The Consequences of Leaving Tech to the Private Sector
Paris Marx is joined by Rosie Collington to discuss the consequences of outsourcing tech to the private sector, how it causes governments to lose important capacities to serve the public, and how the push for open government data empowered large tech firms.
Rosie Collington is a PhD candidate at the Institute for Innovation and Public Purpose at University College London. She’s also the co-author of The Big Con: How the Consulting Industry Weakens Our Businesses, Infantilizes our Governments and Warps our Economies with Mariana Mazzucato. You can follow Rosie on Twitter at @RosieCollingto.
Support the show
Venture capitalists aren’t funding critical analysis of the tech industry — that’s why the show relies on listener support.
Become a supporter on Patreon to ensure the show can keep promoting critical tech perspectives. That will also get you access to the Discord chat, a shoutout on the show, some stickers, and more!
Paris Marx: Rosie, welcome to Tech Won’t Save Us!
Rosie Collington: Hi, thank you so much for having me. I’m a huge fan of the show. I know your listeners are probably going to rinse you for bringing another super fan on. But I’ve really been listening since the first episode. I’m very grateful to be here.
PM: No way! Thank you so much. Of course, we’ve been in contact for a while. I’ve been meaning to have you on the show for ages, and you have a new book coming ou, so I figured this is probably a good time to finally have a conversation with Rosie and hopefully people will go pick up your book as well.
RC: That’s great. Thank you, I’m looking forward to it.
PM: Here’s where I want to start: when we think about data we often associated with tech companies — what these massive digital tech companies have been up to in the past couple of decades — they’re collecting a ton of it; they’re using it to power their algorithms; they’re making the profits, while we have to deal with all the downsides of their “innovation.” You explain that data is also essential to the State and for its ability to deliver for its citizens, that goes back long before the internet and modern tech companies. Can you break down a bit for us how the state has collected and used data to deliver services, just to set that up for us?
RC: Public sector data has a really long history that we can date back to the development of states, the early process of state building and nation-building in Europe and and other countries. In terms of the public sector data that we see today, if we think about what governments do, this isn’t just a committee that is introducing laws. For example, governments today across OECD countries spend, as a proportion of GDP, anywhere between 35% to 50% of GDP. It’s responsible for not just making laws, but producing things like services, regulation, administration, all of these things produce a lot of data.
Digital public assets — or what I’ve termed digital public assets or public sector data — can be defined as all registries, databases, and information that is collected, produced or held by public sector actors, and that’s available in digital format. The actors that are responsible for the production and use of these assets, at present, can include government departments, local authorities, or other bodies, such as the NHS in the UK. A lot of this data is numerical or textual, but it also could include visual audio recordings. This is the core public sector information that is then turned into data s ets, but also lots of data is produced. We can also think about it hypothetical terms of data that could be collected and commercialized, but that isn’t currently collected and commercialized. This is often, also, of huge interest to big tech firms, for example, in med-tech and other fields.
PM: You’ve explained that really well, and it’s fascinating because it’s not how we tend to think about data right now. When we think about data, it’s about what Facebook collecting on us, what Google is collecting on us — not so much what the government is doing. Though, certainly some people are concerned about that as well, and rightfully so in many cases, with how some of the data that the government can collect is used. But as you say, you referred to this as digital public assets — these public datasets that have been built up over many decades. I’m wondering how did the neoliberal turn and the more recent framing of anything having to do with digital technology, having to belong in the private sector, where only these private companies can do anything that relates to digital technology, the governments can’t be anywhere near that? How does that affect how we and our governments think about these datasets and the public data that has been collected?
RC: That’s a really good question, and we need to look back or situate this within the broader development, as you’ve just alluded to, within the broader developments in political economy and in capitalism, and how they have also affected governance and outsourcing. Perhaps, I can begin by saying a little bit about why I became interested, or how I became interested in this area. Before I came back to academia, before I went to do my master’s, I was working for a few years in medical research policy and in some of that work I was working with the British Heart Foundation, which doesn’t have an equivalent in the US, but I was working with the British Heart Foundation. That organization was developing cardiovascular AI policy that will then be used by the NHS and the government in medical research. In all of this work, it was interesting to see how there was this baseline assumption that the actors that would be governing and processing, and even responsible for collecting this data would be in the private sector, and everyone took that for granted.
