volume+0.180Liquidity
Part 2: Odds Ratio
From the table for the logistic regression above, the Odds Ratio for liquidity being present to liquidity not being present is 1.197. This indicates that that if liquidity is present, the direction is 1.197 more likely to go up than if liquidity was not present.
Part 3: Model Prediction
Below is the confusion matrix for the model prediction:
Classification Tablea


Predicted


Direction

Percentage Correct


Down

Up


Observed

Direction

Down

190