Monday, November 16, 2009

Prof. Richard Florida resolves the urban transportation problem

Not since the seminal work by Meyer, Kain, and Wohl on the Urban Transportation Problem, which was published by the RAND Corporation in 1965, has any one been so insightful as Professor Richard Florida of the University of Toronto.  In his blog hosted by the Atlantic,  Professor Florida has found the drive alone mode to be negatively correlated with the following demographics, among others, at the State imagelevel in the United States:

  • Immigrants
  • Gays
  • High-income households
  • University educated individuals

Prof. Florida’s analysis provides us with enough insights to formulate urban transport policy recommendations to discourage drive-alone mode in the United States. These are:

  1. Make people richer. Get rid of the low-income and working class people because they drive alone and cause congestion.
  2. Import rich immigrants because indigenous populations drive alone and cause congestion.
  3. Get rid of the under-educated or send them to the University because highly educated do not drive alone.
    1. Also, increase course offerings in gay and lesbian studies at the universities.
  4. Lastly, instead of focussing on the interaction between the built form and travel behaviour, public policy should concentrate on personality traits because those influence travel behaviour more.
    1. Therefore, as a public policy measure to discourage drive alone mode, one should discourage conscientiousness and advocate carelessness.

On a serious note, I have the following concerns about the professor’s analysis:

  • Prof. Florida  is trying to find correlations at the State level by mixing urban and non urban commuting in the United States where transit ridership outside of large cities is almost negligible. Therefore any correlation between drive alone mode and demographic traits observed at the State level is almost meaningless.
  • Spurious Correlations: Given the strong interaction between the built form (population and employment densities) and the mode of travel, any correlation found between other demographic traits would be influenced by the built form, which acts as a confounding factor.
    • A typical example of this error is demonstrated by the positive correlation between drownings in city’s swimming pools and ice cream sales.  Obviously, it would be erroneous to conclude that an increase in ice cream sales causes an increase in drownings. In fact, the heat wave (confounding factor) is behind the increase in both drownings and ice cream sales.


Prof. Florida’s blog is reproduced below:

Oct 9 2009

Culture / Media

Driving Alone - A Quick and Dirty Analysis

Earlier this week Catherine Rampell posted this map over at Economix. It shows the percentages of workers who drove to work alone by state and is based on U.S. Census data.


D.C. has the lowest rate - a fact which was not lost on D.C. blogging circles. NY did well too.  The worst performers were Alabama, Tennessee, and Ohio, where about eight in 10 workers drive alone -  more than double that of D.C.

With the help of my colleague Charlotta Mellander, we took a quick look at some factors that might be associated with this geographic pattern. It's not an exhaustive list: We examined some key economic factors like income and economic output, human capital and the creative class, and psychological ones like happiness, stress, and personality. We removed D.C. from the analysis because it was such an extreme outlier. We did not develop or run any serious multivariate analysis - just simple correlations, or associations, between variables.

Still the findings point to some reasonably clear patterns.
Income and Economic Output: The richer the state, the less likely people were to drive alone. Driving alone was negatively correlated with state income levels (-.46) and output per capital (-.41).

Class and Human Capital
: States with higher percentages of college graduates (-.47) and the creative class (-.43) were less likely to have people driving alone. Driving alone was much more likely in states with large working class concentrations (.62).

Professional and Creative Jobs:
Driving alone was less likely in states with high concentrations of virtually every type of professional, knowledge-based and creative jobs. But it was least likely in states with large concentrations of artists, designers, and entertainers (-.63), architects and engineers (-.61), scientists (-.56 ), and lawyers (-.55).

Diversity - Immigrants and Gays
: Driving alone was less likely in states with high concentrations of immigrants (-.51) and gays (-.41).

Happiness research tells us that commuting is one of life's least pleasurable activities.  Driving along was negatively associated with state levels of happiness and well-being (-.46) and positively associated with states with higher levels of stress (.29).

Psychologists identify five main personality types. Driving alone was more likely in states with high levels of three of them: extroverts (.29), conscientiousness (.36), and agreeableness (.44). Interestingly, there was no association between driving alone and the two other types - neurotic and openness to experience, which some might say makes it harder to explain New York.

No comments:

Post a Comment