When you speak to people, even within government, today, most people are surprised to learn, or many people are surprised to learn that actually for much of the 20th century, government’s IT infrastructure, and its management were all serviced in-house. This is something Mar Hicks has written about, as well. The question becomes, why do we assume that governments are not able to do this today or this is the inherent function of the private sector? I saw this happening in my work and then when I came back to academia, and I started studying political economy, I returned to this as a way of looking at developments in public-private relationships and the big assumptions that underpin them and how they emerged. I realize that IT infrastructure, public sector-digitalization, opitimizes, in many ways — not just the neoliberal turn — but actually what happened from the 1990s with the growth of, what we might call, the third way paradigms of governance. Unlike the neoliberal paradigm that came before, it actually did recognize the State as a potentially important actor in the creation of value, even defined in social terms.
Whereas the neoliberal period — or Thatcherite, Reaganite neoliberal period — defined the state in terms of not being able to create value for society, not being innovative. The State was conceptualized in governments, such as Clinton’s or the Tony Blair’s government, was supposed to exist and had an important function as the arbiter, or the shaper, the steerer of markets and of the economy. We had this expression that was popularized in a book written by two consultants, actually, and came to underpin the Clinton administration’s policies, which was the role of government is to steer, but not row. This phrase really captures what has happened in the subsequent decades of the third way — not post-neoliberal. There’s huge debates about how we use this term, but third way governments, which assumed that government had an important role, but doing stuff itself was not the function of government.
Actually, it didn’t matter where capacity for delivering the ambitions of government came from, as long as it had that capacity and was able to tick boxes and achieve these goals. It didn’t matter what the political economy of the actors that were doing that work. This is the ideational, or ideological underpinning of how the outsourcing of government IT infrastructure, and crucially, its management emerged. When I first started looking at this, I wasn’t necessarily thinking about bigger paradigms. I was really just thinking: Why isn’t government doing this stuff? This isn’t to suggest either —and perhaps there’s something we can talk about — that governments have always used government data, public sector data, or citizen data for benevolent purposes. Absolutely not, this is a field that has a very dark history and should serve as a warning for lots of different things that perhaps we can talk about in more detail. But just the fact that it’s not questioned — the fact that we assume that governments can’t do this and that the private sector is the only actor that can do this I think is quite interesting.
PM: It’s very interesting, especially when, as you mentioned, you go back a couple of decades, and you see that IT infrastructure is still owned and maintained by governments. It’s quite normal for governments to be involved in procuring and managing the technologies that they rely on. Then all of a sudden, there’s this shift where: Okay, the government shouldn’t have any involvement in this; it should all be handled by the private sector; this is how we create jobs and growth and all these terms that are very popular for politicians to throw around. When does that shift notably happen? When do you start to see countries say: This infrastructure that we have, this IT infrastructure that we’ve controlled, now we should be pushing that into the private sector; we should stop doing these things that we did in the past? When does that really start to happen?
RC: It varies in different countries, but the shift begins in or around the 1960s. Mar Hicks, for example, has talked about this as also happening in relation with the masculinization of the labor force in IT infrastructure management surfacing in the UK. Other academics have written about how the shift also evolved with the increase in the use of management consultants in government departments. That’s something that we write about in the book that I’ve written that’s going to be out. To the extent that we even saw, for example, in the United States in the 1950s, the US government introduced anti-trust laws to prevent IBM from providing infrastructure and providing advice on the infrastructure of government IT and servicing IT, because it considered that it was an issue for competition. It was only actually in the 1990s that IBM was able to provide advice to governments again on its IT infrastructure.
It’s really quite different in different countries. The UK had some really big players that were also in the government IT consulting space, as well at this time, ICL for example. They became really big because these anti-trust measures didn’t exist at this time. This was really a development that picked up pace from the 1960s. Although, the private sector has also been involved in the development of government technology since its inception, since governments began using an integrating and developing technology within bureaucracies.
PM: Absolutely, and there was a lot of public funding to ensure that they were able to do that. Through that history, it was a very important relationship that was there. In a paper that you wrote that you shared with me that got me interested in wanting to speak with you, you talk about how in Denmark, which is one of the cases that you look at, I believe they privatized their IT infrastructure in 1996. That began a process where the incentives in the organizations, or around the technology, started to shift because that was moved privately. Then the influences came with that overtime. It wasn’t an immediate change, but those things started to change. Do you want to talk to us a little bit about what happened in that case?
