Pareto-Optimal

Research Questions

Last updated: October 23rd, 2008

At some point, it would be great to get around to answering questions, but for now, I'll just ask them. Some of the questions that follow are absurd and poorly thought out. Others are less so. Most probably have answers that I have yet to stumble upon. If you know of any satisfactory answers to these questions, please let me know!

Organizational Economics

[9/29/08] I suspect that the need for coordination is more important to understanding the inner workings of firms than we give it credit for being. (Though there is certainly some new and exciting work being done on understanding the trade-offs between coarse communication and coordination.) However, my suspicion is that strategic communication concerns are not the most important downside to centralization of decisionmaking power. There is a lot of merit to the idea that a manager simply has a finite amount of information processing ability. If we interpret this as imposing a constraint on the Shannon entropy of the manager's decisionmaking function, how far can we get in understanding the trade-off between the information processing limitations of managers on the one hand and the improved ability to coordinate decisions on the other? Where do other economic actors fit into this formulation, and can it help us understand the structure of decisionmaking hierarchies?

[10/4/08] This note takes a first step in answering this question. I argue that in a world with no finite Shannon capacity constraints, it is optimal to allocate all decision rights to the least biased party in an organization. However, by introducing these constraints, this allocation produces higher decisionmaking variance. In short, there is a trade-off between diplomacy and overload.

[4/27/08] What is the effect of social learning within an organization? Can a monitor serve to break "incorrect" informational cascades by providing appropriate incentives after having observed the actions of other agents? This is related to the first [4/27/08] question under the topic of general equilibrium theory.
[4/27/08] Ellison and Holden (2008) consider the idea of bounded communication between a principal and an agent within an organization. They are particularly interested in the second-best breadth of "rules" governing the actions of an agent. Throughout, they abstract from the notion of incentives to gain tractability. Are there any meaningful interactions betwen optimal incentive schemes and bounded communication?
- [4/9/08] I examine this question in this note, written for a course on organizational economics taught by Bob Gibbons and Oliver Hart. In short, I find the trivial result that, when a principal cannot perfectly communicate his desired action to an agent, hence bringing up the possibility of an agent making a "mistake," the optimal linear incentive scheme is muted.
[4/27/08] Can an ex post randomization of sharing rules increase efficiency in a partnership?
- [3/18/08] This is the subject of this note, written for the Gibbons/Hart class on organizational economics. I find that, under some circumstances, ex post randomization can increase effort provision, but I have not yet worked out whether or not the costs in terms of increased risk outweigh the benefits.
[3/6/08] How does product market competition affect the optimal ownership structure of a firm? How about non profit organizations?

[5/13/08] The first of these questions is addressed in this paper, written for the Gibbons/Hart course on organizational economics. I find that, under a particular notion of competition, product market competition can indeed affect the optimal decision rights allocation within a firm.

[2/29/08] Have dynamics been incorporated into the theory of the firm? (Not in the sense of relational theories of the firm.) How significant are future expectations of market conditions for today's firm boundaries? Do different theories of the firm imply different first order conditions? Hence, in a dynamic framework, do they imply that different variables should be included in the Euler equation? If so, can we test the different theories of the firm to determine which is most relevant?
[2/29/08] Does it matter to the rest of the economy which theory of the firm is the "correct one?" Do different theories of the firm imply different information content of prices? Can we embed different theories of the firm into a simple general equilibrium framework to determine this?
[12/16/07] How does the form of the contract offered affect the selection of individuals (based on intrinsic motivation/ability/etc?) into a firm?
[10/13/07] Innovation for the past couple decades has occurred primarily inside vertically integrated research teams. Does vertical integration lead to some sort of shielding effect, protecting these vertically integrated teams from outside competition, and hence leading to less innovation than we would otherwise see under a more competitive structure?

