3 Project Management You are the computer network administrator of a company that has three divisions. There has been a proposal put forth to upgrade the company's network infrastructure. Such an upgrade would cost a total of $100. Your boss is something of a tight bean counter and insists that every project you engage in is worthwhile. Consequently, you only want to do the upgrade if the benefit to the company outweighs the cost. Problem is, you can not accurately estimate the benefit of the upgrade to any of the divisions. The manager of each division, however, knows exactly what the benefit to his or her division will be. You will need to design a mechanism to try to figure out if you should carry out the project and collect the money to fund it. In answering the questions below, use these two possible environments as test cases. This table gives the actual benefit to each division under two possible states of the world ( These are not intended to be states of the world over which anyone needs to form beliefs. These are just meant to be taken as possible examples.) Division A Division B Example 1 Example 2 35 60 60 15 Division C 25 5 a. Under which possible set of values should the project actually be implemented? b. Someone suggests to you that the "fairest" way to decide on whether or not to implement the project is to ask everyone what they think it is worth (i.e. each division sends you a value S;) and then if you decide to implement it (if SA + SB + Sc > 100), charge each division 1/3 of the cost. Evaluate this mechanism using the two example sets of values. What should the division managers report and will the consequent decision be efficient? What about fair? Will you recover enough revenue to be able to pay for the project? c. If for one of the example sets of values it was found that the mechanism would not lead to an efficient decision (i.e. project is implemented when it should not be or not implemented when it should be), find a percentage cost allocation procedure that would yield the correct 2 decision. Would this procedure work for the other set of values in terms of generating an efficient decision? Will you recover enough revenue to be able to pay for the project? d. Now assume that instead of trying a percentage cost allocation approach you try to implement a Groves-Loeb tax in which each division is charged a tax equal to Ti= { 100 S; - Sk 0 if implemented else What signal should the division managers send now? Will the resulting decision rule of implement if Sa+SB+Sc > 100 result in an efficient decision for either set of sample values? Will you recover enough revenue to be able to pay for the project? e. What does this suggest to you about the advisability of using simple percentage cost allocation rules vs. Groves-Loeb taxes? 3 Project Management You are the computer network administrator of a company that has three divisions. There has been a proposal put forth to upgrade the company's network infrastructure. Such an upgrade would cost a total of $100. Your boss is something of a tight bean counter and insists that every project you engage in is worthwhile. Consequently, you only want to do the upgrade if the benefit to the company outweighs the cost. Problem is, you can not accurately estimate the benefit of the upgrade to any of the divisions. The manager of each division, however, knows exactly what the benefit to his or her division will be. You will need to design a mechanism to try to figure out if you should carry out the project and collect the money to fund it. In answering the questions below, use these two possible environments as test cases. This table gives the actual benefit to each division under two possible states of the world ( These are not intended to be states of the world over which anyone needs to form beliefs. These are just meant to be taken as possible examples.) Division A Division B Example 1 Example 2 35 60 60 15 Division C 25 5 a. Under which possible set of values should the project actually be implemented? b. Someone suggests to you that the "fairest" way to decide on whether or not to implement the project is to ask everyone what they think it is worth (i.e. each division sends you a value S;) and then if you decide to implement it (if SA + SB + Sc > 100), charge each division 1/3 of the cost. Evaluate this mechanism using the two example sets of values. What should the division managers report and will the consequent decision be efficient? What about fair? Will you recover enough revenue to be able to pay for the project? c. If for one of the example sets of values it was found that the mechanism would not lead to an efficient decision (i.e. project is implemented when it should not be or not implemented when it should be), find a percentage cost allocation procedure that would yield the correct 2 decision. Would this procedure work for the other set of values in terms of generating an efficient decision? Will you recover enough revenue to be able to pay for the project? d. Now assume that instead of trying a percentage cost allocation approach you try to implement a Groves-Loeb tax in which each division is charged a tax equal to Ti= { 100 S; - Sk 0 if implemented else What signal should the division managers send now? Will the resulting decision rule of implement if Sa+SB+Sc > 100 result in an efficient decision for either set of sample values? Will you recover enough revenue to be able to pay for the project? e. What does this suggest to you about the advisability of using simple percentage cost allocation rules vs. Groves-Loeb taxes