Skip to content

The Tipping Point: Fascinating but Mythological?

The “tipping point” is a popular concept covering a whole range of phenomena (and a best-selling book by Malcolm Gladwell) where individual behavior depends on the behavior of the herd.

Its original application was to racial segregation. Nobel Laureate Thomas Schelling developed a beautifully simple model for this. Suppose that whites have different degrees of racism – some would “tolerate” higher shares of nonwhites than others. Schelling showed that the less racist whites would still wind up exiting during tipping because of a chain reaction. At first only the most extreme racist whites exit. But their departure causes the white share to go down, making the second most extreme racist whites uncomfortable, so they also exit. The white share goes down some more, and so now even less racist whites will be uncomfortable being a white minority, and they will wind up exiting too. So the remarkable prediction of the tipping point model is that just a little bit of integration that directly bothered only the most racist whites wound up causing ALL of the whites to exit. So even if the typical white was perfectly happy with integrated neighborhoods, these neighborhoods would be so unstable that the final outcome would be extreme racial segregation. The segregated nonwhite neighborhood will remain permanently nonwhite. Segregated white neighborhoods (with a white share ABOVE the tipping point) will also be stable, because virtually all whites will tolerate a very small nonwhite share. So segregation happens through a chain reaction even though the average white did not want such extreme segregation.

It’s easy to imagine development applications for the tipping point idea. Suppose that people decide to get highly educated based on what is the share of highly educated people in the population. After all, it’s only worthwhile being educated if you can talk to and work with a lot of other highly educated people. If the share of educated people falls below a tipping point, a lot of people will stop getting highly educated, which decreases even further the incentive to get highly educated, and we get the same kind of chain reaction that happened in segregation. So a whole society can tip from high education to low education, below a certain “tipping point” of the share of the highly educated in the population. Assuming that low education causes poverty, this is a “poverty trap” story of low education and underdevelopment.

The tipping point stories are fascinating, but do we observe them in the real world? I got intrigued with this question a while ago, and eventually published a paper testing the predictions of the tipping point story (ungated version here) for its original application: racial segregation of US neighborhoods (reminder to self: my job is not only to blog, also to be a full time academic researcher that must “publish or perish”). The basic prediction is that mixed neighborhoods are unstable but segregated neighborhoods are stable. Data on American neighborhoods from 1970 to 2000 rejected these predictions – it was the segregated neighborhoods that were unstable. There was as much “white flight” out of all-white neighborhoods as there was out of mixed neighborhoods, and there was a white influx into segregated nonwhite neighborhoods. Neighborhoods are still very segregated in the year 2000, but not because of tipping. Maybe segregation exists because most whites really do want segregation, not because of a chain reaction due to herd behavior.

Of course, this is only one test of the tipping point for racial segregation over one time period. Maybe the tipping point is real in other contexts. But think twice and check for evidence before you accept popular stories like the Tipping Point.

This entry was posted in Aid policies and approaches, Economics principles. Bookmark the permalink. Follow any comments here with the RSS feed for this post. Both comments and trackbacks are currently closed.


  1. Jim wrote:

    I was hoping you’d do a post about this, even though it isn’t really about aid, because it is fascinating as you say. I haven’t read all of your paper yet but it looks like a comprehensive analysis. I just want to make a few comments:

    As I understand it, Schelling’s ‘tipping point’ is simply a natural outcome of his very limited assumptions, such as people only caring about racial segregation and not about other issues such as ties to a neighbourhood or preferences for neighbours defined by other characteristics such as income and socioeconomic class. In the real world, people may have preferences consistent with the ‘tipping point’ in one respect which are outweighed by other preferences, or even a ‘tipping point’ process going in the other direction. As you say in your conclusions, the list of other potential factors is endless and empirically, for the US between 1970 and 2000, they seem to swamp any racial tipping point effect.

    But some kinds of tipping point models, used more cautiously, may still be useful in explaining various phenomena. For example, the existence of cities may be evidence of ‘tipping points’ in productivity brought about by increased density.

    Also, Schelling’s model obviously leaves out the effect of any policy measures on segregation. Insofar as government policy tends to reduce segregation, it will make any empirical proof harder to find even if there are social processes going on consistent with the theory.

    Posted June 30, 2009 at 9:13 am | Permalink
  2. Aren’t generational differences likely to skew the results? Every new generation creates diversity out of homogeneity. The children are unlike the parents and want to go in search of themselves (themselves: people like them).

    Furthermore, I would postulate that recent generations are becoming gradually less racially identified, with a long list of reasons and evidence (with the most interesting example being the last election) and more identified with people who share other things in common such as goals and values.

    Posted June 30, 2009 at 9:29 am | Permalink
  3. Adam Hooper wrote:

    “Suppose that people decide to get highly educated based on what is the share of highly educated people in the population.”

    …This example works both ways. Just as you can imagine an educated population migrating towards ignorance, so can you take a completely uneducated population and watch every member get a PhD. (I almost suspect you wanted a commenter to pick up on this…?)

    I’m no economist and I haven’t read any papers, and surely Gladwell points out obvious exceptions, but nevertheless this “tipping point” concept strikes me as a fun retrospective curiosity, not a predictor of human behaviour.

    Posted June 30, 2009 at 9:47 am | Permalink
  4. One of the key variables seems to be the number of different traits used to assess cultural/ethnic parity:

    Posted June 30, 2009 at 11:13 am | Permalink
  5. Adam wrote:

    After all, it’s only worthwhile being educated if you can talk to and work with a lot of other highly educated people.

    Could there not also be a positive feedback loop here? If those who are highly educated make up only a small proportion of the population, would the benefits of such education not encourage others to join and push this effect in the other direction?

