Louise works for BabySoft, a small software company producing computer games for very young children. In designing its software, BabySoft uses a particular kind ofprogram (let's call it D--) that is very idiosyncratic; basically noone else uses it anymore. Unfortunately, much of BabySoft's own internal code is written in D--. Louise does not know D-- and is deciding whether to learn it. To do so would cost her the equivalent of $50,000 in psychic and time costs this year. If Louise learned D--, the expected present value of her lifetime output at BabySoft next year would increase by $80,000. To keep matters simple, assume Louise will only stay at BabySoft for a total of two years (this year and next) and the discount rate is zero. Because BabySoft is a small startup, wage setting is informal. This year's wage is already fixed. At the start of next year, Louise will meet with the CEO to bargain over her wage. The end result is that her wage next year will split the difference between what she is worth to the company and what she can earn elsewhere.

a) In this situation, will Louise decide to learn D--? Why or why not? What if learning D-- raised next year's output by $110,000? Does your answer depend on the sharing rule (i.e. on the expected outcome of her bargaining meeting with the CEO)? Is the company better off if Louise learns D--?

b) Can you think of a better way for BabySoft to run its wage (and/or training) policy than in (a)? (We are looking for a "win-win", or Pareto-improving policy that makes both the company and Louise better off than if the training did not occur). If so, describe it. If possible, de vise a scheme that induces Louise to make the "correct", i.e. the Pareto-optimal, training decision regardless of the parameter values, i.e. regardless of the costs of learning D-- or the actual effects of knowing D-- on her productivity.