It's not as simple as `let them fail'

I've seen some progressive and mainstream voices denouncing the `too big to fail' Fed bailout of Bear Stearns a week ago. I'm much more ambivalent about this sort of thing.

The financial sector of the American economy functions much like a private toll road system. Some roads -- like the one from your savings to your neighbor's new mortgage or small business loan -- are small, low-volume, and highly local. Others, like Bear Stearns, are superhighways, moving tens of billions of dollars back and forth all over the place in both financial and physical space. Bad loans, in this metaphor, are like cars with severe mechanical problems -- no breaks, bad transmissions, maybe a couple cylinders aren't firing. And when lots of them break down in the middle of the superhighway, it gets clogged and slows down. The dead cars have to be cleared out, which adds to the operating expenses of the superhighway.

Now, the superhighway can, as it were, only allow cars that pass mechanical checks to use it. This keeps everything moving along nicely, but it's sacrificing some profit. More risk-friendly superhighway operators aren't going to have such checks, betting that the additional costs from cleaning up after the failures of bad cars will be less than the additional profit from having a higher number of cars move through their system.

Bear Stearns failed, essentially, because it lost this bet. As traffic ground to a halt on the Bear Sterns superhighway due to all the broken cars littering the roadway, the other superhighways closed their connections to keep the backup from spreading to their systems. (Imagine you have an east-west superhighway passed underneath a north-south highway, with a complex series of ramps that allow you to move from one to the other. Now imagine there are barriers that can be lowered to prevent cars from using those ramps. That's essentially what the other giant securities banks did to Bear Stearns late last week.)

This is good, at least in the short term, for the other superhighways, because it keeps them from grounding to a halt along with Bear Stearns. But it's bad, in the medium term, for the economy as a whole, for exactly the same reason suddenly shutting down one of the major superhighways through the middle of a large city is bad. Money/cars still need to move around. Either traffic will be diverted onto other superhighways, which will likely push them beyond their efficient capacity and slow them, and hence the system as a whole, down. Or people will, as it were, just stay home, and not bother trying to fight the nightmare traffic. Money won't move from the people who have it to the people who want it to do things, so people do less things -- don't open new businesses, don't expand existing business, don't build new houses, etc. In either case, economic growth slows down, and there's a potential for recession.

At least, that's the `too big to fail' theory. Closing fifty feet of surface street in the middle of a residential neighbourhood probably isn't going to do much damage -- the failure of small economic players, at least individually, might be inconvenient, but the system will adapt pretty easily. Shutting down a massive superhighway that carries a significant percentage of all the city's traffic will have complex effects that will ripple through the traffic network as a whole, and possibly even damage it severely -- letting Bear Stearns fail could have made the recession even worse.

So, did the Fed do the right thing, in essentially paying another superhighway operator to clean up the mess Bear Stearns made of its system by letting on all the junk cars? Well, they acted to prevent a medium term disaster. The clogged superhighway will be cleaned up. But the incentive to create such clogs will not have been removed. If superhighway operators knew that they would be the primary losers when a massive traffic jam developed on their superhighway, they would be likely to sacrifice some profits and allow only cars that passed mechanical checks onto their part of the system. But they don't know this. They know the exact opposite: that they can create a huge mess, and the Federal government will buy them out. According to a purely self-interested conception of reason, they have no good reason not to just do the same thing again.

It's this last point that critics of `too big to fail' are trying to get at. Or, at least, I think it's the best related point. With a prevalent short term profits mentality, a pure free market system will cease to function efficiently as these sorts of disastrous decisions are made again and again. A more virtuous business culture will help, but the temptation to tempt fate might still be too large. (Remember, Bear Stearns was enormously successful until it finally imploded. Even without the bailout, the former CEO could have walked away from the mess he created a ridiculously wealthy man.) And stopgap, emergency intervention by the government is just that -- it might avert the most looming disasters, but it completely misses the underlying cause of the problem. What the system needs, I think, is vigorous regulation -- the kind that pretty much made these disasters impossible in the '50s, '60s, and '70s.

Does the welfare state hurt us all?