    Posted June 30, 2009 at 11:23 am | Permalink
  6. Simon Pfister wrote:

    I think the tipping point concept is very compelling. And it is used in other sciences such as management (e.g. in change management you should influence the few key opinion leading employees and once convinced they will ‘drag’ the others to follow), environmental studies, or physics (namely the critical mass for nuclear reactions).

    However, I believe that there are different reasons why a person should or should not follow the herd. E.g. if I want to earn above average salary I might still opt for education (unless the definition of ‘talking to others’ includes to earn high salaries). Arguments relating to the neighborhood may include that white move out of white neighborhood because of new job, or if a young person moves to a new city he chooses any neighborhood, only once she/he has a family the selection of the neighborhood enjoys a higher priority.

    In sum: I think it is a concept that has a lot of truth for development work, but other factors may ‘interfere’ and therefore an observation of pure tipping point concepts may be seldom.

    Posted June 30, 2009 at 11:47 am | Permalink
  7. Steven wrote:

    I toiled through Gladwell’s book and it is essentially he just a guy who read a few Psych, Econ, and Sociology 101 books found some support for his term and ran with it. I think the only redeeming (intellectually) point of his book is the last chapter.

    Anyone read any other critiques on Gladwell?

    Posted June 30, 2009 at 4:33 pm | Permalink
  8. Steven wrote:

    Ugh sorry for the typos.

    Posted June 30, 2009 at 4:36 pm | Permalink
  9. Jeff Barnes wrote:

    The “tipping point” is just an intellectual construct, so it doesn’t surprise me if there is no objectively verifiable point at which a social change accelerates. The value of Gladwell’s book is in breaking down the different ways in which diffusion of innovation and social change occur. In particular, the notion of the role of mavens, connectors and salespeople is important to any diffusion of innovation strategy and has clear applications to development. To over generalize, the development community tends to be dominated by mavens (largely academics– sorry Bill) who don’t know how to make the connections and make the “sales” needed to introduce innovations.

    Posted June 30, 2009 at 5:34 pm | Permalink
  10. Dan in EuroLand wrote:

    William Easterly,

    Its interesting that you describe a tipping point model of education. Here is an interesting paper presenting such a model.

    In the paper investments in human capital depend upon the human capital present in an individual’s social network. High human capital in the network translates into the person making the investment in high human capital. Low human capital in the network leads to the opposite result.

    Theoretically I think the paper is very interesting. However as your paper demonstrates, the empirics may no be so clean cut.

    Posted June 30, 2009 at 11:23 pm | Permalink
  11. John wrote:

    Has any used this theory to test knife violence, or even the carrying of knives, amongst teenagers in London.

    Many of the teenagers there now carry blades, which they claim are from self-defence, against the increasing number of young people who are carrying knives. Its a variation on that old chestnut – do guns make us safer? This is Britain’s teenaged alternative, but one that would seem to follow a tipping point type growth. Does anyone know of any research on this?

    I am also not sure of whether the tipping point is a refinement of the path dependency idea? Any one able to explain to me the difference?

    Posted July 1, 2009 at 5:21 am | Permalink
  12. zulusafari wrote:

    I’m surprised not to hear a single example of this in aid or developing country economics. I know of no real world examples but could think of plenty potential/hypothetical examples right off the top of my head.

    Posted July 1, 2009 at 8:36 am | Permalink
  13. kevin quinn wrote:

    Does this mean I should be skeptical about my favorite chapter in the Elusive Quest for Growth (indeed, one of favorite chapters, period) – chapter 8? Say it isn’t so, Professor Easterly!

    Posted July 1, 2009 at 10:47 am | Permalink
  14. The work of Tostan is a rare example of a tipping point in action during a larger process of a shift in a social norm and/or social convention. Over the past 10 years, local community movements to abandon female genital cutting and child/forced marriage have very much followed what one would expect–not only in terms of tipping points, which are of course only one feature of changes in an interdependent behaviors, but also in terms of a core group achieving critical mass and the necessary mechanisms required for collective abandonment.

    Posted July 1, 2009 at 4:55 pm | Permalink
  15. MiGrant wrote:

    I find it a little disconcerting that the housing segregation example seems to assume no capacity for choice on the part of black residents…?

    Posted July 1, 2009 at 6:49 pm | Permalink
  16. Stephen Jones wrote:

    After all, it’s only worthwhile being educated if you can talk to and work with a lot of other highly educated people.

    That’s what the internet’s for.

    Posted July 6, 2009 at 5:40 am | Permalink
  17. Jason wrote:

    I know I’m late to the game, but a bit of insight on the “white flight” narrative in the US:

    “White flight” is conventional wisdom, but it misses the huge policy forces bearing on whites families’ decisions to relocate away from city centers–namely, the incredible incentives extended to them through the FHA’s grants, which catalyzed the development of suburbia. Guess what? FHA grants, while federal in source, were put under the control of local officials–officials that were patently race-biased in determining who received FHA grants.

    An excellent discussion of this can be found in Robert O. Self’s book _American Babylon_.

    Posted July 7, 2009 at 3:37 pm | Permalink
  • About Aid Watch

    The Aid Watch blog is a project of New York University's Development Research Institute (DRI). This blog is principally written by William Easterly, author of "The Elusive Quest for Growth: Economists' Adventures and Misadventures in the Tropics" and "The White Man's Burden: Why the West's Efforts to Aid the Rest Have Done So Much Ill and So Little Good," and Professor of Economics at NYU. It is co-written by Laura Freschi and by occasional guest bloggers. Our work is based on the idea that more aid will reach the poor the more people are watching aid.

    "Conscience is the inner voice that warns us somebody may be looking." - H.L. Mencken

  • Archives