cab691
02-25-2012, 06:26 PM
Hi, I have a pretty basic question about the parallelogram method.
Let's say there's a single rate change of 8% at 7/1, and we want to put the earned calendar year premium on level. Naively, I'd guess that if seven eighths of the premium comes from policies written before 7/1, we'd use an overall rate factor of
\frac{7}{8}1.08 + \frac{1}{8},
the weighted average. But that's not what the parallelogram method prescribes! Instead, we use the harmonic weighted average:
\frac{1.08}{\frac{7}{8} + \frac{1}{8}1.08}.
Why is the harmonic weighted average better than the ordinary weighted average here?
Let's say there's a single rate change of 8% at 7/1, and we want to put the earned calendar year premium on level. Naively, I'd guess that if seven eighths of the premium comes from policies written before 7/1, we'd use an overall rate factor of
\frac{7}{8}1.08 + \frac{1}{8},
the weighted average. But that's not what the parallelogram method prescribes! Instead, we use the harmonic weighted average:
\frac{1.08}{\frac{7}{8} + \frac{1}{8}1.08}.
Why is the harmonic weighted average better than the ordinary weighted average here?