Can Measurement Errors or Seasonal Adjustments Explain the Negative Autocorrelation of Monthly Consumption Changes?

The negative first-lag autocorrelation of monthly consumption changes rejects the continuous-time random walk model of consumption. This paper addresses the question of whether data distortions due to measurement errors or seasonal adjustment procedures may explain this negative autocorrelation, and thus rescue the model from failure. The paper argues that the type of measurement errors that could explain the negative autocorrelation is not plausible, whereas plausible types of errors do not explain it. The paper also shows via Monte Carlo simulations that the application of the X-11 filter may explain the negative autocorrelation or may not, depending on the statistical properties of the unknown seasonally-unadjusted monthly consumption. Overall, the paper offers further arguments against the continuous-time random walk model.