Criticism of the complexity argument, continued

This is part of a series of posts about complexity and implementation. In When Things Break I presented the idea of asking  what happens when one part of a system breaks. In The Broken ATM I discussed experiences during fieldwork in rural Brazil that got me thinking along these lines. Then in Wiring Government Programs in Series or Parallel I summarized Pressman and Wildavsky’s argument that the more “clearances” a program needs the more likely it is to fail.  Criticism of the complexity argument summarized one line of criticism of their approach. This post looks at another response.

Last time, I looked at one response to Pressman and Wildavsky: their assumptions about how often individual components stymie the whole project are completely arbitrary. We live in a world where we are surrounded by complex systems where functioning well is the norm and expectation (and others that provide endless frustration… but we’ll get to that).

Elinor R. Bowen proposed modifications to the joint-probability approach. Again, joint probability is the probability of several things happening together: so the joint probability of two coins both flipping heads at the same time is (.5)(.5)= .25. She offers four addendums which result in a much more optimistic outlook, and which she argues are reasonable descriptions of many policy implementation processes. We’ll look at the first three (I give a reference to the article at the end of this post).

The first is persistence. Those implementing a policy often get more than one opportunity to pass a particular clearance. As the old saying goes “If at first you don’t succeed, try, try again.” To go back to the metaphor of Christmas lights, just because there’s a dead bulb on a string wired in series doesn’t mean we have to throw it out. We could keep replacing the broken bulb until we find one that works. Of course this takes time.

The second is packaging clearances together to get everybody on board for a set of clearances. In other words, look for ways to do several things at once. (Again, a “clearance” is any place where those implementing a program need to get someone else to sign off on something for it to move forward–i.e. a chance for something to break).

The third is bandwagons, with earlier successes increasing the probability of later ones. This makes initiative sense: if things have been going well people are more likely to take a new program seriously and cooperate.

All three take time, energy, and resources. The implication for implementation (say that three times fast) is the following:

…let us imagine a hypothetical implementer with limited resources such that he or she can only enter into negotiations for twenty clearances. The implementer is confronted by three options: (1) he or she can attempt to implement a relatively complex program requiring twenty independent clearances, (2) he or she can attempt implementation of two smaller independent programs, each requiring ten clearances, or (3) he or she can ruthlessly establish priorities and attempt implementation of one small program requiring ten clearances. Only this third option reserves sufficient resources for persistence in the attempt to gain clearances (p 15).

In other words, expect to have to spend resources to be sure to get off to a good start, and to make multiple goes to get past a hangup.

Further reading:

Bowen, Elinor R. 1982. “The Pressman-Wildavsky Paradox: Four Addenda or Why Models Based on Probability Theory Can Predict Implementation Success and Suggest Useful Tactical Advice for Implementers.” Journal of Public Policy 2(1): 1–21.