Evaluating Model Autocorrelation with Durbin-Watson

Very Hard Time-series Analysis Model Evaluation

A financial analyst is examining a time-series model that forecasts the daily returns of a stock. The model, which includes several lagged variables, has produced the following estimated regression equation:

Returns_t = 0.02 + 0.6 * Returns_t-1 - 0.1 * Returns_t-2 + e_t

Where e_t is the error term. The analyst conducts various tests to evaluate the model's adequacy, including the Durbin-Watson statistic for autocorrelation of residuals. After running the regression, the Durbin-Watson statistic is found to be 1.84.

Based on the typical threshold for the Durbin-Watson statistic (DW), which ranges from 0 to 4, where a DW of 2 suggests no autocorrelation, describe the implication of this statistic in the context of model evaluation.

The model shows strong evidence of positive autocorrelation among residuals.

The model does not show strong evidence of autocorrelation, indicating a reliable model.

The model exhibits negative autocorrelation, which suggests instability in the time-series.

Hint

Submitted5.3K

Correct4.4K

% Correct82%