Micromotives & Macrobehavior Brief. In Schelling’s Micromotives and Macrobehavior, he discusses the phenomenon known as “critical mass.” An example would be the outdoor, pick-up volleyball game scenario. If a lot of people join, then games may continue regularly until the weather stops people from playing, or, few people join and there is not enough enthusiasm to keep the games going more than a few days. Also consider applause: if enough people clap at the end of a performance, then the whole audience claps; however, if just a few clap and no one else joins in, it will awkwardly stop. “Critical mass,” A.K.A. “critical number,” can be described as something that relies on HOW MANY people are doing something, and not necessarily WHAT they are doing. Alternatively, it can be thought of as the “minimum number” of people necessarily to maintain a given activity (like clapping, jay-walking, etc). This phenomenon may also be subject to the type of people involved, such as age, gender, national origin, etc. Why? There’s a certain amount of comfort that comes from the anonymity and safety in being a part of the crowd rather than standing out. This does not apply to all activities (E.G: even if many people start to get tattoos, not everyone will). “Critical Mass” also involves the idea of “lemons,” (where there is unequal and unknown information between parties involved), and “tipping” (shifts in populations from one to another [people can “tip-in” (move to a new place) or “tip-out” (leave their current place)]. (J Kyo)
We have, throughout the course, been talking about people reacting to their environments. An element of this is what information they have about their environments. Hayek stressed that people tend to have local information only but that that was "good enough" for social order if people had the right rules in their heads for how to deal with local information.
The claim there was that it's not possible to have a formally rational approach to aggregate order because nobody (e.g., a Leviathan) can have enough information and wisdom to make the right decisions. The best we can hope for is for individuals to have the right rules and let them go about the business of following them.
Schelling comes along and says "look, they can follow the right rules and still produce a socially undesirable outcome!" In other words, formal rationality at the individual level can produce substantively irrational results at the aggregate level. Hence the title: MICROmotives and MACRObehaviors.
But back to information.
1. Describe 'the allegory of the dying seminar'
Colleagues agree to meet weekly to talk about their work. At first most folks come, but over time the numbers dwindle until eventually the seminar is cancelled altogether. (155.3)
2. What conclusion about human behavior do observers draw from the dying seminar?
There just wasn't as much interest in having a seminar as had been thought. People didn't really want to spend time at a seminar. It did not provide them with enough benefits to be worth the effort. (154.5)
3. What other examples of this phenomenon does Schelling offer?
Pickup volleyball game, pedestrians at a crosswalk, applauding students, packing up books at end of class, obeying rules like parking and smoking regulations, neighborhood deterioration, racial "flight" (154.7-155.5)
4. What, abstractly, do all these examples have in common?
Individuals are basing their behavior on how many others are behaving in some manner or how intensely others are behaving in some manner. Generically, we call these [[[soc116:critical-mass
5. Critical mass is really about critical number or density. Explain how it works.
The phrase critical masscomes from nuclear physics. Radioactive atoms shed neutrons, periodically shooting one out in a random direction. One of two things can happen when a neutron gets ejected: it can hit another radio-active atom or it can zoom off into oblivion (eventually being absorbed by some non-radioactive atom). If the neutron hits another radioactive atom, it will cause IT to eject some neutrons. And so on. The more atoms that are crowded into a given amount of space (density), the more likely is the second event. There is some density at which the number of neutrons being produced exceeds the number escaping and the process "cascades": more and more and more of the atoms "decay" (that is, emit neutrons) and give off energy in the process. The situation is said to have "gone critical" when it reaches this self-sustaining state.
6. "Though perhaps not in physical or chemical reactions, in social reactions it is typically the case that the 'critical number' for one person differs from another's." (157.3). Explain.
The simplest model/theory is one in which everyone has the same critical point (I'll only go to the protest if 100 people are going), but a better model might be one in which some people will "do it" no matter what and others will only do it if EVERYONE is already doing it.
As a situation evolves, some people will quit when the number drops below their critical value or join in when it rises above their critical value.
7. Schelling notes (156.8) that people may be reacting to actual numbers or to the effect of numbers. He indirectly also suggests that it may not be the actual number but rather the information a person has about the number (that is, the impression or belief). Compare this to Hayek's notion that people only ever have local information.
