According to the Merriam-Webster dictionary, an infrastructure is “the underlying foundation or basic framework” of a system, an organization or an entire country. The public discourse, though, is usually focused on a narrower definition, encompassing large, physical infrastructures such as roads, railways, energy and telecommunications networks, etc.. From an economic perspective, most of these assets are classified as natural monopolies, i.e. they cannot be easily replicated for physical, technological or economic reasons. Like any other monopoly, an unregulated natural monopoly shall be under-produced and over-priced. Promoting adequate investments in the building and maintenance of infrastructures, while preventing monopoly rents from being extracted, is a socially desirable goal, because infrastructures are often associated with increased productivity, improved mobility of the factors of production, and ultimately economic growth.
While infrastructures have some (or all) of the feature of a natural monopoly, they may compete against each other. For example, over certain distances roads may be alternative to railways; over longer miles, railways may become an alternative to airplanes. Decisions concerning which infrastructures should be built affect the behavior of individuals, the return on investments located in the sites that do (and do not) benefit from the infrastructure, and ultimately the growth potentials of entire territories.
This raises a major issue: How to decide which infrastructures should be financed, and how? What kind of infrastructures should be prioritized, and based on what criteria? What objectives can and should be pursued by increasing an area’s infrastructural dotation? One obvious answer is: the most efficient infrastructures will pay for themselves, so capitals will flow towards the most promising projects. This is, in fact, a reasonable answer to our questions, but before it is accepted as the final one, it should be understood that it only fits under a few proviso.
First of all, it is not always easy to tell in advance whether an infrastructure will pay for itself. Secondly, some infrastructures do pay off, but over a long period of time. Third, the longer the time to pay off, the higher risks and uncertainties, with particular regard to the emergence of potential technological alternatives. For example, investors in natural gas grids used to be very confident, but eventually electrification emerged as a main driver of decarbonization raising the chances that investments in natural gas infrastructures become stranded. Fourth, the economic and financial equilibrium of an infrastructure may be – and often is – strictly dependent on the regulatory framework: the longer the timeframe, the higher the risks of sudden, and possibly disruptive, regulatory or political changes. Last, but not least, some infrastructures may not pay for themselves, but still they may be instrumental to achieve other goals, such as environmental protection, social cohesion, etc. In this case, for the infrastructure to be deployed, governments need to design some subsidy scheme, which would again be subject to regulatory or political risks.
In order to take all these dimensions into proper account, cost-benefit analyses (CBAs) have been proposed as the cornerstone of the evaluation, and prioritization, of the proposed infrastructural projects. CBAs may be extremely useful to answer three questions: a) when several alternatives are available for the same infrastructure (say, alternative routes between point A and point B), which one should be picked? b) when resources are scarce and alternative projects are available (say, a road from A to B or from C to D), which one should be prioritized? c) how to choose between different infrastructures that would partly offset each other (say, a natural gas pipeline to supply a place A or a power line to provide electric heating?).
However useful, though, CBAs suffer from severe limitations. To begin with, the technical life of an infrastructure is of several decades. Uncertainty regarding the future economic growth as well as technical progress may make the infrastructural obsolete or even stranded in the long run, or – at the very least – the scenarios regarding its usage may turn out to be grossly over- or under-estimated. Moreover, infrastructures are often intended to achieve economic as well as non-economic goals, such as improving the environmental quality or social cohesion. In order to account for them, assumptions need to be made regarding the positive or negative externalities stemming from the infrastructure itself, but the evaluation of external costs may change over time as scientific knowledge or social preferences evolve. Finally, infrastructures often respond to political motives, such as the kind of evolution one envisages for a society. Hence they are expected to generate wider effects – say, concerning the people’s way of life or the productive specialization of a region – that can hardly be captured by CBAs, especially when they are supposed to come over time.
Therefore, CBAs are a powerful tool to produce informed decisions regarding infrastructural investments, but by no means they may be the sole (or even the main) criterion. In fact, they should be viewed as an instrument to account for uncertainties under alternative scenarios. They may also serve a political purpose: by estimating the potential economic outcomes of an infrastructure, a well-done CBA may attach a price tag to the long-run political goals of its proponents. Suppose a CBA results in a negative net present value of one billion euro over 10 years for a project, which is intended to guarantee that marginal areas remain connected with the rest of the country and do not fall apart. Is one billion euro a just price to prevent inequality from raising or a region to enter demographic decline? A CBA may help to frame the question in the right way, but only politics can provide an answer.