Why does government work so well? Huh?

In this post on EconLog, Bryan Caplan explores why government enterprises work so well.

He makes a good point.

I think the small-government types (like myself) can overplay disasters of government involvement and we lose credibility when we do so. So, I do think its helpful to recognize when government seems to be doing, at worst, okay.

On the flip-side, I think big-government types can overplay the successes of government enterprises.

But, I think much of this is explained to the extent of what level of government we are talking about and the dynamics of that level, to what extent it is bottom-up or top-down.

I discussed this in more detail in this post back in 2013. I think government enterprises that work pretty well are more bottom-up and the ones that don’t work so well are top-down.

That post was inspired by an apples-and-oranges comparison often made by government-types.  They say that fire and police departments are government, and work pretty good. Then they make the logical leap to use this to support that some Federal government enterprise will work.

The lapse in that logic is that fire and police departments, while government enterprises, operate at the local level. There are thousands of these departments, that operate rather independently of one another, across our country. This makes these enterprises operate much more like a bottom-up organization, than top-down. This allows these enterprises to benefit from the same dynamics of innovationism as businesses.

Advertisements

The spider and the starfish

starfish

starfish = bottoms-up (Photo credit: kevinzim)

I enjoyed listening to this EconTalk podcast with guest Lant Pritchett about education, specifically in poorer countries. I recommend it.

He begins with a nice discussion about bottom-up and top-down systems using metaphors (emphasis added):

…the basic distinction is between a top down organization where the metaphor of a spider is, all of the resources of the spider web, however spread out they are, merely serve to transmit information to one spider, who synthesizes that information and responds with the resources of the system. So, if there is a bug, the spider crawls out and gets it. But kind of all the web is kind of an ancillary to the brains of the top. The starfish is a creature that actually has no brain. It’s neurally connected, but a starfish moves because the individual units of the starfish sense something and if they sense more food they try and pull that way. And if the other side isn’t pulling as hard, the starfish moves. So it’s really a metaphor of a decentralized system, where individual units responding to local conditions create the properties of the system. And the beauty of a starfish is if you cut a starfish up into 5 bits, you get 5 starfish. The danger of a spider is that when the spider dies, it’s dead. The whole system therefore falls into dysfunction [single point of failure].

Later, he proposes why bad school systems may persist:

…one of the conjectures I put in the book is that it persists partly by camouflage. It pretends to be something it’s not and then can project enough of the camouflage that it maintains its legitimacy. So, sociologists of organization have a term called ‘isomorphic mimicry’, which is adapted from evolution where some species of snakes look poisonous but aren’t, but get the survival value of looking poisonous. So, one of the things that’s happened is by this pressure to expand schooling and by the governments’ desire to control that socialization process, they have created all the appearances of schools that provide education but without actually doing it. But have at the same time not produced the information that would make it clear that they weren’t doing it. So they produce enrollment statistics, numbers of buildings, numbers of toilets, numbers of textbooks, numbers of everything. But have, you know, all of which can project the image that there’s a functional system and providing real learning there. But they don’tprovide metrics of learning or incentives for learning or feedback on learning or accountability for learning at all.

There’s much more good discussion throughout, but I’m fond of something Pritchett said near the end:

You know, the United States has always been much more of a starfish system. And the starfish system has enormous strengths, and I think those enormous strengths have led America to be a leader in education in many ways. And one of the examples I use in my book is, if there’s a scaled example of a starfish education system, it’s American universities. And it’s just unbelievable from the data the extent to which America dominates quality universities. It’s just unbelievable compared to Europe, which always took the same approach to universities that other countries want to take to their basic education. And you see the consequences of it. America’s universities–in the book I have numbers of the top 200 universities in the world, what fraction of them are in Anglo countries, and it’s just way disproportional to the population size. And even wealth. Because Europe, which is equally sized and equally wealthy, continental Europe, just has nowhere near. And it’s the result of a starfish system, in my view.

I’ve made the same point several times on this blog. Here and here are a couple of examples.

I do appreciate the spider and starfish metaphors. Those roll-off the tongue better than top-down and bottom-up.