Free marketers like to tell a story that is prima facie plausible, and suggests that life under a system of the redistribution of wealth is worse for everyone, even the poor, than life under a free-market system. The story goes something like this: The high tax rate needed to appropriate wealth from the people who produce it takes away the incentive to produce more wealth. So, under a distributive system, less wealth is produced overall, and since there's less to go around, the people who would have benefited from the redistributive system are actually worse off.

There are obviously some huge gaps in the reasoning there, and I've some suppressed some very, very important ceteris paribus (`all other things being equal') clauses. Ignore those for the moment. The story revolves around the claim that, under a distributive system, less wealth is produced overall. This is an empirical claim. Is it true? Here are two graphs from the US Department of Commerce that suggest not.

These graphs show the percentage change in the US GDP over the past several decades, including during the rise and fall of the Roosevelt-Johnson redistribution system. GDP is a measure of the size of the economy, and is one rough way of measuring how much wealth the country has. The percentage change in GDP over a given period of time shows how much the economy grew -- how much total wealth increased -- over that period of time. If the free marketer's claim is right, then ceteris paribus, percentage change in GDP should be less under a more radical restributive system (with more regulation, higher marginal tax rates, etc.) and greater under a system closer to the free market ideal.

 

 

And the exact opposite is the case. During the height of the Roosevelt-Johnson system, the US economy actually grew a little bit faster than after Reagan started to dismantle it. The '70s, with the highest marginal tax rates the strictest regulation, actually saw the highest economic growth rates of the past half-century. Since Reagan started lowering tax rates and removing regulation, the economy has seen relatively mild growth.

This isn't a Quick Refutation, since I haven't shown that the crucial ceteris paribus clause is satisfied. And it seems quite likely that it's not -- there have been wars and changes in energy costs and all sorts of things that can make a difference. But this does place a burden on free marketers: if regulation and taxation are supposed to be bad for the economy as a whole, then why did our economy do better when it was more highly regulated and highly taxed?

Economics is tricky II

Okay, let's talk about externalities. In a nutshell, an externality is something -- good or bad -- that you don't pay for. For good reason, the examples that always seem to get trotted out are instances of pollution. For example, without the EPA around to correct things, a factory that pollutes a river generally doesn't have to pay to clean up the river. Either the state, and hence the taxpayers, do, or no-one does, and the people living downstream `pay' for the pollution by getting cancer and dying.

The contemporary environmentalist's favourite examples of externalities are `cheap' fossil fuels and `cheap' beef. Both are `cheap' because the cost -- as in, dollars the oil baron/cattle baron pays -- to procure these goods (in the economist's sense, not the ethicist's sense) is much lower than the selling price -- what people will pay to consume them. But, in the quantities which we consume and produce, respectively, both fossil fuels and beef release horrendous quantities of greenhouse gases. Fossil fuels release carbon dioxide, as per Part I, and cows release methane as waste (as well as consuming even more fossil fuels, using up lots and lots of water, and so on).

And yet, because these costs are externalized, it is rational (again, in the economist's sense of maximising one's self-interest) for people who can produce fossil fuels and beef to do so. Which is where carbon taxes and cap-and-trade schemes come in. Since carbon taxes are simpler (cap-and-trade schemes amount to a market on carbon tax thresholds, and that means the analysis is an order of magnitude more complicated), I'll focus on those. From an economic point of view, the idea is quite elegant: if emitters of carbon dioxide (and other greenhouse gas compounds, eg, methane) are required to pay at least a significant fraction of the externalized costs, they will generally shift production to processes and products that emit that much less carbon dioxide. It's no longer rational for (former) carbon-emitters to emit carbon.

I'd also like to argue that carbon taxes can be justified on libertarian grounds. This seems entirely straightforward to me: for every metric ton of carbon dioxide you release, you do US$12 of harm to everyone else. (IPCC report, summary for policymakers, 24) Rather than collecting these damages individually and retroactive to the damage suffered, it makes far more sense (in terms of avoiding the legal costs of everybody suing everybody else) for the state to collect them proactively, and hence use them as a deterrent.

