Good Market, Bad Market

Whenever I hear a politician or other talking head extoll the virtues of the free market, I privately run though a little checklist. Certainly, the marketplace can be a very effective way to run sectors - if they are suited to being run by the market. If they are not so suited, privatisation can be a very grave (not to mention costly) mistake. My private little checklist of absolute minimum requirements goes like this:

Money Moves The World
One often overlooked requirement for successful privatisation is that it must actually be profitable to run the service being privatised. This is surprisingly often forgotten or ignored, with the predictable result that private companies go cherry-picking among the services privatised.

The most prominent example is private health insurances. If the insurance companies are expected to insure everyone against everything, then the insurances will either be hideously and prohibitively expensive or clearly not profitable. Thus the usual result when health insurance is privatised is that companies pick the healthiest patients and then screw the rest of the public over.

Multiple Suppliers
The market has to be big enough (and profitable enough) that every consumer has access to more than one supplier. This really ought to be a no-brainer, given that the entire logic behind the marketplace is that competition makes strong, but you'd be surprised how often politicians and free-market enthusiasts forget this little fact.

If the market cannot be covered by multiple providers in such a way that every consumer is covered by at least two providers, all you achieve is to replace the publicly owned (non-profit) monopoly with a privately owned (for-profit) monopoly.

Direct Contact
There must be direct contact between the provider and the end-user. The reasoning behind this requirement is somewhat more involved, so bear with me for a while.

The 'invisible hand' of the marketplace is, I believe, very much a statistical phenomenon.

The rationale behind the 'invisible hand' is that the customers know what they want better than the government. So if you have two companies producing widgits (A and B), and if company A is better than company B, then the consumers will prefer company A over company B.

Say that there is a 75 % chance that any given customer chooses company A and a 25 % chance that he chooses company B. Then company A gets 75 % of the market, while company B gets 25 % of the market, right?

Yes and no. On average, that is true, but in general, you get a binominal (or, in the case of multiple companies, a multinominal) distribution. This is a well-behaved distribution that has been studied extensively. If we have a x companies (A_1 through A_x) and there is a probability of p_1 through p_x (the sum of these probabilities must of course equal one) that the consumer will chose company A_i. Then company A_i will - on average - get p_i * 100 % of the market. But - and this is the crucial point - the width of the distribution (relative to the average market share) is sqrt((1-p_i)/(N*p_i)).

Now that we're done with the math, let's look at what our results actually say.

Basically they say that the larger the number of consumers, the narrower the distribution will be. So if there are - say - five million consumers, then company A from our example will most assuredly end up with 75 % of the market. But what if the consumers buy widgits through five thousand wholesalers? Then it's the wholesalers who enter the distribution as the number of customers. That's not catastrophic, the distribution will still be fairly narrow, and the market will behave properly.

But what about the situation in which the public outsources services, such as road building and maintainence, or train services? In that case you have only one effective customer, and the with of the distribution blows up in your face. The 'invisible hand' becomes a drunken sailor on a random walk.

Long story, short point. The market only works properly if the consumers have direct contact with the suppliers.

Market Manipulation
Even with the above three criteria fulfilled, it is vitally important to consider whether the market can be easily manipulated by suppliers. While the risk of cartels is ever-present on the marketplace, there are many ways short of outright cartels to manipulate markets and over-bill consumers.

One of the starkest illustrations of market manipulation in recent years is the California Energy Crisis, in which the deregulation of the Californian energy market set the stage for corporations to drive electricity prices sky-high. The basic problem with the deregulation effort was that consumers paid the spot-market price. This meant that suppliers could shut down part of their generating capacity to drive prices up.

That trick worked mainly because the price of electricity goes up sharply when supply is inadequate, but even when that particular trick is impossible, legislature must certainly look out for other, similiar tricks.

The Spice Must Flow
The last criterion on my list is that strategic resources, industries and services should not be privatised, or at least should be subjected to tight regulation.

The oil and armament sectors are two areas that should be either kept in the public domain or subject to heavy regulations. The policing and defence of the country and society should definitely also be a government job. Arguably the entire energy sector should be owned by the public, but one part that should definitely be under public control is the enrichment of uranium for use in power plants.

Railroads and Taxis
So, what does all this mean in practice. In some of the points on my list, I pointed out sectors that should definitely not be privatised, but there are many others. Take, for instance, the difference between railroad services and taxi services.

The former should most definitely remain on public hands - and in all those countries where it has been tried, it has failed more or less spectacularily. Taxis, on the other hand, is a service that we'd never even consider a public job. So what's the difference between railroads and taxis? Let us look at the taxis first.

- Private taxi companies can easily and profitably cover the entire city in which they operate.
- It is not significantly more expensive to run a taxi company that covers an entire city than to run a taxi company that covers half a city.
- There is direct contact between the taxi companies and individual consumers, because the taxis move around on public roads.
- If a taxi company turns down its supply of taxis, prices won't go up enough to compensate the company for the profits lost by docking part of their car fleet.
- The city does not fall apart if the taxis stop moving.

With railroads, on the other hand, it will never be profitable for several companies to construct and maintain their own rails covering the same area.

A way around this would be to let the railroads themselves be run by the public, but let the trains be run by private companies, like taxis on public roads, but there are several constraints imposed by the way the rail works (evasive manouvers, for instance, are impossible), and these constraints mean that traffic must be carefully managed, which cuts out the direct contact between supplier and customer.

Quite apart from these considerations, it is rarely profitable to run commuter railroad services unless they have an effective monopoly, in which case privatisation is simply a way to transfer citizens' money from the citizens' pocketbooks to corporate pocketbooks.

And, lastly, if the rail breaks down, those commuters that depend heavily on it will be left high and dry, with extremely expensive delays as an inevitable result. Thus, if a company neglects maintainence so comprehensively that the system eventually breaks down, the cost to society far exceeds the cost to the company, which may have been palmed off to unsuspecting new shareholders anyway (Enron, anyone?).