Timescale Methods in Economics: Wavelet Analysis of Business Cycle Fluctuations

Business cycles are defined as fluctuations of economic activity between 2 and 8 years. Since market economies are complex and evolving systems subject to internal changes and to exogenous shocks, the statistical properties of economic fluctuations are likely to change over time. Frequency-based phenomena with time-varying features are better studied using time-frequency methods. Wavelet methods are preferable to Fourier analysis because of their optimal time-frequency resolution properties and ability to address the nonstationary features of economic fluctuations. The usefulness of wavelet techniques for business cycle analysis is illustrated by the application of discrete and continuous wavelet tools to certain aspects of the business cycle: output volatility moderation and the cyclical behavior of monetary aggregates.