In all of the various debates around the systematic assault on the US Postal Service (USPS) of late by the Trump administration, one statement crops up now and again. It takes various forms, but boils down to something like this:
The post office is great and all, but shouldn't it be self-sustaining? Things that aren't self-sustaining in the free market don't last, so if something can't be self-sustaining then it probably should get replaced by something that is (like a private company).
There's a whole lot to unpack here, so I will try to be brief, but knowing me I will fail miserably.
The limits free-market
I am far from the first person to observe that the "Free Market(tm)" is the American religion. And like most religions, its most fervent adherents preach one thing and practice another, ignoring the difference. In practice, of course, no western country is truly a free market, not (just) because of government involvement but because truly free markets do not exist naturally and tend to degrade when they do exist, thanks to the win-more dynamic of wealth/power acquisition.
Those who argue that all aspects of society should be governed by the free market, without exception, are religious fundamentalists, pure and simple. And just like any religious fundamentalist, arguing that "My Religious Text(tm) should be the one twue law of everything" is short-sighted and dangerous.
That's not to say the free market is intrinsically a bad thing; it just needs to be understood and leveraged in a non-religious way. Specifically, it is an abstract model of how people behave under certain conditions. The problem is those conditions are not universal, and in practice not everyone acts that way even if in aggregate most people do. Those conditions are things like:
- Very low barrier of entry in all industries
- Fully-informed consumers
- Perfectly elastic transfer of resources from one endeavor to another
- Competing goods are perfect substitutes
- Involved parties are acting in good faith
- Involved parties are considering future impacts of an exchange, not just immediate impact
and so on. Even a cursory look around will tell you that those are rarely entirely true, and are often extremely not the case. That makes the model break, and need extra epicycles to adjust and adapt.
Additionally, bear in mind what supply-and-demand (the core pillar of a free market model) actually optimizes for. It doesn't optimize for social value, or social benefit, or personal benefit, or long-term efficient use of resources. It optimizes for price. Specifically, it optimizes for an equilibrium between price and demand, such that there is no surplus demand or supply. That's it. It doesn't even tend toward an optimal equilibrium. It tends toward a stable equilibrium, which may or may not be optimal.
(That is, the supply and demand curves are not straight lines. In practice they're wavy and flip around a lot, such that you can find plenty of places where a given intersection point is stable, because it's at the bottom of a trough, but not optimal, because there are other intersection points that have a lower price point. They won't be reached, though, because the current equilibrium is stable.)
Of note: A real free market does not optimize for profit. In a truly elastic, pure free market, net profit for a firm will always be zero. If it raised its profit margin even slightly, that would raise its price and consumers would immediately switch to some other supplier, forcing the price back into equilibrium. That tells us two things: One, since profit clearly does happen there is no truly free market because the criteria for it do not exist; Two, those who model the world through profit maximization are not free-market capitalists. They're just mercantilists.
When that isn't what you want
A free market model is all well and good and very useful, when what you want to optimize for is minimizing surplus supply or surplus demand. Often that is what you want, and the assumptions above are close enough to correct that you can mostly fudge it and let the free market do its thing.
But there are numerous situations where optimizing for minimal surplus is not, and should not be, the goal.
That is particularly the case for "infrastructure," which has very different characteristics from most abstract goods-and-services. Infrastructure refers to something that is itself not a particularly valuable good or service, but it enables others to build other, more valuable goods and services on top of them. (From one point of view, therefore, one person's infrastructure is another person's goods, at least sometimes.)
A social infrastructure is those cases of infrastructure that support the society; not the economy, but the people in the society directly. It's a form of non-financial support. The free market is, in practice, absolutely horrible at figuring out how to optimize social infrastructure.
Social infrastructure
What qualifies as social infrastructure is somewhat fuzzy and subjective, and people can, and frequently do, have a good faith disagreement about what should or shouldn't be considered social infrastructure. As an example, the following are all usually understood as social infrastructure today (modulo the hard-core anarchist crowd):
- Public education. It is to the social benefit that as many people as possible are literate, capable of critical thinking, and well-informed about both history and the world today.
- Clean water. Without ready access to clean and safe water, most of modern society is simply not possible. It's a problem space that people simply should not have to think about on a day to day basis.
- Roads. Whether you personally have a vehicle or not, the presence of safe and well-maintained ground transportation is a net benefit to society as a whole.
In each of these cases, what you want to optimize for is not "finding a price point at which you minimize excess supply or demand." You want to optimize for "maximum availability." That is, net profit is not the goal: Ensuring that everyone has access to that infrastructure is the goal, whether or not that is profitable.
The free market simply doesn't do that. It can't. That's not what it's designed for, and that's OK. There are other models besides the free market that can be applied to get to the optimal state for social infrastructure.
The problem is two-fold: One, agreeing on what is and is not social infrastructure; Two, the fundamentalists who believe that the Free Market(tm) is the One Twue System(tm) to solve all problems, and therefore any optimization other than the one the free market shoots for is immoral. You can probably tell my opinion of the second stance; I consider it fundamentally morally bankrupt and toxic, and its adherents the same.
