There is a growing empirical literature that seeks to determine country growth and demand. One possible shortcoming of those studies is their sensitivity to the estimation strategy, which is linked to the time horizon. A profit-led regime has been found to be the most likely outcome when adopting an aggregative approach, while a wage-led regime is the most likely outcome when utilizing a structural estimation. Another potential weakness is that the empirical approaches adopted are mostly linear. To overcome these criticisms, we adopt wavelet analysis, which allows the decomposition of a time series into short- and long-term components, thereby enabling investigation of whether the growth regime switches over time. This paper tests an extended version of the Goodwin model for the US economy using data from 1967 to 2016. The results show that both the growth and demand regimes are sensitive to time and are profit-led in the short term and wage-led in the long term, thereby confirming Blecker's (2016) insight.