Bottom-up vs. Top-down: Don’t put all your eggs in one basket

This is the 3rd post in this series. Here are posts 1 and 2.

In post 1 of this series, I explained what I think are the key differences between top-down and bottom-up organizations and why it’s helpful to think in those terms, rather than other common organization characterizations like government vs. private sector.

At the end of post 1, I listed three reasons why I think bottom-up systems work better than top-down. In Post 1, I elaborated on the first reason. In this post, I elaborate on the second: No single point of failure.

This has been conventional wisdom for a long-time and you may know it better as the phrase, Don’t put all your eggs in one basket.

Why? Because you might drop the basket. This sage advice helps reduce the risk of breaking all the eggs should you drop that basket.

We use this advice when investing. No matter how much homework we do on a company, there are no sure bets. Best not to bet everything on one company.

We consider the advice when planning careers. We train for one career path, but we know it could be automated or outsourced, so it’s best to have backups.

Sports teams try not to bank too much on a single player. Great players are good to have, but they can get injured.

Engineers try to avoid single points of failure when designing systems. Bridges are designed with redundant supports, so they won’t fall if a single support fails. Systems without single points of failure save lives.

Bottom-up systems do not have single points of failure. Baskets can be dropped in such systems. Eggs will be broken, but there are plenty of other baskets to go around.

Why is this good? Because failure happens and it happens more often than success. We live in a trial-and-error universe.

Capitalism is a good example of a bottom-up system. When one business fails, there are others to take its place. It doesn’t take down the whole system. We survived Enron’s collapse, for example. It was not ‘too big to fail’.

Local government is also a good example of a bottom-up system. Local governments can and do fail. Detroit is failing, but it’s not bringing down the whole system. There are thousands more cities, counties and townships. 

School districts don’t yet have a single point of failure. Failing school districts do not bring down the whole system.

Though, school districts have moved toward becoming more top-down over the past few decades as a small group of folks in DC use taxpayer dollars to encourage school districts to deliver on what the folks in DC think is a good education.

This has moved accountability away from parents toward a central point of failure, the ‘common core’ curriculum.

Of course, ‘too big to fail’ is code for ‘single point of failure.’ If it is true that some organization has become ‘too big to fail’ (though I don’t think that was the case in the financial crisis), we should spend more time thinking about how we let a single point of failure crawl into our lives, much the same way the common core curriculum is doing now.

Bottom-up systems are not painless. Failure can be painful. But, bottom-up systems deal with pain and failure better than top-down systems.

Attempting to avoid pain and failure is one reason people advocate for top-down systems. Unfortunately, they soon learn that was a fairy tale.

One reason bottom-up does better: Try, try, try again

In this post, I wrote that bottom-up systems tend to do better than top-down systems. In the following post, I expand on the #1 reason why I think that is.

I use to think feedbacks were the most important difference between productive and unproductive organizations or systems.

I don’t anymore.

I still think feedbacks are important and much can be gained from addressing feedback problems. But, I think something else makes the difference between good and bad: The number of trials.

I credit Nassim Taleb for planting this thought in my head. In one of his books he said something like capitalism doesn’t work because of profits and losses [feedbacks], it works because it induces a lot of trials and some of those happen to work.

I think Jeff Bezos understands this idea about trials. But, I don’t think many others do.

We see success stories after the fact and credit visionary leadership, brilliant insights and clever innovations, but this is too simple of a view of what really happened.

Since Taleb planted this idea I started noticing other things about success stories.

I noticed that for every success, there were dozens or more failures. Successful people are often brilliant. We conclude that to be the cause of their success, but that doesn’t account for the failures that are often led by equally bright people. Why did one succeed while the others didn’t? That gets tougher to distinguish.

A story Taleb tells about dolphins illustrates this well. We hear about dolphins pushing to shore people who are stranded at sea and conclude dolphins know what they’re doing and are kind and gentle. Maybe that’s true. But, we don’t hear from the folks when the dolphins took them the other way. Maybe the dolphins that pushed to shore just got lucky.