But that still leaves the entirely valid question of how to determine if something is social infrastructure or not, and what the alternate resourcing mechanism is.
Questionable social infrastructure
Most of the debate in the US around "Medicare for all" or other forms of socialized health care fundamentally boil down to this point: Is access to modern medicine a social infrastructure that should be handled through social infrastructure models, or a conventional good-and-service that should be handled by the free market? Every major industrialized nation in the world has decided the answer to that question is "Yes" except for the United States, and I would argue the relative quality and cost of healthcare in different countries makes it blatantly obvious that the US has the wrong answer here, but that's what the debate is.
In the US, electricity was determined nearly a century ago to be social infrastructure. The Rural Electrification Act of 1936 set out to ensure electricity was available to rural areas where it was not financially viable for private companies (the free market) to do so. The net result was expanding farm electrical access from 3% to over 90% over the next 20 years. Whether or not money could be made doing so (whether a satisfactory equilibrium price could be found) wasn't the question.
Telephone access followed similar regulation later, because it was determined that telephone access had become a type of social infrastructure.
Any industry where there is very high up-front costs but relatively low ongoing costs (and thus a very high barrier to entry against an incumbent) is a potential candidate for social infrastructure, although by no means should all such industries be considered social infrastructure.
Public transportation used to be viewed as social infrastructure in most cities a century ago, until it was deliberately undermined and sabotaged by the private automobile industry. (The film "Taken for a Ride" is an excellent and horrifying look at that sordid tale.)
The least debatable social infrastructure in the US is... the Post Office. It's right there in Article I, Section 8, Clause 7 of the Constitution: "The Congress shall have Power... To establish Post Offices and post Roads;" Even from the beginning of the country, it was recognized that mail delivery, a critical form of communication and commerce both then and now, was a social infrastructure that needed to just damned well exist, regardless of whether it was profitable or not. The post office today delivers to every house in America, even the ones that do not cover the cost of stamps for the messages being delivered. There are thousands of addresses for which USPS delivers mail that are a net financial loss for it, and that's OK. Profit is not the goal, finding an optimal equilibrium price is not the goal, minimizing waste isn't the goal; the goal is "ensure that absolutely everyone has access to the postal service, period."
That is a social benefit, a social infrastructure for which the rules of the free market simply do not apply. FedEx and UPS frequently hand-off final delivery of packages to USPS in areas where it would cost them more in gas to drive (or take a boat) to a location, but the USPS does it anyway because that's its job.
Ensuring social infrastructure
Once something has been determined to be a social infrastructure, there are many ways of going about maximizing access to it. Different models make sense in different cases, and have been used in different times in different places. Often hybrids get used, but here are some of the most common archetypical models.
One model is "the government just does it." You pay taxes to whatever level of government is going to provide the service, and government employees carry out that service. Local schools, fire, and police service, and roads usually (but not always) fall into this model. This is also the model used by the military (the presence of which is, also, a social infrastructure for mutual defense).
Another is "the government just does it, but also charges for it." In this case, a government agency provides the service but charges some low fee for it, with exceptions for those who are otherwise unable to pay. Water service often uses this model, as does the USPS itself. The reason for the fee can vary, from reducing tax burden to ensuring a higher cost to those who use the most.
A third model is "the government grants a private company a semi-monopoly on a given service, provided that they ensure everyone gets it." That's often how electricity works today, and landline telephone service. Sometimes these private companies are non-profits with a government charter, other times they're ordinary for-profit corporations.
None of these models is necessarily right in every situation, or wrong in every situation. I would argue that the 3rd is the least effective and most subject to abuse, but all of them have potential pitfalls and avenues for abuse. Which... is also true of the free market, which has its own long list of pitfalls and innumerable avenues for abuse. There is no silver bullet.
In numerous cases, a free market alternative coexists alongside a social infrastructure service. There's nothing inherently wrong with that in theory, but in practice it always seems to result in the private firms working to undermine and destroy the public structure so that they have one less competitor, even though it means access to the service is then reduced.
Social infrastructure matters
Which model is ideal for which case is a question I am not going to get into here. That's an entirely separate debate. We could also debate which services ought to be viewed as social infrastructure and which should not. I am not getting into that here, either.
The point at hand is that social infrastructure matters. It's a social need. Not every market can or should be optimized for minimizing surplus supply or demand. Not every market can or should be optimized for maximizing profit.
Some markets should be optimized for maximizing access and participation, even if, and especially if, doing so is not profitable or an efficient use of resources. In those cases, all of your free market intuition and training is simply not applicable. You need to instead think about "how do we ensure that everyone has access to good or service X, because that matters, and how do we ensure that most efficiently?" (Efficiency always matters, but not always the efficiency you think.)
An economics education that doesn't include social infrastructure, and the various complementary models to manage a market other than a strictly free market, is incomplete. If your economics education focused solely on free market models, you should ask for your money back because it ignored critical parts of the picture and did you harm as a result.
But in the meantime, please don't pass that harm along to others through a lack of understanding of social infrastructure and social benefit.