Valve Software's popular game Counter-Strike: Source uses prices to balance more and less desirable pieces of equipment purchased by players. Because scarcity in the game is only an illusion, a price adjustment algorithm was implemented to simulate a global weapons market. This simulated market mechanism, its operation, history, and results are examined. In an attempt to better understand what is going on, particularly in regards to how the mechanism produces undesirable results, simple simulations are conducted. Some changes are proposed that may improve the behavior of prices over time.It's currently a working paper. If anyone has an idea for a journal to which it could be submitted, let me know. It's a strange paper; I'm not sure where it would fit in.
Monday, April 27, 2009
Friday, April 24, 2009
Thursday, April 23, 2009
This blog entry is an article written by one of my best students, Brent Butgereit, who wanted to get it into the school paper, but it seems not to have been accepted. It's a clever solution to persistent housing problems on campus. I hope current students will think about this and consider implementing some version of this. If you're interested, contact Brent, or contact me and I'll put you in contact with Brent.
A couple of weeks ago, roughly three-fourths of the current student body went through “Housing Registration” which has all of the anxiety of uncertain living conditions combined with the tedium of waiting in long lines. This system is theoretically designed to maintain ‘fairness’ in room acquisition and yet in practices proves to be inefficient and largely non-negotiable.
Each student is randomly assigned a number within differently ranked class brackets. Upcoming seniors have priority above all other grades in housing registration because they have, in some sense, “earned it” by going to Rhodes longer (upcoming juniors are ranked below them and so on). The random numbers are meant to take the responsibility of ‘who gets to pick what room first’ out of the administration’s hands. Favoritism can be considered less of an influence with merciless Chance at the wheel.
But because Housing Registration is a lottery, it is unable to allocate the rooms to the people who would value them the most. As well, it may prevent the allocation of rooms to people who have no other housing options. The limited number of rooms and obligation for all first-year students to live on campus can result in upperclassmen – who because of their poor lottery numbers are unable to obtain on-campus housing – to have to transfer. Rhodes may be losing excellent students at the roll of a die (albeit a very large one).
At least once a week I have lunch with a table of economists (professors and classmates) to discuss the economics left out of the 101 and 102 models – or the economics of everyday life. The answer to the housing dilemma was simple (but not simplistic): establish a market for housing. One possibility could be that we randomly assign rooms to students and allow them to trade until they settle with a room they like. Students would at least have the opportunity to trade up to a room they value more. And while this could produce a slightly better outcome than the status quo, it faces the same problem of a random ‘number’ generator deciding who is going to stay on campus.
There might be a better way through the use of an auction; we could just bid for rooms. The person with the highest valuation of a room will be the one to get it. Of course, financial means are not equally distributed throughout the Rhodes student body – so bidding with actual dollars will not work (the students with more money have a clear bidding advantage over the students with less money). So what if we started using academic points instead? Simply put: the better you do in your classes (and the higher level the classes are), the more points you get. (For example, an ‘A’ in a 400-level class would give you 440 points while an ‘A’ in a 300-level would give you 340 points.) Students who are more successful in more rigorous classes would be rewarded by being able to use the points they earned to bid for better rooms. Moreover, this would still grant some degree of seniority to upperclassmen (who are more likely to take upper-level classes anyways). Rhodes could effectively establish a market for academic points where students could buy and sell each other’s points or simply use their own in the housing bidding process. Of course, this does not mean students have to sell their points.
Under the point-bidding system, students who are excelling academically are rewarded for their work while more-passive students will have an incentive to work harder. It is possible that the well-to-do students just buy a large number of points from the good-grade students and outbid everyone for the best rooms easily. But bear in mind that the good-grade students are able to decide whether they would rather have a monetary bonus or better housing arrangements – and that because they face this set of trade-offs, there will not be the flow of points from the hands of the many to the hands of a few. And regardless of whether or not the points accrue in the hands of the well-to-do, it will not be easy for any bidders to know the valuation others place on a room until the actual auction. Not knowing how many points it takes to win an auction means that attempts to amass large quantities of points could have the same outcome as if moderate amounts of points are amassed. However, it would be prudent to post the previous year’s bids for each room so that students can have an approximation of its price.
The most effective way to operate this auction would be via computers (leading to what I see as the biggest hurdle in changing systems – designing the software for bidding). But I don’t see this as being too much of a problem. Each student could place a bid for multiple rooms and the highest bidder would get the room. Should one student win multiple rooms, they select the room they want and the other rooms go to the next highest bidder. It should be made clear that this system would not fix the housing shortage; but it would make it less likely that the better students would have to leave because of a poor housing structure.