Economics is tricky I

Some bad news about ethanol:

Researchers led by Timothy Searchinger at Princeton University said their study showed greenhouse-gas emissions will rise with ethanol demand. U.S. farmers will use more land for fuel, forcing poorer countries to cut down rainforests and use other undeveloped land for farms, the study said.

Searchinger's team determined that corn-based ethanol almost doubles greenhouse-gas output over 30 years when considering land-use changes. Bob Dinneen, president of the Renewable Fuels Association in Washington, said the study used a flawed model and overestimated how much land will be needed.

The intuitive case for ethanol is based on the fact that the carbon you release in burning it was trapped in the corn just a few months ago.

Let's back up. The oxygen combustion of carbon fuels, no matter how efficient (ie, no matter how little particulate pollution is produced) produces carbon dioxide. Yearly averages of carbon dioxide levels are rising because we burn lots of fossil fuels -- essentially, highly compressed prehistoric swamps. This is carbon that was in the atmosphere millions of years ago, but has been out of the atmosphere since the beginning of human civilization. Hence we say that fossil fuels result in a net emission of carbon dioxide.

The intuitive case for ethanol is based on the fact that, rather releasing carbon that was trapped by plants while our shrewlike ancestors were trying to steal eggs from dinosaur nests, burning ethanol releases carbon that was trapped by corn a few months ago. So, over the scale of a couple years, ethanol results in zero net emissions of carbon dioxide.

However, you can't just magically find some unused tract of land and grow lots and lots of corn for use in producing ethanol. At least, the land that's not currently being used is in environments that are too hostile for corn to grow -- the middle of the Sahara, the tundra, etc. You have to take land that's currently in use -- namely, for growing corn or other crops for food -- and change its use -- to growing corn for ethanol.

This results in an increase in food prices -- nothing's changing the demand for food (if anything, it grows as the global population increases), and the supply's just dropped a few percentage points. And, because of globalisation, an increase in food prices in the US increases food prices in places like rural Brazil.

The people of rural Brazil are, generally speaking, already quite poor. So if food prices go up, there's now a (larger) gap between how much food they can afford to buy (it's not like they've gotten any less poor) and how much food they need to survive. It's a lot easier for them to clear-cut another X number of acres of rainforest and grow enough food to fill that gap than it is for them to scrounge together enough additional money to maintain their level of food consumption. But the X number of acres of rainforest would have absorbed a certain amount of carbon from the atmosphere. This counted against the net emission of carbon from burning all the fossil fuels (and, I might add, raising all the methane-producing cattle we do). Effectively, a mandate to switch a certain amount of US land from food production to ethanol production does actually increase net emissions of carbon.

The question is how big of an increase, and how that compares to the decrease in net emissions from burning ethanol instead of petroleum. And the answer, it appears, is enough to double the emission rate in 30 years.

Part II, on externalities, when I feel like writing it. Link from Paul Krugman.

Pretending that economics is the right way to think about problems of justice

Background: NARAL PCA wants to use text messaging to send action alerts to members who sign up to receive such alerts. Verizon, citing a longstanding internal rule prohibiting text messaging about `controversial' topics, refuses to let NARAL PCA run this program with Verizon customers. (Cont'd below the fold.)

What's to be done? One response suggests expanding something called the common carrier rule to include text messaging, now that text messaging has become a major form of interstate communication.

Professor Wu pointed to a historical analogy. In the 19th century, he said, Western Union, the telegraph company, engaged in discrimination, based on the political views of people who sought to send telegrams. “One of the eventual reactions was the common carrier rule,” Professor Wu said, which required telegraph and then phone companies to accept communications from all speakers on all topics.

But this isn't the only response.

Some scholars said such a rule was not needed for text messages because market competition was sufficient to ensure robust political debate.

“Instead of having the government get in the game of regulating who can carry what, I would get in the game of promoting as many options as possible,” said Christopher S. Yoo, a law professor at the University of Pennsylvania. “You might find text-messaging companies competing on their openness policies.”

Well, yes, you might find text-messaging companies competing on their openness policies. Just the same way you might find organic foods being served at McDonald's and Burger King. Or the way you might find $500 in your winter coat next week.