8. Some people's behavior may not depend on numbers. Explain with examples at both ends of the behavioral spectrum (157.9-158-3).
There may be some who will join no matter what and there may be some who will never join, no matter what. Sometimes, the former group may be enough to "get the ball rolling" — because they will show up, the folks who have low critical number will also show up. Together, these two groups are enough to entice the folks with slightly higher critical numbers. And so on.
9. "The model does not tell us which outcome is preferable" (158.5). Explain.
This situation can produce a macro-outcome that everyone wants, that no one wants, or that some do and some don't want.
10. "And even if one of the outcomes is unanimously chosen, we cannot infer that it is preferred from the fact that it is unanimously chosen" (158.7). Huh?
One might select an outcome on the expectation that others will select that outcome and doing something less preferred but that everyone else is doing may be better than doing what one prefers and being alone in that choice. You might prefer not to have a cell phone or not be on Facebook, but once everyone heads down that path, you might decide to as well. See network effects in Wikipedia.
11. Describe Akerlof's "market for lemons" (159)
12. Two new concepts are added in Akerlof's model: averages as information about the world/behavior and "asymmetric information." Explain.
Rather than estimating or counting how many people are doing something, buyers, knowing that some cars are lemons "average that in" to the price they'd be willing to pay for a car that COULD be a lemon.
Asymmetric information refers to the fact that the seller knows more about the condition of the car than the buyer does.
13. What is tipping and what real world phenomena was it first connected with? (159-160)
The metaphor comes from things like see-saws: you can add weight to the raised end for a while but then you get to a certain level and one more milligram causes the balance to shift and the system completely switches to a different configuration — what had been the high end becomes the low end and vice versa.
The original use was in terms of the racial composition of neighborhoods. A few newcomers of a different race would arrive in a homogeneous neighborhood. A few residents would leave. A few more arrive and a few more leave. Sometimes a point is reached when there is suddenly a mass exodus and the racial composition of the neighborhood shifts in a big way in a short time.
- 0201 Schelling's "dying seminar"
- 0210 Hayek: "Cosmos and Taxis"
- 0213 Hayek: "Cosmos and Taxis"
- 0214 Schelling, Akerlof, Lemons
- 0235 Schelling "Micromotives, Macrobehaviors"
- 0238 Fehr and Gintis
- 0410 Schelling Models
- Exam Markets 2012
- Life Of A Pencil
- Hayek "Cosmos & Taxis" (R014)
- Adam Smith : The Division of Labor (R016)
- Axelrod: "The Evolution of Cooperation" (R017)
- Axelrod "Live and Let Live" (R018)
- Annotated Full Text, Smith: Division of Labour
Chong, Howie. "[Why it makes sense to bike without a helmet"
- A set of observations
- dying seminar
- volleyball game
- end of class
- (155.7) "…people's behavior depends on how many are behaving in a particular way…"
- So, probably "critical number" not "mass" (156.5), but in social world can be number, density, ratio, etc.
- Lots of variables: actual numbers, effect of numbers, people's impression of numbers; contact (rumors and diseases) required or not; reversible or not; discrete or continuous behavior (156-157)
- "critical" level may vary across people 157.4
- Collective irrationality (= by following our individually rational inclinations, we create a situation that almost nobody wants) 158.6)
- LEMONS, TIPPING (158.9)
- SELF-FORMING NEIGHBORHOOD MODEL
- CHAIN REACTION
Thomas C. Schelling Symposium - Micro Motives and Macro Behavior (UC Berkeley 2008)
Modeling Racial Segregation
Rubbernecking, Innovation, and Bringing Down a Governments
EXERCISE: what would happen if you clapped at the start of a class? why not? can I get you to clap? Or, imagine this is us
We call behavior "social" when it takes into account other people. But what do we mean by that? One way we can take into account other people is when we interact with them directly - if we want someone to do us a favor, we ask nicely. Another, perhaps more common, way we take others into account is when we base our decisions to do something or not do something on the basis of how many other people seem to do it.