RC: Denmark is a really interesting case for a few different reasons. I like looking at Denmark, because it’s widely viewed as a good model of capitalism, and good form of social democratic capitalism. If we look at what’s going on in Denmark, and we interrogate Denmark and the Danish government, then we can think about — because this is a bit of an extreme example of this nice good form of capitalism — the flaws or we can see some flaws very clearly in this model. Denmark is a very interesting cas because it has adopted e-government reforms and public sector digitalization strategies very early, relative to other countries. Today, it is always ranked very highly in the public sector digitalization indexes, the DESI which is the European one. It was also one of the first countries to collect population level data on citizens within health systems.
Famously, or maybe not famously, but within this academic literature it’s been described as the epidemiologist stream, because it has population-level registries in health data across the whole 6 million person population. It is also interesting because its developments and the political economy of its public sector digitalization reflect broader trends, and perhaps are much more explicit in the Danish case as well. As you mentioned, Denmark was relatively late in privatizing its public sector data storage and servicing company, which is called Datacentralen, and it did that in 1996. Then we see from the e-government reforms that were introduced in the 1990s — particularly, in the public sector, digitalization reforms that were between 2002 to 2019. The ambitions and goals of public sector digitalization began to evolve.
At first, the government was framing this and these policy reports as being an issue of improving efficiency, improving transparency, making sure that citizens have better communication with the government — that government is better able to communicate with citizens. Increasingly, this became framed as a way of making Denmark a more friendly environment for businesses who want to do work there and decrease the administrative burdens of businesses. Then after the financial crisis, within a broader export-led growth strategy that the government adopted, digitalization, and the use of public sector data as well within that, was really framed and the policies around were structured around boosting growth in the private sector and in the digital technology sector. We really see through this history, or through this lens, of Danish public sector digitalization transformations in governance and thinking about the role of the public sector in the private sector, not just within digital technologies, although that’s the focus of the paper, but also more widely in the economy.
PM: It’s absolutely fascinating, that evolution that you talked about. I want to drill down on it a little bit further because I think it’s really illustrative of how these things slowly evolve. As you say, in the beginning, the idea with the digitalization bringing in new processes is around: We’re going to make this more efficient; we can better deliver services to you the public, because we are adopting these new technologies. Then post-2008 or 2009, when the economy has been hit, but also in this moment where the tech industry is positioned as: This is how we are going to grow the economy into the future, and all of us, various countries — Denmark being one of them, but countries around the world — we all need to find our way to develop our domestic tech economies to encourage companies to grow so that then they can export their tech products. Then that becomes part of the goal, even when it comes to public services. How can we use our public services in a way that is going to encourage this growth, technological development, the growth of this tech capitalism? Can you talk to us a little bit more about that and the ways that this transformation plays out, but also the impacts that come of having the shift in the mindset around how these technologies are going to be used and what the goals of using them actually are.
RC: There are a few different dimensions to this. On the one hand, we have this period that I was looking at in this paper from 2002 to 2019, also being the period, as you’ve just alluded to, when we saw the explosion of Silicon Valley as the heart, or the core of not just digital innovation, but technological innovation more widely. We had this development happening. At the same time that we saw this tendency across many governments that I’ve already discussed, towards outsourcing core functions and infrastructure, and their servicing. So that being framed not as a way of rolling back the state. As I mentioned before, it’s quite important to differentiate between this view of the state versus the previous anti-State rhetoric of Thatcher and Reagan, even though we know that the state did actually increase in terms of spending under their governments.
PM: Just to drill down on that point for a second, you’re not saying that the government was coming out and saying: We’re privatizing these services by bringing in these technologies. Rather, they’re saying we’re bringing in these technologies, and that’ll allow us to deliver these services better to you.
RC: Exactly. That’s really important to recognize that difference. I guess, this is more of a reflection on how I probably once and how on the left we often talk about public private partnerships and privatization and outsourcing. There is an assumption that this is done in order to shrink the State, and there is an explicit agenda to do that. I think it’s more interesting, as I have suggested in this paper, to explore the ways in which these developments are happening, and then perhaps they are co-opted, or perhaps they have this consequence, and that’s inadvertent. Or perhaps this happens because there is lobbying or because there is involvement of companies that are working in this area that are working with the State. And perhaps they recognize that this is going to be good for them in the long run, but within the State, at a time that we recognize there is hollowed out capacity within it, like within IT management, or within other parts of public services.
Perhaps this isn’t an agenda of the bureaucracy — perhaps privatization isn’t what the government wants. It’s not what it’s aiming at and it’s aiming other things like improving citizen communication. That doesn’t mean we always have to take this in good faith, that’s not what I’m suggesting. It just means that we have to recognize that the ambitions can be quite different, which makes the politics of this and thinking about how the different actors play into it across the public and private sector, how they influence this as well. We were talking about drilling down into this development, and how between 2002-2019, we saw this transformation of how public sector digital infrastructure and public sector data became resources that the public sector wielded for the growth of markets in health tech and digital technologies.
