50th Anniversary of Muddling Through by Lindblom

I've been putting up pictures instead of writing about things like Mayor Sullivan's veto of the ordinance to add gays and lesbians et al to the Anchorage anti-discrimination ordinance or about health care. It's not because I haven't been thinking about both, but I have been waiting to say something that adds to the discussion.

We got up at 5 am today to get DZ to the airport, so my head isn't totally clear (well, that implies that it would be otherwise and I don't want to give any false impressions), so let me offer one short perspective on the health care issue.

Back in 1959 - I guess that makes this the 50th Anniversary - Professor Charles Lindblom of Yale published an article that is still in the reading lists of public administration, public policy, and political science students. It was titled, "The Science of Muddling Through". (The link will only get you the first page at JStor. If you have access to a university library system, you ought to be able to get it all free. If not, check with your local library to see if they can get it for you. Alaskans should be able to get it through SLED.)

In this article, Lindblom argued that the traditional way of thinking about policy - rationally setting up a goal and then a plan to reach that goal - was nice theoretically, but that Congress didn't work that way.

Instead, changes were made incrementally, because Congress was made up of a 100 Senators and 400 and something Congress members who all had a vote, didn't agree on much, and had local constituents to please. Therefore, if you wanted people to agree on the goal, nothing would ever be accomplished.

But, he said, they could and do agree on the means through which each committee member can fulfill their individual goals.

He called the first method - the one in the academic journals - The Rational Comprehensive method, or Root method. Root, because you started all the way from the roots to build something completely new.

He called his model,

"I propose in this paper to clarfiy and formalize the second method, much neglected in the literature. This might be described as the method of successive limited comparisons." (p. 199*)

He also called this the Branch method, because you aren't going all the way back to the roots. You work from the existing system and just do some work with the branches.

[Double click to enlarge]

So, basically, he's saying that (and I'm reducing his points to fewer ones, so the numbers don't match the ones in the chart.)

1. The root method assumes that first you agree on goals, then design the means to meet those goals. But the branch method says that the means (policies to achieve goals) and ends (the goals) are intertwined and can't be easily separated and articulated that way. (1 and 2 in the chart)

2. The test of a good policy isn't whether it meets the objectives (since they were never agreed on) but whether there is agreement on the policy. (3 on his chart) One of his 1959 examples was that the agreement in Congress to extend old age insurance

. . .stems from liberal desires to strengthen the welfare programs of the Federal Government and from conservative desires to reduce union demands for private pension plans. (p. 202)

A different example of agreeing on the means (but not goals) might be a Senate committee agreeing to support a new ship for the Navy. One senator wants to have parts for the ship built in his state. Another will vote for it if it includes the Navy giving wetlands it owns for conservation. Another wants minority businesses to get at least 10% of the contracts. A couple actually think the ship is needed for national defense. And perhaps another will vote for it if the son of a particular constituent gets into the Naval Academy. They all have different goals, but they agree on the means (the ship) to their separate goals.

3. While the Root ideal is for comprehensive examination of the new policy, the Branch reality is acceptance of the basic existing system and just reviewing the additional incremental changes. This is the "successive limited comparisons" notion.
Complete analysis is not humanly possible and so simplification needs to be achieved.

First, it is achieved through limitation of policy comparisons to those policies that differ in relatively small degree from policies presently in effect. . .(p. 203)
The second method of simplification of analysis is the practice of ignoring important possible consequences of possible polices, as well as the values attached to the neglected consequences. (p. 204)

Lindblom recognized that ignoring the consequences may sound shocking but argues that in fact it may lead to better policy than "through futile attempts to achieve a comprehensiveness beyond human capacity." (p 204)

4. The root model relies on theory while the branch method reduces the reliance on theory through successive comparisons. Lindblom argues that our knowledge in the social sciences is, at best, limited. We can make practical comparisons of the consequences of incremental policy changes - we can look at how some past programs worked or didn't, for example - without necessarily having a strong theory.


A lot has happened in 50 years, including increasingly sophisticated policy analysis techniques. Furthermore, there are some assumptions in Lindblom's model here that I have problems with. As a descriptive model - describing what is actually happening - I think he did a good job. Adopting it prescriptively - saying this was a good way to go - was probably under[over]-ambitious. And Lindblom himself got tired of people not getting past this early piece of his and seeing that he'd done more than that in later years.

I'd note that as a doctoral student I took an intensive -week class on the policy making process with Professor Lindblom. He did acknowledge that individual policy makers may use a more rational comprehensive model in their individual strategies to achieve their own goals. But when power is dispersed, the agreement on goals becomes very difficult.

He, at that time, tended to believe that while it wasn't perfect, it was as good as we can get. I agree that having representatives of a cross section of the US population making decisions is definitely better than trusting a single decision maker to make the best decisions for us all. But I also think there are better ways than we have for funding and communicating with our elected officials. One of the problems I had then and have now, is that a major goal of members of Congress is to get re-elected. This means they are always in search of money and thus they are always in debt to large donors and key supporters.


I started this intending to use it to comment on Obama's health care proposals. I'm not sure it tells us a lot beyond the fact that Obama's health care program is hardly a minor incremental change and that makes it much harder to get through. People can legitimately question whether it will work. They won't be making minor moves toward some new program. It will be big time. On the other hand, incremental changes on something that works so poorly for so many and costs so much isn't good policy either.

People who are getting good health care now, fear changes will hurt them. So they don't agree with Obama's goals. People who don't like Obama may oppose anything he proposes to make him look bad. And if Obama moves toward normal incremental changes, he'll lose his base which expects him to go for the gold.

But politics today are not what they were 50 years ago when Lindblom wrote the piece. We'll just have to see if Obama writes a new page in the policy making literature.

[Yeah, I know, this is a lame ending. I got up early and I can't pull it together. So if you've read this far, you'll will have to mull on it yourself - is Lindblom still relevant? If so, what can Obama do at this point? If not, what's different now?]

* I had trouble getting the article through JStor. While it shows the first page on Google, trying to get it through the library didn't work. Probably the librarian can make it work, but I have a copy of Shafritz and Hyde's Classics of Public Administration and the article is in there. So the page numbers come from there, not the original.