Arnold Kling  

Size, Scope, and Fragility

What I'm Reading... How Free Were American Women i...

Andrew Haldane writes,

Complex control of a complex system is a recipe for confusion at best, catastrophe at worst. Complex control adds, not subtracts, from the Knightian uncertainty problem. The US constitution is four pages long. The recently-tabled Dodd Bill on US financial sector reform is 1,336 pages long. Which do you imagine will have the more lasting impact on behaviour.

It is a long speech, but worth reading in its entirety. He describes a financial system that became more fragile as banks became larger and more complex. He describes the issue as "neurological"--humans lack the ability to manage such large, complex systems. He does not invoke Hayek, but he easily could.

I believe that the same holds for government. Our government has become more fragile as power has become more concentrated at the center and the scope of government has increased.

This fragility of tumescent government is what many otherwise thoughtful commentators fail to notice. They think that we can solve our problems with a value-added tax (VAT) or with a fiscal reform commission.

Instead, I believe that the problem is structural. American government will not be stable and sustainable until it is broken up and most of its functions are devolved, so that spending decisions are decentralized. A robust society would have more public goods delivered by the voluntary associations that Tocqueville noticed were characteristic of America. These associations are much more numerous and diverse than our governmental units.

This thesis would suggest that China is fragile. It may do very well for a time, but it will be subject to sudden and catastrophic failure. On the other hand, Europe remains highly federalized. At the moment, it appears to me that the fate of individual European countries is less tied to any central authority than the fate of individual American states. Sweden is somewhat insulated from mistakes made in Athens or Brussels; Maine is not so well insulated from mistakes made in Sacramento or Washington.

Comments and Sharing

CATEGORIES: Political Economy

COMMENTS (14 to date)
JPIrving writes:

China is so interesting, so much is run from Beijing but for years their regulators were toothless, only able to make token interventions. Despite illiberal laws much of the economy, especially small business, was able to work through informal arrangements.

As an example

The crack downs disrupted a well functioning informal system. It is as though a fairly free market economy was able to sprout up within the cracks of the state, and now the cracks are closing. It will be interesting to see what happens.

The Chinese are so entrepreneurial, and now many can afford to do business in the Tiger economies and see what real capitalism looks like. I cant see the middle class putting up with the current system forever, especially when they can see how much better the alternative is in Taiwan/Hong Kong/Singapore.

Boonton writes:

As I pointed out on the Krugman thread, the problem is backwards. Complex systems aren't prone to failure, simple ones are.

A market economy is a very complicated animal. When some slight change happens, like a sudden popularity for a particular toy, the incentive structure for thousands of actors is changed. A planned economy is the opposite, its not complicated, its simple, too simple. When some silly toy suddenly becomes popular, the 'toy planner' probably issues updates once a quarter. Sortgages develop, black markets form, the system is stressed and can eventually break apart.

Yes the planner may seem to have a complicated job. He will have plans thousands of pages long, numerous Excel sheets and so on. But that individual complexity masks a very simple system. His numerous books only capture a fraction of the complexity of the whole economy.

Now the US Constitution appears simple by the simplistic metric of 'how many pages is it?' but that's all wrong. By this silly standard a typical Stephen King novel is more complicated than, say, James Joyce's Ulysses. No one would say simply reading the 4 page Constitution would qualify you as a Constitutional lawyer. The 'simple' Constitution masks a lot of complex historical, cultural and literary baggage. The Constitution was kept simple for the same reason the description of a market economy is simple, it has to contain a lot of complexity!

Contemplationist writes:


The "planned economy" is simply a masking of complex individual interactions with a simplistic gloss of false control. Though i guess you could argue this is all semantics.

Regarding China, I think ethnic homogeneity definitely plays a role in a better-managed centralized state. The US of course is the polar opposite of China on those terms.

Doc Merlin writes:

Also, China isn't that centralized. It has a federalized system similar to ours in many ways.

Loof writes:

Relative to central planning and socialization: The melting-together metaphor to develop a homogenous society appears as apt for China as it is for America. Both promote extreme nationalism of the absolute nation-state with hand over heart idealism. Media and schooling systems are permeated by propaganda.

Troy Camplin writes:

A decrentralized polity would in fact be more complex than a centralized one. Again, there is a mistake being made in calling complicated systems complex. The kind of rigid hierarchies that necessarily come with centralized government is simple and simplifying, not complex. Complex systems are more robust than are complicated/simple systems. The U.S. federal government hasn't grown too complex -- it has grown too complicated. It has grown too big for what it can actually accomplish. To decentralize and distribute political power to the states, counties, cities and towns would be to make the system both more complex and, as a result, more robust. The same is true of economies. If we allow them to become more complex, we'll have a more robust economy, less prone to these kinds of catastrophic collapses. The evidence is that it is simplifying factors which cause problems, not the complexifying ones.

