Which of the following statements is true about autoregressive models of order one, AR(1)?

An autoregressive model of order one, commonly referred to as an AR(1) model, is characterized by its dependence on its immediate past value. When considering the characteristics of AR(1) models, it can be beneficial to analyze the individual statements to identify which one is true.

The first statement describes an AR(1) model as a "meandering process." While it does exhibit characteristics where values can drift over time, a proper description of an AR(1) process requires understanding the conditions of stationarity and the underlying data generation process, which does not inherently imply a meandering characteristic.

The second statement notes that a stationary AR(1) model generalizes white noise and random walk models. In actuality, while white noise has zero autocorrelation and a random walk is a non-stationary process, an AR(1) model does not generalize these models. Instead, it includes an autocorrelation structure that does not align with the characteristics of white noise or a random walk, particularly regarding their statistical properties.

The third statement asserts that the lag k autocorrelation of a stationary AR(1) model is always non-negative. In fact, the autocorrelation function of a stationary AR(1) model decays exponentially and can take on