Interpreting Business Cycles as Generalized Two-Dimensional Loops Using Penalized Splines Regression Techniques

The two-dimensional circular structure model by Kauermann, Teuber, and Flaschel (2011) will be extended to estimate more than two time series simultaneously. It will be assumed that the multivariate time series follow a cycle over the time. However, the radius and the angle are allowed to smoothly change over the time and will be estimated using a Penalized Spline Regression Technique. The model will be put to life using the Leading, Coincident and Lagging Indicators provided by the Conference Board. The model confirms known observations and sheds new light in the business cycle discussion concerning the long-term trend of economic time series, duration and dating of the business cycles, and the measurement and confirmation of leading and lagging characteristics of time series.