Moving Average MA Time Series Models скачать в хорошем качестве

Moving Average MA Time Series Models

In this lecture, we explain moving average (MA) time-series models from first principles, with a strong focus on intuition, econometric logic, and empirical finance applications. We clarify what MA(q) really means as a moving average of error shocks, not past values, and show how shocks propagate over time and generate short-memory dependence. We apply a moving average (MA) model to month-to-month returns for the overall market using the Finance Research Gate website. The estimated MA(1) coefficient (θ₁) is statistically significant at the 10% level but not at the 5% level. This provides only weak evidence of short-term dependence in market returns. Therefore, while the results do not strongly support violations of market efficiency, they also do not fully rule out mild short-run predictability at the monthly horizon.