The cul-de-sac layout so ubiquitous in suburban neighborhoods can be partly explained as a collective-action problem.  But any explanation of its dominance must also acknowledge that developers use the cul-de-sac layout because it is cheaper.  Developers get more homes for less asphalt.

Take this neighborhood off Convict Hill in southwest Austin (outlined in blue). This is the prototypical suburban subdivision -- it's got several cul-de-sacs (blue pins) and just a few entrances (marked by Cookie Monsters). The interior streets (green) have little connectivity.  As a result, the neighborhood is inhospitable to through traffic. But it is also difficult for its residents to get from one point to another, or even to exit their own neighborhood. For example, someone on Vol Walker Drive (the yellow pin) has to take a circuitous route to get to the Convict Hill entrance (the red pin).

View Larger Map

But this neighborhood illustrates why developers like winding roads and cul-de-sacs:  they save infrastructure.

In order to estimate how much infrastructure the cul-de-sac layout saved, I've drawn in streets to complete a pseudo-grid (red).  It takes a lot of them. Gmaps tells me these "missing" streets together measure 4,905 feet long. The existing (green) streets are 7,390 feet long, which means the developer would have needed 66% more asphalt, street lamps, curbs and sidewalks to complete the grid.  Since sewer and water lines follow the streets, he would have needed 66% more sewer and water pipes as well.  And all that extra asphalt would have left less space for houses.

This subdivision would have cost a lot more to develop around a grid. This additional expense likely would have been passed back to the property sellers  or passed forward to the home buyers.

But if we are trying to determine the optimal street configuration, this can't be the end of the analysis.  The cul-de-sac layout imposes costs on the rest of us.  Laissez faire allows developers to socialize these costs while internalizing the benefits.  My view of these things is to price everything properly and then let God sort it out.  

The price of disconnectivity should include the higher fire station and EMS costs. On the other hand, the developer should get a credit for shorter water and sewer networks, which save the city money in maintenance.

The main problem with disconnected street networks, though, is that they exacerbate congestion and reduce mobility.  It might not matter if only one subdivision here or there eschewed the grid.  But when every subdivision does, traffic is channeled into overcrowded through streets and everyone is forced to take longer trips.  Because the cul-de-sac layout provides purely private benefits, the city should leave road maintenance costs on the subdivision.  (Note that the costs of the subdivision's design to the subdivision's own residents are reflected in their home prices, and are therefore not social costs.)

Shrugging off maintenance costs might not cover the city's loss, though.  Extending the grid rather than lopping it off might save a couple of minutes per day for hundreds or thousands of people.  That adds up.  Providing better opportunities for other modes of travel is worth a lot, too.  Saving a paltry few thousand dollars per year in maintenance might not offset these losses.

I'm not sure how to price the value of connectivity to the city and its residents -- "a lot" is the best I can offer.  I do think it's important to try, though, rather than simply mandate it.   As the example above shows, a street grid might need massively more infrastructure than a cul-de-sac layout.  We should give the developer the incentive to lay the right amount of asphalt rather than simply assume a grid is always the optimal solution.