Solve the following problem:
Given that IXi≤t is a Bernoulli random variable equal to 1 with probability Φ(t), show that the variance of the normalized estimator IXi≤t/Φ(t) goes to infinity when t decreases to -∞. Deduce the number of simulations (as a function of t) that are necessary to achieve a variance less than 10-8.
Make sure you use enough details to support your answer.