A Measure of Comovement for Economic Variables: Theory and Empirics
Summary (1 min read)
Summary
- Wkh phdvxuh lv gh qhg lq wkh iuhtxhqf| grpdlq dqg lv dssursuldwh iru surfhvvhv zklfk duh frvwdwlrqdu|/ srvvleo| diwhu vxlwdeoh wudqv0 irupdwlrqv1.
- Frkhvlrq lv xvhixo wr vwxg| sureohpv ri exvlqhvv f|foh v|qfkurql}dwlrq/ wr lqyhv0 wljdwh vkruw0uxq dqg orqj0uxq g|qdplf surshuwlhv ri pxowlsoh wlph vhulhv/ wr lghqwli| g|qdplf foxvwhuv1.
- Zh xvh vwdwh lqfrph gdwd iru wkh XV dqg JGS gdwd iru Hxurshdq qdwlrqv wr surylgh dq hpslulfdo looxvwudwlrq irfxvhg rq wkh jhrjudsklfdo dvshfwv ri exvlqhvv f|foh xfwxdwlrqv1 MHO Fodvvl fdwlrq=.
Did you find this useful? Give us your feedback
Citations
24 citations
24 citations
Cites background from "A Measure of Comovement for Economi..."
...The codispersion coefficient has also been studied in time series to address how two time sequences change concurrently; it is a geometrically natural comovement coefficient since it compares proportional slopes at matched pairs of points across sequences (Croux, Forni, and Reichlin 2001)....
[...]
24 citations
[...]
23 citations
23 citations
Cites background from "A Measure of Comovement for Economi..."
...Thus, while dynamic correlation cohesion – as proposed by Croux et al. (2001) – can be used to study linear similarities of two processes at each frequency ω, the suggested measure provides information on the relative importance of frequencies with respect to the indicators’ overall fluctuations....
[...]
...For instance, Croux et al. (2001) show that while a white noise process and its one-period lagged value have zero correlation, the co-spectrum indicates perfect positive correlation in the long run (at frequency 0) and perfect negative in the short run (at frequency π), overall, however, integrating to zero. Thus, dynamic correlation makes it possible to gain more insights on the linear relation of two stationary processes than the standard correlation measure. Finally, note that dynamic power cohesion differs from cohesion based on dynamic correlation (see Croux et al. (2001)) by relating the co-spectrum to the overall standard deviations and not the respective auto-spectra at frequency ω. Thus, while dynamic correlation cohesion – as proposed by Croux et al. (2001) – can be used to study linear similarities of two processes at each frequency ω, the suggested measure provides information on the relative importance of frequencies with respect to the indicators’ overall fluctuations....
[...]
...For instance, Croux et al. (2001) show that while a white noise process and its one-period lagged value have zero correlation, the co-spectrum indicates perfect positive correlation in the long run (at frequency 0) and perfect negative in the short run (at frequency π), overall, however, integrating to zero....
[...]
...For instance, Croux et al. (2001) show that while a white noise process and its one-period lagged value have zero correlation, the co-spectrum indicates perfect positive correlation in the long run (at frequency 0) and perfect negative in the short run (at frequency π), overall, however, integrating to zero. Thus, dynamic correlation makes it possible to gain more insights on the linear relation of two stationary processes than the standard correlation measure. Finally, note that dynamic power cohesion differs from cohesion based on dynamic correlation (see Croux et al. (2001)) by relating the co-spectrum to the overall standard deviations and not the respective auto-spectra at frequency ω....
[...]
...Finally, note that dynamic power cohesion differs from cohesion based on dynamic correlation (see Croux et al. (2001)) by relating the co-spectrum to the overall standard deviations and not the respective auto-spectra at frequency ω....
[...]
References
27,170 citations
3,983 citations
"A Measure of Comovement for Economi..." refers background in this paper
...In this category belong the following three concepts: (i) the idea of co-integration (Engle & Granger, 1987): two processes are co-integrated if the spectral density at frequency zero has rank one; (ii) codependence (Gourieroux & Peaucelle, 1992), which refers to linear combinations of correlated…...
[...]
1,128 citations
550 citations
511 citations