can verify that with Ht defined according to these rules the demands for equities are passed through and the demand for bills (lending) reflects the logic explained in the two-country case-namely that the demand for borrowing in foreign countries reflects currency hedging and the demand for lending domestically is 1 minus the sum of allocations to domestic equity and foreign lending. It will also be useful to note that the 2n X (In - 1)-dimensional matrix formed by taking the product, H^I'1)', is a constant matrix for all i. We denote this matrix, which we use later in equation (6.22), /. The form of/ for the six-country case is:   1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 -1 0 0 0 0 0 -1 -1 -1 -1 -1 0 -1 0 0 0 0 1 0 0 0 0 0 0 -1 0 0 0 0 1 0 0 0 0 0 0 -1 0 0 0 0 1 0 0 0 0 0 0 -1 0 0 0 0 1 0 0 0 0 0 0 -1 0 0 0 0 1 The next step in developing the general model is to put the relationships among expected excess returns into matrix notation. The expected excess returns for country i have two components corresponding to linear and nonlinear effects, respectively. The first component is the linear transformation of the expected excess return vector of the home country to the perspective of country i. The second component is the addition of a column from the covariance matrix of country i. As was discussed in the two-country example, the covariance component arises from the Siegel's paradox effect of the nonlinearity of the inverse function relating exchange rates. Normalizing on country 1 as the home country, then the covariance component is exactly the n + 1st column of the country i covariance matrix. We can pick off this column by postmultip lying the covariance matrix by the vector of l's and 0's, 1^"+"|, defined earlier. Thus, the formula for the expected excess return vector for country i is given by: H.^ + fl.ZJX-1 (6.20)