Ron writes:

[Comment removed pending confirmation of email address and for policy violations. Email the to request restoring your comment privileges. A valid email address is required to post comments on EconLog.--Econlib Ed.]

James A. Donald writes:
  1. China has unofficial decentralization: The emperor is over the hills and far away. The emperor's minions are happy to have it that way, and the emperor's ability to do things that make his minions unhappy is less than it seems.
  2. Chinese business is run out of foreign tax havens through encrypted servers. This is illegal, but efforts to change the system are not going anywhere and not likely to go anywhere.
  3. Democracy is failing, as illustrated by the thousands of pages of Basel II banking regulations that no one comprehends and no one obeys.
  4. Europe is ludicrously centralized, as Brussels daily spews out gigantic volumes of regulation binding on everyone in Europe, regulating strawberries and ATMs in hyperminute detail.
  5. A system that produces the sort of laws and regulations that democracy now produces cannot long endure.

Kurbla writes:

I do not think that complexity of a market economy is unreachable by central plan.

It was maybe the case 30 years ago, but today, central planner in Moscow could have data about every apple sold in fruit shop in Vladivostok practically same moment the capitalist fruit shop owner has that data, and run calculations equivalent not to just thousand excel sheets in three months, but billions of sheets in a second. Fruit shop owner has no data or incentives to do so.

It doesn't mean that all problems related with central plan can be solved today. For example, I can easily imagine fruit shop employee in Vladivostok sending false data, because he has some private interest in making central plan worse.

Boonton writes:


You clearly don't work in a huge corporation. I can assure you it very well may be true that once you have a huge amount of data today's computers can churn thru it and do trillions of calculations in seconds. Getting data, though, is hardly free.

For example, in clinical studies data is collected on patients including things like the drugs or therapy they received, various medical exams and test results. There are specialists whose job it is to design and administer the data collection and processing of this high priority data. Even so data must be double checked, scrubbed for errors and so on. Even with such important data and so much attention being paid to it for reasons of safety and getting valid study results, there's still a lot of problems.

If getting critical information like medical tests produces so many problems how easy will it to collect and transmit critical information like an apple sold at a fruit stand to a central database? Stores like Wal-Mart do build databases where they can track individual products and tell you where they were on their shelfs, when they were sold and which cashier rang them up.

This data collection isn't free, though and it still leaves a lot on the table. For example, how about large numbers of people going into Wal-Mart, looking at the milk and eggs but leaving without buying them. This might be noticed by a manager who might discover people are coming in hoping to get savings on milk and eggs but are finding the products not very fresh. His observation might lead Wal-Mart to reallocate its trucks so that the milk and eggs arrives faster.

Would a central computer and planner see that or would they just see that demand for milk and eggs was oddly off and opt to replace them with something else?

Troy Camplin writes:

Computer scientist Stephen Wolfram says that things as complex as the economy are inherently incalculable. In other words, no computer of any speed or of any size would be able to predict what will happen in the economy -- or what will be invented (the big thing a central planner cannot do). That's the thing about complexity: it is inherently incalculable and inherently unpredictable. Wolfram points out that it would take longer to calculate the outcome than it would be to just sit around and wait to see what happens. In other words, incalculable. Not only is there no way any person can have enough knowledge, there is no way any computer can have enough information. Hayek's information theoretic objection to socialist calculation applies equally to computers' abilities. Enough said.

DB writes:

One thing about the economy that has always held basically true is that the economy follows basic patterns, but the way it fluctuates cannot be predicted perfectly. Trying to create system upon system to regualte it seriously complicates things and makes it more difficult for businesses to function. The sheer size that the federal government has grown to be has become a hinderance for the market, in my opinion. By expecting the federal government to regulate the entire economy gives the government both too much power and too much to try and handle. If you allow state governments to take on more of the responsibilities for the economics of their own states then there will be a smaller, more controllable system. The larger government gets, the more complicated and flawed the system will be.

" A robust society would have more public goods delivered by the voluntary associations that Tocqueville noticed were characteristic of America."

I wish I could agree, but this is a counterfactual statement in view of the long history of failures to solve public goods problems through private means. A review of the early history of the U.S. is enough alone to underscore the problems with coordinating private solutions to public problems. The ultimate issue is envy: the people with talent who work hard don't want to share with those less fortunate. The reason we have government is to solve this problem by giving these talented and hard working people a neutral-as-possible mechanism for keeping a rising tide of social problems at bay -- taxation and the provision of public goods.

And that's without getting into the difficulties of governing a confederation of local governments in the absence of a overarching central authority. Just once I would like to see a minarchist take seriously the problems faced by the colonies under the Articles. How or why was the Constitution an underoptimal solution to those problems?

Troy Camplin writes:

I know of no such history to support your contention. There is a long history of people with political power acting that way, but not of private citizens -- especially the rich.

Comments for this entry have been closed
Return to top