Page 1 of 1

question about time-varying trigonometric cycle

Posted: Sun Feb 22, 2026 8:59 am
by hardmann
Dear Tom:

I am familiar with using the UC model to decompose trends and cycles. But the cyclical component uses ar2 instead of trigonometric functions. If there are multiple different cyclic components that need to be stacked, is it necessary to use cyclical components in the form of time-varying trigonometric functions with different frequencies alike (Harvey & Jaeger, 1993). "Detrending, stylized facts and the business cycle". Do you have any relevant example?

Best Regard
Hardmann

Re: question about time-varying trigonometric cycle

Posted: Mon Feb 23, 2026 5:35 pm
by TomDoan
Have you looked at the harveyp090.rpf example? That does the cycle as a separate trigonometric cycle rather than an AR(2) with complex roots.

Re: question about time-varying trigonometric cycle

Posted: Tue Feb 24, 2026 6:48 pm
by hardmann
Thanks