The digitalisation of infrastructures and infrastructure-based (public) services: what are the problems?
The paper ”The digitalisation of infrastructures and infrastructure-based (public) services: what are the problems?” (Finger, M., Montero, J.) will be presented at the 11th FSR Annual Conference “From Data Spaces to Data Governance” (9-10 June, 2022).
Abstract:
Infrastructure and the corresponding infrastructure-based services are currently being digitalised (Montero and Finger, 2021a). For both, this raises a so-called ‘digitalisation dilemma.’ On the one hand, they need to digitalise, as digitalisation yields huge efficiency gains for their operations and therefore for their competitiveness in the light of the liberalisation of their sectors. On the other hand, digitalisation makes them vulnerable to platformisation by big global private digital platforms, and even more rapidly if they are forced by governments to make their operational and commercial data freely available.
Such digitalisation and perhaps platformisation of infrastructure and infrastructure-based services raises a series of questions about what can be called ‘public service.’ In this short article we discuss what we consider to be the two most relevant questions, namely (1) that of what happens to infrastructure and infrastructure-based public services once they become platformised, and (2) that of public services by digital platforms. In the conclusion, we briefly discuss two recent attempts to define the ‘digital layer’ in public service terms. But before doing so, we briefly define the applicable concepts.
Definitions
Infrastructure is a vague concept that technically needs to be limited to the traditionally monopolistic physical infrastructure per se, i.e. roads, railway track, gas and district heating pipes, electricity grids (high voltage transport and distribution), water pipes, telecommunication cables (copper and fibre), air traffic control, airports, ports and railway stations. These are sometime also called “foundational infrastructure” (Foundational Economy Collective, 2018) and constitute the so-called ‘layer 1’ in an unbundled network industry (Finger, 2020). Using this infrastructure, infrastructure-based services can then be provided (layer 2), such as rail passenger and freight services (by train operating companies), air transport services (by airlines), energy services (electricity, gas, heat), telecommunication internet services (by telecommunication operators and internet service providers), postal services, water services, etc.
These types of foundational infrastructure typically have a monopoly over a given territory, even though there may be some exceptions, notably in the telecommunications sector. However, they do not necessarily have to be publicly owned. However, they are typically operated in the general economic interest. Services based on this infrastructure may or may not be public services. After the liberalisation of infrastructure in the 1990s, a process that was typically accompanied by unbundling, it became necessary to distinguish between public services and commercial services. Consequently, public services had to be explicitly defined, and their provision had to be assigned to a particular public service provider by way of a public or universal service obligation. Typically, companies providing such public or universal services would be compensated if these services could not be self-financed. Public or universal service obligations typically conform to criteria such as continuity (of service), equity, accessibility and affordability.
All the above remains valid in the era of digitalisation. However, with digitalisation a new and third layer comes into existence above foundational infrastructure (layer 1) and infrastructure-based services (layer 2) (Montero and Finger, 2021b). This third digital layer, and particularly its interactions with the original two layers, raises a series of questions, especially when it comes to public services.
What happens to infrastructure and infrastructure-based public services in the age of digitalisation?
Typically, traditional physical foundational infrastructure and infrastructure-based public services continue to be relevant and will be provided even when they are platformised. This is valid for all the above-mentioned types of infrastructure and infrastructure-based public services. However, in this article, we limit ourselves to transport. For example, a mobility-as-a-service (MaaS) platform will offer train, metro, tram and bus rides, among others. Similarly, trains will use railway infrastructure and buses will use roads. However, they will not only continue to exist. In fact, their very existence is actually a condition for the existence and success of digital service providers. Indeed, without physical delivery, the value of the platform would be nil. The same is true when it comes to the delivery of electricity, gas, heating, water, parcels, snail mail, telecommunications, government administrative services and media services. Most of these are either entirely or partially public services, and they will be provided by their service providers even though they are now intermediated by a digital platform.
When a traditional infrastructure-based service is intermediated by a digital platform – with the customers now being customers of the platform rather than of the physical service provider – two questions regarding public services arise. The first pertains to finance, more precisely to the compensation that the traditional public service provider receives from public authorities for fulfilling its public or universal service obligation. Indeed, platforms usually take a cut for their role as intermediaries which can be as high as 20% to 30%. In other words, up to 20% to 30% of the public funding to compensate a public service obligation now goes to the (private global) platform, meaning that public authorities now indirectly subsidise platforms. Such subsidisation of private platforms is not acceptable if it means that taxpayers’ money goes to private platforms without any public service in return. However, such subsidies could perhaps be acceptable if in exchange the platform does something that contributes to public services such as promoting public services on public transport rather than driving, or any other public policy objective for that matter, such as decarbonising transport.