The supply and demand model (which I presume Prof. Yoo is utilising here, the Times not having quoted him citing any more sophisticated model) assumes, among other things, that the marketplace is in a state of perfect competition. Primary features of this state include perfect and complete information (everyone knows everything), equal access (it's easy for consumers to move from one producer to another, among other things), free entry (it's easy to start up a new company in the market), and the independence of consumers and producers (the fact that my best friend chose to go with company A in no way influences my choice between companies A and B).

It should be clear that none of the assumptions made in the previous sentence obtain in the case of today's telecommunications giants. A quick Google search isn't turning up a nice chart or graph, so I don't have evidence, but AT&T, Verizon, and Sprint (I might be forgetting one or two prominent others) have almost complete control over the cell phone network within the United States. Critically, this includes the infrastructure, the physical network itself. A new carrier trying to enter the market must either pay one of these companies to use their infrastructure, or invest billions of dollars (and, most likely, engage in tedious legal fights) to build their own infrastructure. The big telecoms require customers to sign two-year service contracts, so buyers cannot easily move from one service to another. And, of course, the service contracts are specifically designed to encourage customers to use the same network as their friends, family, and co-workers, with significant discounts for in-network communications.

Again, a quick Google search isn't turning up the information I need to make good on this claim, but I think the cell phone networks in the US would be better described as an oligopoly. If this is the case, then both competition and collusion are likely, and a simple model can't predict with any reliability. For example, the telecommunications giants might start competing with each other to let NARAL PCA send action alert text messages. Or they might just as well all decide to ban NARAL PCA from their networks. Critically, in neither case do consumers have any real say over what happens. The ability of NARAL PCA to effectively communicate with its members using text messages depends on decisions made at the highest levels within the big telecoms.

Which leads to the real problem I have with Prof. Yoo's suggestion that we just let the invisible hand sort things out. There's a difference between NARAL PCA communicating action alerts to a list of members and what kinds of frozen pizzas we can find at the local megamart. The Times reporter recognises that difference, albeit only in one short sentence: `Messages urging political action are generally thought to be at the heart of what the First Amendment protects.'

The real, effective ability of citizens to express their political beliefs is at issue here. It's true that there are plenty of other ways for citizens to engage in political speech -- NARAL PCA has a website and sends out messages to its members by both e- and snail mail. But if Verizon is allowed to prohibit the communication of controversial ideas over its network, then mutatis mutandis so is FedEx, and so are landline phone companies and television stations and cable television and internet companies. Hence, by contraposition, if we accept a principle requiring these other common carriers to give NARAL PCA access to their networks (whether that principle is grounded in the First Amendment or something else), then we must accept a principle requiring Verizon to give NARAL PCA access to its text messaging network.

Finally, once we highlight this aspect, we realise that Yoo's economic argument (such as I assume he had; the Times just quoted his conclusion, not his reasoning for that conclusion) is a huge red herring. When we start debating the proper economic policy, we end up completely ignoring the deeper, and far more important, issue of justice: What impositions can and should we place on members of our society to ensure that everyone has a real ability to express their political beliefs?

Chaos, the stock market, and the ends of political philosophy

I just finished teaching a class on chaos theory to a group of 15 young people (ages 12-16). The subject is fascinating, and I highly recommend picking up a copy of James Gleick's Chaos for an non-technical and historical introduction. One of the major themes we have learned in the development of chaos theory over the past 40 years is that order and disorder are not, as traditionally thought, opposites. Gigantic, destructive hurricanes -- to pick a vivid example -- are caused by essentially the same processes that produce refreshing light rains.

One of the most important and interesting features of chaotic systems is a tendency for a system to transition between order and disorder spontaneously and without any outside influence. The same processes keep happening in the same way, and predictable stability suddenly turns into completely unpredictable randomness -- and then, just as suddenly, settles down into predictable stability again. Freakish weather -- Indian summers, snow in April and May (in the US) -- is a good example of this sort of thing.

I spent all day yesterday travelling, which means most of the news I saw was business news. And all the business news was going on and on about the turmoil in the stock market. And I was reminded that chaos theory has turned out to be rather successful at explaining the behaviour of stock markets. It's not that the abstract mathematics gives a causal account of how real economic events influence the way the wealthy trade money. Rather, those causal processes, whatever they are, have a mathematical structure that is accurately expressed using chaotic dynamical systems.

