Economics 174

Problem Set 4 -- Solutions

1. What are routines, in the context of organizational theory?  What roles do they play within organizations?  Applying evolutionary theories of the firm, what makes firms perform particularly well in circumstances where their competitive environment changes quickly?

Routines are regular and predictable activities within or between firms.  They serve primarily as the organization's memory, but also as its intermediate objective (routines as target) and internal incentive structure (routines as truce).  Firms with routines which are robust across states of the world or which allow the invention or adaptation of new routines tend to perform particularly well in dynamic environments.

2. During the late 1880s, new, more efficient, means of producing (packing) meat were invented. Relatedly, Chandler writes that: "In 1882, Gustavus F. Swift, a Chicago meatpacker...began to build a nationwide distributing organization which owned, besides many [refrigerated railroad] cars, a network of refrigerated warehouses that also served as branch offices for the company's wholesale marketing forces."

From the analysis in class and in the readings regarding asset ownership and organizational change (two different topics), what forces provided Swift the incentive to a) own his own refrigerated railroad cars (as opposed to the railroad owning such cars), b) integrate downstream into the distribution and marketing of his meat? What other investments would you expect that Swift made at approximately the same time and why.

Refrigerated railroad cars are a specific asset -- if the railroad owned them, Swift could hold up the railroad and opportunistically demand lower rates than those agreed upon in the original contract.

From the analysis in Chandler, Swift would integrate into distribution and marketing of his meat in order to capture economies of scale in production. Integration aided this by helping maintain a constant flow of inputs through the production process. Also from Chandler, one would expect that Swift would invest in a more sophisticated organizational structure and hire middle managers to coordinate input and output flows.

3. Woodruff (1999) examines shoe retailing in Mexico.  Non-contractible investments undertaken by retail store managers are important in certain circumstances.  These investments are knowledge investments -- managers who invest in learning about individual clients' tastes can sell more shoes.

Managers may underinvest if separation from their store would lower the value of their knowledge investments.  If they do not own their store, then the manufacturer could threaten to fire them -- and appropriate the value of their knowledge investments.  Fear of appropriation would blunt their incentives to undertake such investments.  If managers do own their store, this is not as much of a threat and investment incentives are stronger.

(Note that an important assumption here is that knowlege investments are specific to the store.  If the knowledge investments were equally valuable, given that the manager works at a different store, then appropriation problems would not appear.)

When managers are not owners, they are not 100% residual claimants.  They therefore have less than full incentives to undertake value-increasing investments.  If the manager of the store is the owner, he or she has full incentives to make value-increasing investments. There are several ways of answering this question -- economists would generally disagree on this.  (This is a really difficult and subtle question at some level.  I am struggling with it, in fact.)

On one hand, both theories assume that investment levels are non-contractible.  This is a similarity.  If they were contractible, there would not be an underinvestment problem in either theory.  In both theories, non-contractibility means that managers who are not owners are unable to fully appropriate the value of their investment.

My own jbeliefs about the core of the difference is the following.  In some agency theories, contracts cover all states of the world in the sense that the contract specifies transfers between the prinicpal and agent for all values of the contractible variable (the contractible variable is often "output").  In contrast, there are some states of the world that are "uncovered" by contracts in the property right model.  So the root of the problem in agency models is that one cannot distinguish between all states of the world finely (for example, we cannot distinguish between high output caused by high effort and luck).  The root of the problem in property rights model is that some states of the world are uncovered by the contract.

4. Salespeople at retail apparel stores generally have several responsibilities.  One is that they are responsible for serving customers and encouraging them to purchase goods.  Another is that they are responsible for keeping the store looking good.  Part of this involves restocking clothes and making sure they are displayed neatly, either on hangers or folded on a shelf.  It generally takes little effort to make clothes on hangers look neat; it takes much more effort to make stacks of folded clothes look neat.

Assume the following.  Managers of apparel stores care both that individual salespeople sell clothes and undertake effort toward ensuring that their stores look neat.  It is relatively easy for managers to keep track of the sales generated by individual salespeople.  Because of this, it is feasible to reward salespeople with commissions. It is extremely difficult for them to develop measures of how neat displays look.  It is so difficult, in fact, that it is impossible for them to directly reward salespeople on the basis of how well they maintain displays.  While payment on commission indirectly provides salespeople incentives toward maintaining displays -- they sell more if their store looks nicer -- these indirect incentives are extremely weak, weak enough to be ignored completely.  Any effort expended by salespeople toward maintaining displays primarily comes from their pride from working at a nice-looking store.

Hidayatallah (1997) surveys 37 retail apparel stores in the South Coast Mall, and finds the following.  First, slightly over half of the stores pay their salespeople commissions.  Second, there is a negative correlation between paying salespeople commissions and the fraction of clothes in the store that were displayed folded.  That is, stores which displayed a high fraction of their clothes folded tended not to pay their salespeople commissions.  In contrast, those which displayed most of their clothes on hangers tended to pay them commissions.

Applying Holmstrom and Milgrom's theory, explain Hidayatallah's second finding.

 Applying H&M directly, if it is impossible to directly reward display maintenance, the only way to encourage display maintenance would be to provide weaker incentives for everything else.  Therefore, if effort toward display maintenance is more important at stores with folded clothes than at those with hung clothes, one would expect lower sales commissions at stores with folded clothes.

At some retail stores, "team-selling" is used.  For example, different employees greet, serve, and ring out customers.  Would you expect commissions to be more common at stores which "team-sell" than those at which single salespeople greet, serve, and ring out customers?  Why or why not?

Commissions are not going to be particularly effective in stores with "team selling" because of free-rider effects.  Individuals' compensation are going to be based a lot on factors outside of their control -- such as other team members' effort levels.   Commissions may be more effective at stores where single salespeople greet, serve and ring out customers -- their sales are going to be a function of how they themselves treat customers.  Factors outside of their control have less of an impact than where team selling is used.  One would predict commissions to be more common at stores where team selling is not used.