I also started noticing that success stories often includes information about how the discoveries were not planned, but were accidentally stumbled upon while the great leaders were trying to do something else. Sometimes these leaders even resisted going the direction the discovery led them because it didn’t fit their original vision.

Finally, I started seeing the dumb luck that is a part of almost every success.

When I look at success, I try to remember its causes can be elusive and that the easy explanations are usually incomplete.

Capitalism works well for us because it encourages a lot of trials.

I believe organizations can improve (not guarantee) chances of long-term success by employing the same incentives as capitalism. Encourage a lot of trials. Let failures fail and reward success.

We even have evidence of this respect for trials embedded in common, inspirational phrases like:

  • If at first you don’t succeed, try, try, try again
  • Fail often to succeed sooner
  • If you fall off the horse, get back on

But, perhaps one quality of the successful that I may have overlooked is that they aren’t afraid to fail.

You miss 100 percent of the shots you never take. – Wayne Gretzky

I didn’t fail the test, I just found a 100 ways to do it wrong. -Ben Franklin

I’ve never been afraid to fail. -Michael Jordan

Bottom-up vs. Top-down

Forget comparisons like ‘government vs. free market’ or ‘public vs. private’. We can all pull out examples from each realm that works and doesn’t work.

I think a better distinction of human interaction systems can be drawn between ‘bottom-up vs. top-down’. Bottom-up systems tend to be driven by decisions and interactions very close to the individual. Top-down systems are driven by decisions made by a few individuals at the top of the hierarchy.

I often hear proponents of a central government — a top-down system — cite bottom-up examples to support their view, demonstrating that they don’t understand the top-down/bottom-up distinction.

Fire departments, police departments and roads are prominent examples of this and sound like, “These things are run by government and they work pretty well, let’s expand Federal government.”

Yes. These are usually run by government entities. But local government is not equivalent to Federal government. Local government is more bottom-up than top-down.

I think it’s helpful for discussion and policy-making to be able to recognize the two different systems. Here are a couple of thoughts.

  • To what degree do individuals have the ‘power of voice’?
  • To what degree do individuals have the ‘power of exit’?

Power of voice is an individual’s vote, purchasing decision or the ability to have the decision-makers hear your voice. Votes carry more weight in local elections than national elections. It’s easier to get in touch with city councilmen or the mayor than it is the Congressman or President. A purchase is a vote indicating that the individual values the product at that price.

Power of exit is the individual’s ability to leave or not choose a certain product. If I choose not to buy a product, I’m signaling that it’s not worth it. If I don’t think the city council is steering my city in the right direction, I can move a couple of miles to the next city over. Same if I’m looking for a better a school district.

Power of exit is a function of how many choices are available to me and the dynamics of competition among those choices.

I have lots of choices of products. Competition among those products is generally high. I can incur little cost for choosing one store over another or one product within a store over another.

In my area, I have lots of choices on cities, townships and counties that I can live in. Not as many choices as I have for different types of beer, but still quite a few. Since I do have a fair amount of choice, competition is relatively high and these cities and townships try to do things to voluntarily attract residents — like providing good fire departments, police departments, schools and roads. The cost to move is higher than the cost of choosing one product over another, but still not terribly high if things get really bad (this is why suburbs suck the life out of mismanaged urban areas).

I have 50 states to choose from, but the costs to move between states is relatively high. I have many countries to choose from, but the costs to change countries is even higher. So, competition goes down as these costs increase and so does my power of exit. I still have it, no doubt, but things have to get really bad before exercising that power.

I think there are a couple other elements of the bottom-up vs. top-down systems that are worth mentioning. I don’t think they are as useful in helping us recognize these systems, but I think they explain why bottom-up system work better than top-down.

1. Lots of independent trials/experiments. It’s hard to tell what will work and what won’t, so it’s best to try a lot of stuff.

2. No single point of failure. Since it’s hard to tell what will work, it’s best not to have a single point of failure, which top-down systems have.

3. How far apart are the benefits and the costs? Or, another way to think about it, how far removed are the decision-makers from the consequences of their decisions? This gets down to, how much skin-in-the-game do decision-makers have?

I will discuss more about these last three in future posts.