... 0.¹

The argument, $t$, must be centered around 0 because Hermite polynomials are orthogonal on the symmetric interval $\left(-\infty,\; \infty\right)$ with respect to the weighting function $w(t) = e^{-t^{2}}$. See Appendix A for details.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

....²

The current model selects a Hermite series of order 6.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...app:hermite.³

I do not wish to imply that Hermite polynomials are necessarily superior to other polynomials such as Legendre or Chebyshev. The use of Legendre polynomials does not significantly alter my results.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... time.⁴

See SSW (1997b, 207) Equation 10.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... inflation.⁵

See King and Watson (1994) footnote 18. I had previously included the NIXON variable that Gordon (1990, 1167) suggests as a supply shock, but it is not significant and is rejected by various model selection criteria--LR test, Mallows' C$_{p}$ and BIC.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.