Examples: crossing the street, bullying, buying a stock, getting a smart phone, joining facebook, voting for Bernie, switching back to Hillary.
Of course, we are always also evaluating the action on its own, but we trust one another as a way of
One of the most important findings in the social sciences is humans' capacity to be individually rational and collectively irrational. Now, by "rational" we mean something simple, but specific: rational behavior is when you take an action that will get you what you want. You want to get into college so you do some SAT prep. You want to be better at your sport so you work out regularly. You want to be able to buy that special something so you save a little each week.
But think about some other examples: you pass the scene of a traffic accident and you are curious so you have a quick look, maybe even just a sidewards glance (especially, perhaps, since you've already been slowed down and you wonder what the fuss is all about). That's rational, individually rational. But let's see what you are a part of collectively. Suppose each driver pauses for the equivalent of 2 seconds.
But now let's imagine I have Waze or Google maps in my car and so I learn about the traffic problem well in advance of hitting it and so I decide on taking the other way to Mills. What happens?
In many situations, each person doing something that seems rational can contribute to an outcome that no one wanted at all:
How do these things work?
Let's imagine we have a population of, say, 30 young people sitting in a room. They are faced with an opportunity to make a choice about some action. And they can see each other. Now some of these folks are pretty much ready to do it but others are holding back, looking around the room to see what others are doing. If it appears lots of people are going to make the move, they are inclined to follow, but if it appears few folks are taking the plunge, then they will refrain too.
To be a little more precise, let's imagine that each person in the crowd has a threshold - a level at which she will get on the bandwagon.
If a person cares about whether other people are doing something as a part of their decision as to whether to do it - order dessert at a group dinner, cross the street even though the light is read, join a standing ovation, come out for the protest march - it is reasonable to assume that they have a threshold, each person has their own definition of how many is enough to make it worth their while to join in the activity.
In any community there will always be a few people who will join in no matter what. They are ready to riot, always in the mood for dessert or available for a protest. But everyone else has a number; if fewer than that are expected, then they won't come.
You've all been given a number, this is your threshold - ask a few what their numbers are and how they would characterize themselves.
Let's record come data about folk's numbers:
Let's draw a histogram of the data of the folks in the room.
Now let's look at this diagram in more detail - how many people, altogether, would come to the protest if they believed 8 people were expected? None at all. Nobody has a threshold that low. What about 14? Two people.
Suppose that by hook or by crook on the first week of September 25 people showed up to the protest and this fact was well publicized along with the news that next week another protest was planned. How many would we expect next week?
We can assume that everyone will base their decision on what happened last week. We look at our charts and we see that there are 10 people who are willing to come if at least 25 show up. So, next week we get just 10.
And the next week? How many people are happy to come out if the attendance is 10? Answer: NONE.
But what if by some miracle, we had had 45 people the first week? We see that 37 people have a threshold less than 45 so we expect them to come next week. And the week after that? 21. And then? 7. And then? 0.
But what if the first week we'd managed to get 49? We look at the chart and see that there are 50 people with a threshold at or below 49 so next week we get 50. And then? Again 50. And so on.
But what if that first week were just 2 more at 52? We have 63 willing to come out at that level. And then? 79 are willing to come out. And then? 93. And then? All 100.
Let's examine what happened. Below a certain level we had a dying protest. At 50 we got stuck. And above a certain level the protest took off.
To see better what is going on, let's look at the cumulative distribution
This curve is sometimes called "the S curve" and it describes a lot of things in the social world.
One of the most important is the rate at which innovations are adopted.
We can understand the dynamics of this process a little better if we add something to our chart: the "45 degree line." This line simply represents the points where the number of folks who are expected is exactly equal to the number who will come at that threshold.
Now we can see that the place where we got stuck is a location where the 45 degree line crosses the S curve. And the other two places we ended up - no one coming at all or everyone coming - are also places where the line and the curve intersect.
Let's tweak our data slightly, shifting some folks into the innovator category - these folks will try it even if no one else is doing it.
Our curve now looks like this:
Work out for yourself what happens if 1 person is expected. Two? Four? Forty?
Next we shift a few more folks - now we have a small number who will not attend no matter what. Work out for yourself what happens when we start with 60.
