Threshold autoregressive models in which the process is piecewise line
ar in the threshold space have received much attention in recent years
. In this article I use predictive residuals to construct a lest stati
stic for detecting threshold nonlinearity in a vector time series and
propose a procedure for building a multivariate threshold model. The t
hresholds and the model are selected jointly based on the Akaike infor
mation criterion. The finite-sample performance of the proposed test i
s studied by simulation. The modeling procedure is then used to study
arbitrage in security markers and results in a threshold cointegration
between logarithms of future contracts and spot prices of a security
after adjusting for the cost of carrying the contracts. In this partic
ular application. thresholds are determined in part by the transaction
costs. I also apply the proposed procedure to U.S. monthly interest r
ates and two river flow series of Iceland.