This means that stock markets, like other chaotic phenomena, will tend to alternate -- unpredictably and on all time-scales -- between periods of ordered and disordered behaviour. The current turmoil on Wall Street may be the result of exactly the same processes behaving in exactly the same way as two months ago. Predatory and self-destructive lending practices may have been an incidental factor -- more like the straw that broke the camel's back than a ton of bricks -- or may have even been completely irrelevant -- the system went from order to disorder entirely naturally, with no outside influences.

But we don't think in chaotic terms, at least not generally and not yet. When we see order turning into disorder, we think some external force has intervened in the system. And so we see pundits talking about this or that possible cause, without ever considering the possibility that there was no cause, that this is just the sort of thing that happens in chaotic systems.

Now, to get to something that's actually important. I don't want to try to formulate the definitions that would make empirical investigation actually possible here, but I do want to suggest that `political stability' may also be a chaotic dynamical system.

This hypothesis would have some interesting explanatory power. It would explain, for example, why the 1960s and '70s were so tumultuous, and our era is so placid.

But, for this post, I'm really interested in thinking about what implications this hypothesis would have for political philosophy. We tend to think of the just society as at least internally stable -- it has all its problems solved, so the only way instability could arise would be for some outside event to occur, like a drought or plague or alien invasion. Some political philosophers have even considered stability as the single most important issue for political philosophy -- Hobbes, for example, defends a totalitarian state on the grounds that it's the only way to guarantee stability, and Rawls makes the transition from A theory of justice to Political liberalism by realising that an ineliminable political pluralism will be a feature of any liberal democracy, and hence a potential threat to stability that must be dealt with.

But if this chaos hypothesis is right, then a perfectly stable society is a pipe dream. We may be able to achieve stability for a while, but it will never be completely ineliminable, and indeed will crop up unpredictably and spontaneously on all time scales. Even Fall of the Roman Empire-level disorder may be inevitable. And then political philosophers may be completely wasting their time, looking for the totally stable society.

We should therefore go after justice directly. A stable society is either impossible, or will be a secondary achievement of a truly just society. Perhaps we can even develop a conception of the just society that can survive a Fall of the Roman Empire-level disorder.

Two kinds of justice

Capitalist justice: An apple farmer sells his crop to the person who's willing and able to trade them for the most amount of money.

Actual justice: A community uses an apple orchard to feed itself.

Inspired by this exchange.

Libertarianism and violence

Over at Pandagon, Chris Clark has a nice post up about the history of libertarianism, that pro-capital anarchism so popular among upper middle class teenagers of a certain intellectual and rebellious stripe. And, of course, a certain stripe of political philosopher, including a few of my own fellow grad students.

I wanted to highlight this bit of Clark's in particular:

Libertarian Cranial Detonation Technique #3: Mentioning Libertarianism’s blindspot.

That accumulation of serious political power is the end result of the Libertarian political wankdream, and yet somehow boss-based coercion escapes the Libertarian scrutiny to which municipal zoning boards and feminist bloggers with itchy banning fingers are routinely subjected.

Arguing with some libertarian colleagues a time or two, I've developed a hypothesis, which I'm neither going to defend or critique here.

The hypothesis is that libertarians are working with an excessively narrow understanding of violence. Ask a libertarian about, say, taxation, and you'll suddenly find yourself in a rhetorical whirl of gun-toting IRS agents kicking down your door. But try to draw a parallel with the threat of economic violence -- the kind, say, Wal-Mart uses to keep its employees in line -- and the libertarian will look at you as though you're speaking Martian.

For libertarians, I suspect, physical violence is real, and the state is suspect because it is the only agent authorized to use physical violence, or the threat of physical violence, as a means of coercion. (The idea that the state is an agent is also, I think, suspect, but that's far too complicated an issue for this little note.) Economic violence is not real, `because you can also go get another job', so the use of economic violence, or the threat of economic violence, is likewise not real, and hence there's no reason to be suspicious of the wealthy.