7.7 Appendix: TAR modeling

Let the deviations from the LOP for a sector i (i.e., the sectoral real exchange rate for sector i) be qt(i). An econometric model that captures the predictions of a discrete switch in qt(i) is the following symmetric three-regime TAR model:18

7.9 7.9

where Δ denotes the first difference operator; θi is the threshold parameter for sector i; and qtd(i) represents the threshold variable for sector i, with d denoting an integer chosen from the set Ψ ∈ [1, d]. The indicator function, given by l( · ), takes a value of unity when the bracketed expression is true, and is zero otherwise. The error term ε it is assumed independently and identically distributed (iid)Gaussian. The autoregressive order of qt is images. Note that essentially the TAR specified has two regimes given that the behavior of the outer regimes is identical because of the symmetry assumption.

The integer d represents the delay parameter and reflects the possibility that market participants react to deviations of the LOP from equilibrium with a lag. As long as |qtd| ≤ θ, the time series qt follows a unit root process. Thus, in this regime (or region), there is no tendency for the series qt to move back toward equilibrium. Once |qtd| > θ, however, qt becomes a stationary process and has a tendency to ...

Get Handbook of Exchange Rates now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.