Industrial Organization and Pricing

[3/6/08] The central result of Jonathan Vogel's "Spatial Competition with Heterogeneous Firms" paper is that more productive firms will be more isolated in the product space. Can this be tested?
[3/5/08] The new two-sided markets literature defines a market as two-sided if, in essence, tax (platform fees) incidence matters for total quantity traded. Unsurprisingly, given how new the theory of two-sided markets is, there has been very little empirical work supporting it or estimating such models. One issue is that the standard theory does not separately identify membership fees and per unit fees. By explicitly modeling competition on one side of the market for the customers on the other, can we pin these values down?
[12/12/07] Prior to an item going on sale, do list prices increase to take advantage of the "you save x percent" advertising?
[2006] Do unions compete with one another for membership and for clients? I have difficulty understanding what economic forces keep a monopsony in place. Is it a coordination issue?
[2005] Why did gas prices not increase until after hurricane Katrina despite the fact that it was well-known prior to Katrina's landfall that the oil-refining capabilities of the United States would be damaged? Doesn't economics predict that prices are a function of expectations and therefore, shouldn't prices have increased more smoothly? After all, if storage is possible (and it certainly is!), then today's gasoline is a fairly decent substitute for tomorrow's gasoline.
[2005] How do firms behave in response to serving predictably irrational consumers? The pricing question posed above is potentially one example. (Also see question under decision theory below) Another example involves the promotion dynamics of a firm which hires people who get utility from "feeling accomplished." If we account for this, it might be possible that the Lazear/Rosen theory of tournaments could be more efficient than the piece rate standard of efficiency.
- [2/29/08] Michael Grubb has a nice new paper "Selling to Overconfident Consumers" which seeks to explain the pricing behavior of cell phone companies.
[2004] Why do people tip? There is plenty of justification for the supply side (why would it be in the best interest of a company to offer tipped positions?) of the tipping phenomenon including worker sorting and improved monitoring (that is, the customer is a better monitor than the principal), but the justification for why the customer tips is weak. It suffers greatly from the last period problem. The usual band-aids for the last period problem, unknown end times and reputation effects, can probably solve this theoretically, but I doubt the predictions such a model would give would be supported in the data. Perhaps some sort of peer pressure model could serve a decent explanation. A potential data source would be credit card receipts which give the customer's name, the restaurant name, and the server's name. Those who return often would be expected to give a higher tip. Those who return often and have the same server would be expected to give an even higher tip.
[2004] Independent of tips, is it possible to find a proxy for "quality of service" (e.g. frequency of computer access for a particular table). Is there a drop in "quality of service" when serving a group of eight or more where gratuity is included? Is there a drop in service given to different groups? (i.e. if people get a senior discount, does this decrease the "quality of service"?) Is a higher initial tip an indicator about whether a customer is likely to return?
[2004] Why are items priced $X.99 or $X.95 instead of $(X+1).00? How widespread is this phenomenon? Is it practiced in other countries? Is it practiced in catalogs, online purchases, "wealthy people" stores, "dollar" stores? Is it preserved when prices increase? (Consider online video game worlds, auctions with less strict pricing)

