A Note on Modelling Dynamics in Happiness Estimations

This short note discusses two alternative ways to model dynamics in happiness regressions. A explained, this may be important when standard fixed effects estimates have serial correlation in the residuals, but is also potentially useful when serial correlation is not a problem for providing new insights in the happiness of economics area. The note discusses modelling dynamics two ways the note discusses are via a lagged dependent variable, and via an AR(1) process. The usefulness and statistical appropriateness of each is discussed with reference to happiness. Finally, a flow chart is provided summarising key decisions regarding the choice regarding, and potential necessity of, modelling dynamics.