| Description |
This dissertation consists of three independent essays on cyclical fluctuations of functional income distribution and effective demand in the Post-Keynesian tradition. The first essay investigates the longer run relationship between wage share and measurements of economic activity. Our main tools are wavelet covariance and multiresolution analysis. Results indicate that (1) Goodwin type cycles are observed even at longer run and (2) when considering smooth trends for periodicities longer than 32 years, a long Goodwin cycle seems to appear from the 1940s to mid-1990s that collapses afterwards. The second and third essays are related in the sense that they empirically investigate the possibility of strong internal dynamics in the business cycle model of effective demand and income distribution. Specifically, in the second essay, we study wage share and output gap in an univariate setting. Each time series is examined through chaos theory. The main tools are the nonlinear autoregressive neural network model, the dominant Lyapunov exponent, coefficient of determination, and local Lyapunov exponent. Results indicate that output gap might behave quasi-chaotically and wage share noisy-stable. Finally, the third essay inquires into the possibility of limit cycle in the two-dimensional model on wage share and output gap. For that, we use the multivariate nonlinear autoregressive neural network model. Our results indicate that limit cycle behavior describes well their dynamics and, furthermore, the instability is located on the wage share isocline. Chapters 1 through 3 open several questions that we hope further research will address. In Chapter 1, we conjecture that globalization plays a crucial role in the stagnation of the trends in the late 1990s. However, further research is required. Chapter 2 concludes that instability is rooted in the goods market dynamics rather than the distributive dynamics. Results on Chapter 3 indicate that the demand regime is stable and wage share locally unstable. This possibility remains largely unexplored both in the theoretical and empirical literature, and it creates a contradiction with the results found in Chapter 2. Further research is necessary on the robustness of the result, and possible mechanisms. |