The current registration system does not work anywhere near as well as it should. And while the system suggested by the Economist’s Table is not perfect, it provides a way to allocate resources more efficiently. When trying to improve policy anywhere, our goal should be to minimize the inefficiencies knowing that we will never be able to completely eliminate them. With the relatively low cost of switching systems, I don’t see any reason why we shouldn’t try this out; the ability to improve efficiency and experiment with policy design should be welcome – especially in a small, liberal arts college where the objective is to expand learning beyond one’s major. I would be more than happy to talk with anyone who knows how to get this proposal implemented.
Thursday, April 02, 2009
This is an expanded version of a short article I was asked to write for the college newspaper.
We are currently in a recession. Why? How do we make it end sooner? The causes are complicated, and there is some disagreement over them. A bubble appeared in housing. A bubble occurs when people wrongly believe that the price of an asset will keep going up, so they buy now in order to sell later and make a profit. Eventually bubbles pop—everyone realizes that housing can’t go up at such a rapid rate forever, so they scramble to sell before everyone else does (According to Larry Lindsey, there are about 3 million surplus empty housing units on the market, for a total of around 18.5 million--and that was as of last June. The numbers have surely climbed since then). As the supply of houses on the market increases, prices drop rapidly. People who own these assets (some of whom did so in lieu of saving) experience a decline in their real wealth. When this happens, they are poorer, so they spend less and save more. In some macroeconomic models, this reduced spending results in recession due to the inability of prices to fully adjust downward in the short run.
When businesses begin to close down on a large scale (for example, many financial businesses closing down, or many homebuilding companies closing down), many people lose their jobs. If wages were fully flexible, employers could just offer these workers a pay cut to stay on, but this doesn’t seem to be the case. As a result, unemployed workers are poorer, so they reduce their consumption, which affects other people who depend on that expenditure for their own livelihood. The overall reduction in confidence, combined with sticky wages, and reduced expenditure, results in a general economic slowdown. It turns out that the biggest reductions come from investment, not from consumption. That is, consumption is somewhat smooth over the business cycle. Investment is very unstable, dropping dramatically during slowdowns, both because lenders are unwilling to lend, and because borrowers are afraid to borrow for expensive investments when times look bad.
Why did the housing bubble happen? Some blame the Community Reinvestment Act, although that is probably not to blame. It has too small an impact, and happened too far back in time. Some blame changes in regulation, or deregulation, although that is also not clear. Mark Jickling has a great list of possible causes and criticisms of those causes here. The sad truth is that we do not really know why bubbles happen at some times, and not at others. That may be the very nature of bubbles—that they are unpredictable and inexplicable.
In addition to the housing bubble collapse, there is a related credit crunch. As a result of some strange financial dealings, many companies bought financial instruments—such as credit default swaps—that turned out not to be a good idea. Again it is unclear why they took off at this time (but not earlier), and it is unclear if regulation could have stopped this from happening (remember regulators aren’t perfect either). So now many companies hold financial assets that are dangerous—toxic assets, they are sometimes called. That is, some companies hold bundles of “securitized mortgages” which may be worthless, because the mortgagees are likely to default. Companies are reluctant to reveal whether or not they are holding such mortgages, because they do not want to frighten their business partners and potential lenders. This has at least partially frozen some credit markets, however—because lenders are worried that they may be lending to people holding toxic assets, they are reluctant to lend to anyone. Why didn’t stockholders restrain companies from engaging in this risky behavior? We don’t know; they seemed to be asleep at the wheel. Why didn’t these assets blow up earlier? We don’t know.
All of this leads me to a general complaint about macroeconomics: We lack data. Yes, we have years of data, but we only observe the world as it existed. We do not get to run back the clock, try something else, and see what would have happened. In microeconomics we often have a wealth of tiny natural experiments, as different counties or states do different things, or as different firms do different things. Not so at the federal level. To put it another way, we’re trying to do statistical work on depressions with only one observation: The Great Depression. It is very difficult to conclude much from this. Perhaps in a few years we will have another depression to look back upon (let us hope not), giving us two data points.
As a result there are many economists providing many different explanations for how we got here, and how to get out of this mess. There is a lack of consensus on basic macroeconomic variables. Which ones are endogenous? Which are exogenous? What direction does causation flow? I do not think I am exaggerating. Read the various reputable economics blogs—Paul Krugman, Brad Delong, James Hamilton, Arnold Kling, Greg Mankiw—and you will come away bewildered. Are there aspects of macroeconomics on which economists agree?