Social Economics

[3/10/08] Bullying can be a problem both in the school yard and in the work place (http://abcnews.go.com/Health/MindMoodNews/story?id=4410909). A large part of the problem is that bullying usually goes unreported. Taking into account the relevant social concerns (among others, reciprocity, peer pressure, and desire to fit in), is there an incentive compatible mechanism for bullied individuals to reveal who is bullying them? More generally, how robust are mechanisms to the introduction of social preferences?
[12/16/08] When a landlord accepts a housing voucher, do the housing prices in the neighborhood decline as a result? (Landlords often turn down prospective tenants out of "fear that a Section 8 tenant will not properly maintain the premises." (Wikipedia) Conversely, others believe that a section 8 tenant is "a perceived higher quality of tenants, since [s/he] can be permanently removed from the Section 8 program if s/he damages the rental unit and/or fails to pay his/her share of the rent." (Wikipedia)) Does crime increase? (Negative peer effects) Are there any good natural experiments related to housing vouchers?
[2005] Many outcomes of economic situations are dependent on the degree of anonymity present. The more anonymous the economic situation, presumably the less likely we are to see people behave altruistically and the more likely we are to see people engage in "shirking" behavior. These claims are likely supported in the experimental economics literature. As a society gets larger, the more likely that the degree of anonymity between any two randomly chosen individuals is higher. (Having never seen a formal definition and not being so bold as to create one, I am purposefully being vague about the definition of "degree of anonymity") From Professor Ostroy's class, we see that, in the limit, it is in the individual's best interest to act as a price taker (and hence not "shirk") by virtue of such behavior being the Nash equilibrium strategy. This "flattening effect of large numbers" is what makes economics the beautiful subject that it is. In the other limiting case in which individual actions have by no means an infinitesimal impact on an economy (e.g. the workings of a family, workplace, etc.), we do not see the crippling presence of artificially created transactions costs that we would expect. That is not to say that we do not see these costs at all but just that they do not destroy a small economy's ability to "work."
[2005] How does the degree of anonymity affect the results of the divide the dollar game? If the players do not see each others' faces or learn their names before playing, do we get different results than if they have a lengthy discussion on their common interests before playing?
[2005] Is there a theorem which shows that, as an economy expands, the degree of anonymity between any two randomly chosen individuals is increasing? (Or at least non-decreasing?) Alternatively, is there anything to support the six degrees of separation hypothesis? If this hypothesis is true, does it have any economic significance? (Or are people who are separated by more than one or two acquaintances effectively strangers?)
[2005] The degree of anonymity present should affect the level of reciprocity (especially with respect to gift-giving) of an economy. Is it possible to model this? Have experiments been conducted showing this result?
[2005] Why are gifts given in-kind? According to standard economic theory, an optimal gift would be monetary (except where search and transactions costs are sufficiently high, as in rare items), but this is clearly not the case. Why do people prefer to receive in-kind gifts of readily available items, especially gift certificates? Especially curious is the existence of American Express gift cards that have a pretty substantial fee.
[2005] People clearly pay to have social ties to other people. This payment often has both monetary components (paying money to throw a party, paying to eat out at a restaurant) and non-monetary components (the time it takes to plan a party). What is interesting here is that the monetary payments go to a third party who derives no other benefit from the relationship. Combining marginal productivity theory and social networking theory, can we derive a system of prices for social ties? Why does it make sense for these payments to go to those not involved in the social tie?

Decision Theory

[2005] What does Arrow's Impossibility theorem have to say about whether or not we can aggregate the preferences of a single individual over time? Why do we ignore this and simply assume that everyone has additively separable intertemporal utility functions? If we temporarily heed Arrow's theorem and give up our ability to aggregate preferences over time, how do we define the concept of optimality? Arrow's theorem allows for a coherent dictatorial decision-making scheme; how, then would we pick which "age" should be granted this decision-making power? I also think that if we allow ourselves to imagine a single person as a composition of several decision-makers in different periods, we will see that the young self imposes externalities on the older self; how do we deal with such externalities? Does the government aggregate preferences by only giving weight to the age-individual (that is, a specific person at a specific age) who will generate the largest tax revenue or vote in a specific way?

Experimental Economics

[2005] Would it be possible to perform economic experiments within a massively multiplayer online role playing game? Benefits of doing so would be: 1) It would be possible to frame experiments in such a way that they are not overtly seen as experiments, 2) It would be easy to longitudinally study a player's actions, 3) There could be ultimate control over almost all factors, and truly controlled experiments could be possible, 4) People take these games very seriously (See Edward Fullbrook's work) and are able to make very educated decisions. The biggest downside would be the potential selection bias: what members of society play these games? (Though there are millions of people across the world that do participate in these games) In addition, it would be difficult to control for the fact that some people sell stuff through online auction sites (eBay) and hence are able to convert outside wealth into inside wealth. This could possibly be detected by examining a series of trades by an individual: those who regularly make trades that are exceedingly in their favor are likely "culprits."
- [4/27/08] On a related note, about two months ago, several executives from the company that created Eve-Online gave a talk at MIT. They apparently have a good amount of rich data from their game, and they are looking for people to analyze it. They are also interested in setting up a central bank and would like some input from economists.

