## Numerical Examples

The results are as follows:

The ex ante expected prices at ¿=1,2 are the same:

The market impact at time 2 is smaller than the one at t=1: P (°F + ^P; ) -> P^'F •

Therefore, it is best for the nations that want to minimize the expected costs to trade only at t=2. The trading partner of the small traders is the noise trader at t=1.

The price at time 1 has an impact on the price at time 2. This is because the private brokers are risk averse and hate buildup of inventory. In the case where there is heavy selling at time 1, the price at time 1 becomes low and small traders are long inventory by a large margin. Then they will prefer to sell in order to avoid the inventory risk, and place lower price at time 2.

25.5.2 N=2, lx=h=p=CF=onl 2=1, anl 2=6, W,i= Wja=W The results are as follows:

The price at time 1 has an impact on the price at time 2. The market impact at time 2 is smaller than the one at t=1, as in the first example. The nations dare to sell at time 1, drive down the price, and make the price at time 2. The ex ante expectation and the variance of the cost of a nation on the equilibrium are WF+(5/2) W and ((15)/4) W2, respectively. If the nations do not trade at time 1, fixing the behaviors of the small traders, the ex ante expectation and the variance of the cost are WF+2W2 and 7W2, respectively. That is, the expected cost of the nations become larger and the variance smaller, compared to the case of risk neutrality.