Abstract:
This paper introduces two co-movement measures based on the Thick Pen Transform into the macroeconomic literature: the Thick Pen Measure of Association (TPMA) as well as Multi-Thickness Thick Pen Measure of Association (MTTPMA). Both measures are non-parametric, time-varying, and flexible. These methods are used to analyse the co-movement of, first, US long- and short-term interest rates, and, second, growth rates of per capita GDP and consumption. As methodological benchmark, this paper also applies the recently pro-posed measure of long-run covariability. The paper finds, first, the co-movement of all series to be stronger the more long-term the components of the time series are. Second, the co-movement of GDP and consumption growth rates is not only generally higher, it also fluctuates considerably less over time than that of the interest rates. Third, the co-movement of the interest rates is sensitive to choosing how long-term the components are. This is attributable to the different extents to which the interest rates exhibit cyclical behaviour. The benchmark method confirms this pattern of the results.