Behavioral Economics

[2006] Do people with economics training do a better job at understanding and learning from the information contained within a price? For example, in a prediction market, do different groups of people differ in how they adjust their beliefs about the probability of events after seeing the current market clearing prices? This could an important consideration for the proposition that policy makers should use information generated by prediction markets to inform policy decisions.
[2005] Why don't people lie more? Economic theory suggests that only in the presence of a well-designed mechanism would individuals necessarily tell the truth. This does not appear to be true. Perhaps there are some sort of implicit social mechanisms which serve to elicit the truth from individuals in many settings.
[2005] It is often claimed that the assumption of rationality is a "good first approximation" to how people actually behave in the sense that individual idiosyncrasies in effect "cancel each other out." To what extent is this true? (i.e. for what classes of idiosyncrasies does this hold?) I would expect that for the preferences suggested by some of the more recent advances in behavioral economics, this property would not hold. (Unless, of course, we assume that each idiosyncratic individual is of zero measure in a world filled with a continuum of rational types.)

General Equilibrium Theory

[9/29/08] Do the new econometric tools of partial identification have anything relevant to say about the existence or robustness of partially revealing rational expectations equilibria?
[4/27/08] What is the relationship between the vector of optimal state-contingent taxes in the social learning framework of Bikhchandani, Hirshleifer, and Welch and the probability of informational cascades on the wrong action? Would a group of agents optimally decide ex ante to set up a set of taxes to help early-movers internalize their informational externalities in order to decrease the probability of "bad" informational cascades?
[4/27/08] Given the price taking assumption, an economy with a price-taking equilibrium is necessarily "replica invariant" in the sense that, by replicating the economy and rearranging the resources among the now-larger economy, no per-capita Pareto improvements can be made. Arguments that the "extent of the market" determines the returns to specialization rely on the idea that, if one were to replicate the economy, per capita gains would emerge. That is, an economy for which there are limits imposed by the "extent of the market" does not exhibit replica invariance. Thus, it appears that the notion of price-taking equilibrium is incompatible with the idea that there exist any further returns to specialization.
[4/27/08] Typical criticisms of the first welfare theorem are that it does not hold when there is asymmetric information, public goods (a special case of which is externalities), or social preferences in the economy. This says nothing about the robustness of a competitive equilibrium to the introduction of small amounts of these concerns. If there is an epsilon amount of asymmetric information (i.e. a set of small measure of individuals has private information), is the outcome of the economy approximately efficient? Similarly with public goods and social preferences.
[2005] In an economy with individuals with interdependent preferences (of the type specified by Gul and Pesendorfer), what modifications would we need to make to the welfare theorems? Could prices exist? What would they mean?

Various Topics

[4/27/08] When individuals care about who else attends their school with them (preferences can others can be vertical, as in the idea that students may want to have higher ability peers, or they can be horizontal, as in the idea that students may have friends who they would like to go to school with), it seems unlikely that the "student optimal stable matching" of the Gale-Shapley matching algorithm is still stable or student optimal. Does the optimal mechanism when we allow for these types of preferences have the property that group assignment is optimal?
[2/29/08] What are transactions costs?
[9/25/08] Do the clever mechanisms of implementation theory, which often rely on arbitrarily large fines and ignore individual preferences for reciprocity, still function properly when behavioral agents attempt to use them? Can we test this experimentally?
- [9/25/08] This is the underlying question of the work in progress by Fehr, Powell, and Wilkening.
[2005] Is it impossible to endogenize priors?