Most economists prefer to rely on the Federal Reserve to stimulate the economy. The Fed’s preferred method for doing this is to increase the money supply in order to drive down interest rates. As the Fed reduces interest rates, some businesses are encouraged to borrow to build factories, invest in new equipment, tools, training, etc. Ideally this increase in investment and employment would restore consumer confidence and get us out of recession. At the moment, however, the Fed has effectively reduced its target interest rate to zero. There is little more it can do on this front. The Fed has some other tools available, and is trying new tools, but it may have reached the limits of its influence.
Because of this, many economists turn to fiscal policy. Fiscal policy consists of changing government spending and taxes to stimulate or slow down the economy. The goal is to spend money in ways that put currently unemployed resources—workers—back to work. It may be necessary to borrow money to fund this sort of stimulus (tax receipts are down in recessions, and government spending goes up), but it might be worth it to get out of the recession. After all, once the economy speeds up again, the borrowing can be paid back with higher taxes.
Now two problems appear. First, in the real world, governments spend money according to who has political influence, and not necessarily according to what will stimulate the economy. So there has been a great deal of effort by lobbyists to get pieces of the stimulus bill. Politicians want to send money back to their home states to get votes to help them get reelected. Spending on many of the projects—such as green energy—will probably not put to work those workers who are currently unemployed. That is, an unemployed Wall Street guy is not going to become an environmental or electrical engineer building a new clean power plant (not fast enough to stimulate the economy, anyway). Other areas of spending, such as road construction and bridge repair, may indeed re-employ currently unemployed workers (so construction workers who are currently idle may go work on building roads and bridges instead). A portion of the money won’t provide a stimulus now because it won’t take effect until next year. Even the money spent now will not have an immediate effect because these things take time. It is prudent to be skeptical regarding the effectiveness of the stimulus program.
The second problem is that resources devoted to uses dictated by this government spending are not available for alternative uses. That is, there is no free lunch. If we spend $2 billion on building a green power plant today, some of that $2 billion represents resources that will not be available in the future. It depends, again, on which of those resources are currently idle and which are not, and on which resources are finite, like concrete (once used, it’s not available for other uses), or not finite, like labor (that is, using it for one thing now doesn’t mean it can’t be used for other things later, and if it’s idle now, it’s probably better to use it than not). In this sense, dollars are really irrelevant; what matters are resources. Government spending either moves money around from one person to another (like social security) or consumes real resources (like building a dam). It may be worth it to build this power plant, or it might not, but if it is not going to put to work currently out-of-work workers, then the criteria we should use for deciding whether or not to expend resources on it is cost-benefit analysis, and it’s not stimulus.
Bob Higgs has argued that the Great Depression’s duration and end were caused by “regime uncertainty”—businesses were unsure about how much Roosevelt would change the rules of the game. This made long-term investments risky. The same may be true today. Businesses may be holding off on investment spending because they are unsure of what the government will be doing. Will the stimulus package help them or their competitors? Will regulation in their industry increase or decrease? Will health care reform reduce their costs or raise them? This can lengthen the adjustment process, as firms wait to find out how this uncertainty is resolved. For this reason it might be best for Congress and the administration to hurry up and do something, however bad, and stick with it in a predictable way.
A further, more subtle point: Arnold Kling points out that the finance industry (and I would suggest the construction industry) got too big during the bubble. It needs to shrink. Trying to reinflate them by bailing out banks and trying to keep them operating, by encouraging people to take on more debt and buy houses, and generally act the way they did three years ago is simply trying to reinflate the bubble. This is a bad idea. It may be the case that the only way to readjust our resources is to suffer a recession for a year or two. We need to recognize that we built too many houses, and spent too much time and effort trading pieces of paper, and stop doing that, not try to do more of it.
Other proposals have included more income tax rebates or cuts in income taxes. Based on recent evidence, these do not seem to work, because workers save most of the money (often by paying off debt), rather than spending it.
A more interesting recent proposal is the elimination of the payroll tax. Payroll taxes are taxes taken out of everyone’s paycheck up to the first $102,000. Income over that amount is not subject to payroll tax. Payroll taxes are essentially a tax on employment; they are a disincentive for workers to work, and a disincentive for employers to hire. Getting rid of the payroll tax would instantly put money in the pockets of both workers and employers, encouraging the former to spend (although they might just save most of it) and, more importantly, encouraging the latter to hire. Several economists, including Greg Mankiw, suggested eventually replacing payroll taxes with a higher gasoline tax, or a carbon tax, in order to reduce pollution. I think this sounds like a much better idea than a messy spending bill, which may do very little. Or maybe it will do a lot. I’m not a macroeconomist, and my opinion on this may not be very valuable. On the other hand, I’m not convinced that the real macroeconomists know, either.