Physics 3333 / CFB 3333 Local Optimization


Local Optimization

Local optimization describes a process in which individual entities (people, companies, governments, etc.) attempt to optimize some situation (maximizing profit, reducing costs, maximizing production, etc) in the short term without effective consideration of the global picture.

Such activity can have downsides, or possibilities for losses, but the losses are not thoroughly analyzed or are assumed to fall on someone else. Also - short-term optimization assumes that optimum short-term actions are the same as optimum long-term actions. Furthermore, long-term results are often not even considered, making them a surprise when they finally do occur.

Assumptions

Implicit in this kind of approach to things is the concept of assumptions. A lot of short-term local actions are taken under a set of assumptions about what will happen in the long term. Remember that an assumption is something you take as fact although you have no proof of it.

The problems arising from local optimization result from a few basic assumptions that prove wrong.

  1. That you can really understand and describe the world and its interactions.
  2. That people will continue to behave as they do now, even as things change.
  3. That what is good in the short term is also good in the long term.
  4. If something is possible it is a good idea to do it.
There are other assumptions, of course, but these cover a lot.

Assumption Analysis

Most people are not accustomed to analyzing their assumptions, but doing it could save a lot of misery. Scientists have to detail their assumptions in their scientific papers. Why shouldn't we do something like that?

When planning some action, do the following:

  1. Write down everything you are assuming. This is not easy and will take a good bit of practice.
  2. Look at each assumption and try to figure out what happens to your results if that assumption turns out to be wrong!
  3. If the consequences of a wrong assumption are quite bad, you might benefit from investigating that assumption very carefully.
Failure to analyze assumptions is part of short-term thinking.

What's Going to Happen

It seems that failure to answer this question is part of short-term thinking. What will happen if...? Answering this is hard. It often requires predicting human behavior, which is usually difficult. Also - the person who should be asking the question is often focused on the near term and cannot imagine what might happen in the future or on a larger scale. It's sort of like looking at the world through a paper towel tube; your view is constrained so you don't see everything.

It means, among other things, that you just think you can control the world around you. Any action to accomplish one limited purpose will likely produce effects you didn't expect. Some of these effects can be quite unfavorable, in some cases worse than the original problem.

Remember that, in the context of this discussion, assumption #1 above is almost always wrong.

Static Analysis

The term "static analysis" is often applied to government actions. There is a nice economic definition of it, but it really boils to assumption #2 being wrong. It means that, for example, if you change a tax structure in some area and make your revenue projections assuming that the peoples' behavior pattern won't change, your projections can be wildly wrong. Such analysis does not account for how variables may change.

The Tragedy of the Commons

This concept is traced back to a 19th century English writer on economics, William Forster Lloyd. The theme was later developed in 1968 by ecologist Garrett Hardin. The basic premise is that, given access to a shared depleting resource, individuals will attempt to maximize their own return/wealth in the short term while the long-term result is the destruction of the resource. It is interesting to note that the individual participants may know perfectly well what might happen but go ahead and do it anyway on the grounds that "If I don't do it someone else will." Often a few people reap the benefit of the overuse while every pays for it.

Examples

The Law of Unintended Consequences

These unintended side effects result from the fact that the world, and human behavior, is very complex. It is not possible for any individual to understand all of the possibilities and interactions that can occur. It has been characterized as the result of attempting to regulate a VERY complex system like a society or an economy with a simple system like government. You often get results you didn't anticipate. Rob Norton notes that "Economists and other social scientists have heeded its power for centuries; for just as long, politicians and popular opinion have largely ignored it." This is not good, as the unintended consequences can be quite bad.

When we started researching this we found just how powerful and ever-present the Law of Unintended Consequences really is. It is seen coming into action often enough that one wonders just how deeply policy-makers are thinking about what they are doing. Stephen Dubner and Steven Leavitt, in a 20 January 2008 article in the New York Times (see References), say that "... if there is any law more powerful than the ones constructed in a place like Washington, it is the law of unintended consequences."

Examples

Charles Murray's Laws of Social Programs

For the whole story, see the full column here. Murray is discussing empirical "laws" that affect the results of social programs in spite of the good intentions of those who design them. Assumptions 1 and 2 above prove false.

The Law of imperfect Selection

Any attempt to craft rules for program eligibility will invariably exclude some people who should be eligible. The world is too complex for rules to work perfectly, yet rules are all we can assemble.

The Law of Unintended Rewards

"Any social transfer increases the net value of being in the condition that prompted the transfer." (Charles Murray) For example, a program which gives substantial benefits to homeless people immediately increases the value of being homeless; if you are not homeless you cannot get the benefits

This effect is, as Charles Murray suggests, often seen in social programs. Suppose there is a problem with some social pathology. Your government (or some other body) wants to "solve" the problem. This is often done by offering some benefits to those who have the problem with the intention of helping them improve their condition. The hazard is that, if the benefits are particularly attractive, other people will get themselves into the condition that allows them to get the benefits. You wind up with more of the undesired behavior for the simple reason that you are paying for it.

A good example can be found in programs to assist the homeless. People have been known to move out of a relative's home, live in their car for a while, then apply for housing. The fact that they are "homeless" results in their being kicked to the top of the waiting list, ahead of everybody else who are not "homeless." That is a valuable benefit.

The Law of Net Harm

"The less likely it is that the unwanted behavior will change voluntarily, the more likely it is that a program to induce change will cause net harm." (Charles Murray) It can take a long time for people to understand the damage that is being done and finally stop the program.

References







Outline