Things we could do: pay a few folks to come; lower everyone's threshold by X; falsify reports so that people think X more people showed up; lower the thresholds of the early adopters
If we pay a few folks to come, it would make sense that we would pay off the folks at the bottom since they'd be the cheapest to persuade? Or should we go for some hard core folks?
Last problem was 0480 (HELP)
Q170. Suppose you have a population of one hundred persons. It is divided into five categories of willingness to join a protest all of which depend on people's expectations of how many others will appear at the protest. The thresholds range from very low (I'll go if anyone else is going) to the very high (I won't go unless basically everybody else is going).
Assume the population is divided among these categories as follows:
|Challenge of Recruiting||Very Easy||Easy||Average||Hard||Very Hard|
|Number at this threshold||10||20||40||20||10|
a. If news reports suggest that 15 people will show up, how many actually will?
b. If last week saw participation of 41 and this is widely reported so that everyone knows, how many will come out this week? And then next week? And after that?
c. What if 91 came last week?
Q171. A common phrase to describe processes in which people engage in imitative behavior is "bandwagon effect." Explain the appropriateness of this metaphor.
Q172. Consider this data on the thresholds in a population. Draw a frequency histogram and cumulative frequency diagram. If news reports suggest participation will be at 20 people, how many people's threshold is met or exceeded? How about if the number is 70?
Q173. Consider this data on the thresholds in a population. Draw a frequency histogram and cumulative frequency diagram. If news reports suggest participation will be at 20 people, how many people's threshold is met or exceeded? How about if the number is 70?
Q174. Consider this data on the thresholds in a population. Draw a frequency histogram and cumulative frequency diagram. How does this system behave when the expected number is 10? 20? 50? 60? 90?
Q175. Consider this data on the thresholds in a population. Draw a frequency histogram and cumulative frequency diagram. Plot the cumulative distribution on a chart with a 45 degree line.
Q201. (1) Explain what 'the allegory of the dying seminar' was in Thomas Schelling's work. (2) What (incorrect) conclusion about human behavior might observers draw from the dying seminar? (3) What other examples of this phenomenon does Schelling offer? (4) What generic phenomenon are these all examples of?
Q214. Explicate and comment:
Akerlof generalized this model to a number of markets in which there is unequal information on the two sides - insurance companies know less than you do, usually, about whether you are accident prone, or susceptible to hereditary diseases, or are contemplating suicide. Life insurance rates for sixty-five-year-olds must allow for a large fraction who are not long for this world. And those who know they are healthy and have a family history of longevity and are exposed to few risks have to pay the same premium as the poorer risks; life insurance being unattractive [to them] at that price, few of them buys it. The average life expectancy of the customers goes down, the rates go up further, and the bargain now looks poor even to those of normal life expectancy. And so forth.
…. It is akin to, and sometimes coincides with, those situations in which the below average, or the above average, withdraw or won't join, causing some potential market or institution to unravel. Because people vary and because averages matter, there may be no sustainable critical mass; and the unraveling behavior, or initial failure to get the activity going at all, has much the appearance of a critical mass that is almost but not quite achieved (Schelling 1978: 159.5-7).
Q235. Schelling’s piece, “Micromotives, Macrobehaviors,” is included because it demonstrates some specific conditions under which market interactions may lead to coordination but not cooperation. What are his two examples and what are the conditions that can affect whether market interactions lead to cooperation? Explain the role they play, perhaps using our class chairs and offices simulation as a point of reference, in limiting the optimism of Smith and Hayek for markets as a source of social order.
Q410. Show what you know about Schelling's "micromotives macrobehavior" models by explaining this diagram.
Q450. Centola, et al. describe a process whereby people collectively "produce" something that is contrary to their individual beliefs/interests/preferences. Thomas Schelling, similarly, described a process whereby socially irrational results emerged from individually rational action. Identify points of similarity and difference, using it as an opportunity to show what you know about the two thinkers' ideas as well as your ability to compare markets and groups as generators of social order.
Q451. If the phenomenon described by Centola et al. is common, what are the implications for Schelling's critical mass and tipping models?