This brings us directly to the second question arising, namely that of the role platforms – in this case MaaS platforms – play in incentivising and more generally shaping mobility behaviour. Indeed, mobility is an area that is necessarily permeated with public policy objectives, such as more or less mobility, public transport versus private mobility, road transport versus rail transport, hard mobility versus soft mobility, etc. Regardless of what the platform is programmed to optimise, it always incentivises behaviour towards some objective, which may or may not be in line with public policy, namely the public policy objective that was at the origin of the public service obligation to begin with. In other words, digital platforms intermediating, among other things, traditional physical infrastructure-based public services will necessarily incentivise and shape behaviour, be it in mobility, energy consumption or any other area. It is therefore only logical that such platforms intermediating public services should be regulated in accordance with the original (physical) public service objectives. In concrete terms this means that a platform using recommendation algorithms has the power to shape the mobility ecosystem, and therefore has to reflect these original public policy objectives. This means in practice that data sharing obligations, obligations concerning ticketing and even the very same authorisation of MaaS platforms should be conditioned on agreement to reflect (the original physical) public service objectives in their algorithms.
Are platforms intermediating infrastructure-based public services also (in part) public services?
From the above considerations, the question logically follows of whether platforms that intermediate public service offerings (such as MaaS platforms) should also be considered public services, meaning that they would be tasked with a public service obligation. The answer to this question, it seems to us, is no because a public service obligation does not seem an appropriate instrument in the digital layer (layer 3). This is even more so because platforms are profit-making and do not need subsidies, whereas public service obligations are typically tied to subsidies. Indeed, platforms are typically global private companies which, in principle, offer their services in competition. Public authorities, be they at the national or the local (city) level, may or may not authorise these platforms to operate on their territory, especially if a platform (negatively) affects a public policy and even more so if it diverts corresponding public funds to private profits (see above).
We think that there are two possibilities here, again taking MaaS as the example. The first possibility is the one discussed in the previous section. In this case, a (private and potentially global) MaaS platform is intermediating, among other things, physical transportation services that are tasked with a public service obligation and receive subsidies for doing so. However, such a MaaS platform also intermediates and aggregates on its platform other transport offerings, such as car rentals, scooters, bikes, parking, etc. In this case, as we said above, we think that regulating these platforms is the most appropriate way forward, so that their algorithms reflect the public policy objectives which gave rise to the public service obligation to begin with. The way to do this would be to condition their access to data and their right to resell tickets on them agreeing to run the algorithm according to public service and general interest objectives. If they do not accept these conditions, they will not be able to benefit from regulation facilitating access to data and ticket reselling. Still, they will have to negotiate bilaterally with each underlying physical service provider. By formalising such a ‘public and/or general interest operation’ of the algorithm, an objective difference is introduced between types of platforms, which in turn would justify why access to data and ticketing is given to some platforms but not to others.
However, there is a second possibility, which is most likely to be limited to the local, urban or metropolitan geographical level. Indeed, cities and metropolitan areas want to organise their entire urban mobility, and also their urban logistics, their urban energy consumption and other policies following very concrete public policy objectives, such as decarbonisation. A single urban mobility platform that aggregates all transport modes, be they public or private, can indeed be the most effective mechanism for a public authority (such as a metropolitan transport authority) to organise urban mobility (and also urban logistics, urban energy, etc.) on a geographically bounded territory (e.g. a metropolitan area) according to its public policy objectives. In this case, the public authority wants a platform provider which translates its public policy objectives into its algorithm(s) as its privileged so-to-speak ‘policy implementation partner.’ While it will be difficult to grant exclusive rights to such a privileged partner, it will, however, be possible to tie the rights to access data and to resell tickets (in the case of mobility) in such a partnership agreement (see the previous paragraph).
Conclusion: recent attempts to define the digital layer in public service terms
In this short article, we have tried to establish a link between foundational public interest infrastructure (layer 1) and infrastructure-based public services (layer 2) on the one hand and the digital layer (layer 3) on the other hand. Basically, we have aimed to clarify whether digital platforms intermediating physical public services could or should also be considered public services as they have public service features. Our basic conclusion is that such digital platforms are not public services per se, but that they could and in certain cases should be regulated so that their algorithms reflect public service, public interests and even more generally public policy objectives. We are aware that this is a position which has not been taken so far, neither at the academic nor at the political level.
Still, the question of public service or public interest digital platforms is clearly in the air, and it is not limited to digital mobility. For example, Christoph Busch has recently published a report which claims that digital platforms should be seen as services of general interest. And the Swiss government has just launched a public consultation on a proposed law that would establish mobility data infrastructure as a “third infrastructure next to rail and road” and that aims to create a new “government entity to build, operate and further develop this infrastructure.” Our article should therefore be seen as a contribution to this more general debate about the public interest dimensions of digital platforms.
