As President Biden continues to promote the administration’s historic proposed $2 trillion-plus American Jobs Plan (aka “the infrastructure bill”), there is significant debate as to what is and is not infrastructure. I thought I would weigh in to offer a definition: infrastructure is a set of physical things a society deems as necessary to enable that society’s economic and social activity at a way of life that the society desires. By my definition there are three important elements to infrastructure: 1) it is physical, 2) it is deemed necessary by a society in its time and place (and not always a “public good”), and 3) it is primarily an intermediary to other activity, and not an end unto itself. Because of (2), what ends up being or not being infrastructure is always socially construed.
Etymologically, “infra” means “below,” or “underneath” the structure, so in a sense the word implies something important. Infrastructure is rooted as a word as being underneath the built environment; it supports it. It is not the “structure” itself but things beneath it. Infrastructure is not an end unto itself, and its main purpose is to facilitate other economic and social activity.
By my definition, transportation meets this criteria (railroads, transit networks, buses, ports, highways, bridges, bike paths, etc.). I’d easily add potable water and wastewater facilities; power generation and transmission; and information such as telephone wires, airwaves, and now fiber and wifi. I’d be more open to many other items, but I want to focus on a conceptual definition rather than a list. Most importantly I want to stress that just because something isn’t infrastructure doesn’t mean it isn’t worthy. We need to stop making “infrastructure” synonymous with “the set of all things we like and want to do.” It’s OK if your thing doesn’t make the cut, and we can be comfortable with that.
1. Infrastructure is physical
While I believe the idea of “social infrastructure” to be valid, I tend to consider that more of a metaphor than part of what we’ve traditionally considered infrastructure. For public policy, I see infrastructure as physical things — as physical capital stock— for which bonds are issued, which require labor to construct, that depreciate, that require operations and maintenance, and which have imprints on our built environment.
The intangible world of necessary systems for a society may not rise to this definition because, if it does, then everything the state provides for or that we consider important is “infrastructure.” Clearly we need systems for the rule of law; we have norms of reciprocity and enforce social standards; we support family structures and social networks around hobbies. Many agree (myself included) that policies and outlays for a social safety net are necessary, but I won’t elevate that to infrastructure lest it lose all meaning. If everything is infrastructure, nothing is, and I want to defend its definition as being part of the physical things we build.
Of course, there are physical things that we view as necessary, but many of those are ends unto themselves and therefore not infrastructure. Schools are built as ends unto themselves, such that you believe education is an end unto itself. Networks of grocery stores are surely physical things and necessary to modern societies, but the point of a grocery store provides an an end unto itself: you go there because, for the most part, that’s where you have to go to get most of your food. Later I’ll expand on the “intermediary” nature of infrastructure, but hang with me for now.
2. Infrastructure is necessary, as deemed by a given society
We are currently caught up in a heated “infrastructure” debate because there is a misperception that “infrastructure” is some natural law with an inherent, cosmic, objective definition. There is no such operating definition, and infrastructure is inherently a social construct and reflective of a society and place.
We used to need horse facilities and hitching posts as part of infrastructure in cities and towns, but we no longer do. Ditto for telegraphs. Japan collectively considers its high-speed rail Shinkansen as critical for its society and economy, and provides for it. The U.S. has not come to that same consensus. Amsterdam surely considers its bike lanes essential to its society; many American cities, clearly, do not. Broadband is “infrastructure” to many of us now, but it wasn’t imaginable in 1850.
Some societies may even reject what we in the U.S. would broadly call infrastructure. We always talk about highways, bit I’d say they are not even absolutely necessary for humans to function if we want to dig back to the idea that we are capable of existing as a species in pre-modern conditions. Anti-technological societies exist, and they function as they choose. So defining something as “necessary” infrastructure is always a subjective choice, political even.
Every society in its time and place creates a definition of what physical things it thinks are necessary, and that becomes what they call “infrastructure.” The reason the present debate is so contentious is because many feel that once something is elevated to “infrastructure” then there is automatic justification for state provision. It has become a silly rhetorical trick (in my opinion worthy of ridicule) to consider anything you like infrastructure because, presumably, therefore it must be funded. It has become a bit of an annoying political branding game.
An important corollary I offer is that the word “infrastructure” is not synonymous with “public good” in the economic sense: something that is non-rivalrous and non-excludable. Conflating the two has corroded popular sentiment around infrastructure’s definition because we have so often associated infrastructure with public goods and therefore public intervention.
But let’s recall that the private sector is capable of, and often does, provide some items we consider infrastructure. Railroads were very important 19th century infrastructure, and in the U.S. it was private enterprise that provided for freight and passenger rail (of course with generous land grants and federal and state regulations). We’ve generally arrived at our present provision of internet services via the private sector, whether or not we think that is a good thing.
The question of privatization or public provision of infrastructure, however, is wholly political and normative. Ought we privatize airports? Should we privatize ports of trade? Should the state offer public broadband? I offer this to suggest that we need to decouple our idea of infrastructure as automatically meaning “public good.” To define something as necessary is a function of a sociological, moral, and political process. It is just a coincidence that in the U.S. we’ve had somewhat bipartisan general consensus on many of these items for the past few decades, even though that consensus is not as durable as we think.
3. Infrastructure is primarily an intermediary
We drive on roads all the time, and few of us (for the most part) drive on roads for the sake of driving around. We drive to get somewhere: to work, to school, to parties, and so on. To the extent that roads are infrastructure it is because roads facilitate our social and economic lives. The intermediary nature of infrastructure is critical in its definition because when we debate what to provide and not provide, we often want to focus on the substrate of or social and economic world.
To roads we now consider broadband as elevated to infrastructure. I don’t know about you, but I don’t tend to use my high-speed internet to try and watch zeros and ones fly by. What my internet service does is enable activity: being entertained on Netflix, posting on Twitter, Zoom calls with work colleagues, and so on. In this sense it is a intermediary service built on physical backbone of undersea cables, satellites, fiberoptic, copper, and coaxial cables.
It is fine to focus on other ends themselves, and there are many, but let’s not call them infrastructure. An example: Public parks are both physical and deemed “necessary” by a broad consensus of people. They also happen to be a classic example of a public good, and most people have an expectation that their government provides for parks. However, I argue that public parks are not infrastructure because they are mostly an end unto themselves. We go to them to enjoy public space, beauty, engage in recreation and enhance our health and social lives. But a public park is not part of the substrate of society, it is part of the thing itself. Let me add to this (and drop a grenade as I leave the room): housing. Its not “infra,” it is the structure.
If its not “infrastructure” that’s OK!
By offering a definition of infrastructure, and creating these boundaries, I’m clearly excluding many things that you, dear reader, may like. That is fine. Be comfortable. Defining infrastructure doesn’t mean that the thing you’re talking about isn’t worthy of attention, or worthy of funding, or worthy of moral consideration.
In fact, the vast majority of the federal, state and local dollars are not spent on what I would consider infrastructure, and there is absolutely nothing wrong with that. We should be comfortable with the idea that something can both be worthy of political attention and funding, but not have to play silly games to add it to the list of infrastructure.