In Denmark throughout the 2000s, we still saw the rhetoric and the language of these policy documents that, at least what was written, the aim of these documents, very much reflected these broader third way governance trends around making government work better for citizens, making states more efficient, making governments more transparent. That was a huge thing, and we have to remember, again, looking at this in a good faith way. We’ll talk about the open government data movement as well at some point. At this time, there was a lot of distrust. I mean, there’s always distrust around governments, but in the wake of governments heading off to wars, lots of citizens were very frustrated, and angry about this. There was a lot of distrust about what governments were doing and citizens wanted to scrutinize this and opening up government data was seen as a way of making governments more transparent, and therefore accountable.
Then in the wake of the financial crisis — which affected all governments, affected all economies — we saw governments then looking at the resources that they had internally. This included public sector data in Denmark, for example and in other countries as well, and the digital infrastructure as a resource that can be wielded and used by actors, including SMEs but also larger companies to develop technologies that might reduce costs internally. That was one framing of them. But crucially in Denmark, that could also be used through of these public-public or triple helix partnerships between the university, a private sector actor, like an SME and public sector body like a hospital. They could then use services to develop new technologies that could then be used in export strategies.
PM: For me, it’s very concerning to read how that history has played out, and how these things have evolved in such a way as to in the beginning, it’s just about efficiency, but then later it becomes: Okay, we’re building companies on top of this public data. Now these companies are being built into the healthcare system because the healthcare system is then buying their services back because it’s become something that they have become dependent on. Then it creates lock-in so the service — whether it’s the healthcare system, as we’re talking about now, or other public services, that are using other technologies developed by private companies — now there’s a lock-in so it’s harder to get off of this service, regardless of what changes into the future. Then that creates restrictions on what this public service can actually do, because it’s dependent on this tech product that’s developed by a private company, and the private company probably doesn’t have — or we know doesn’t have — the same sorts of incentives and things like that actually driving it.
RC: Absolutely. This issue of lock-in came up a lot in interviews that I was doing, as part of this research on on Denmark, and actually the Danish government did recognize some of these consequences of lock-in. Normally, when we talk about platform capitalism, and the technological lock-in that happens in platform capitalism.We’re thinking about how these different platforms then become embedded or how the State becomes embedded and enmeshed with these different systems that then can’t get out over, or it’s going to be very expensive to get out of. One thing that became very clear in this research was not just a technological lock-in, but also there’s a capacity lock-in that develop. Where in the process of outsourcing not just the infrastructure — also the management and servicing of this infrastructure to private actors — governments lose the capabilities. They’re no longer having teams internally that are able to do this — let alone even manage the contracts. Often, they lose control of contracts, or governments can lose control of contracts, because they don’t have the capabilities for the people working within them, who are actually able to know what’s going on
Another great case of this in Denmark that came out— I’m not sure I wrote about it in this paper — but the Capital Region government, so Copenhagen, the local government for Copenhagen, developed a partnership with IBM or entered a partnership with IBM Watson that was one of the most expensive healthcare technology partnerships that has ever existed in Denmark. It was a five-year partnership to develop healthcare AI technologies that could be used in the health service. That sounds really broad because it was a really broad contract. Within a few years, it became clear that it wasn’t going anywhere. Some of the people who had been involved in this contract early on were saying to the press: There was really something of the Emperor’s New Clothes about this; we didn’t know; we couldn’t assess, we couldn’t evaluate what IBM was saying to us. So it’s not just that these capabilities don’t exist in house to do this in house, but also that the government has lost capacity to assess whether what companies are saying to them and make sense, or if it has any potential. This whole partnership, even though it was really expensive, ended up folding after just a few years.
PM: Back in the moment when there was all that hype around Watson and what it was going to deliver. So what you identify there is so important. Especially, if we think about the types of technologies that these public services, public sector institutions, are contracting to bring into their institutions, and whether they’re able to assess whether that is actually going to serve the needs that they have. But it also identifies something larger, where if you lose that capability, that knowledge in house, then it also becomes more difficult to regulate these technologies and to think about those questions, which are obviously things that we’re still grappling with right now. It’s fair to say it’s been proven that governments in many cases have lost these capabilities, have lost this kind of institutional knowledge, and that is left them less prepared to look at what is coming out of Silicon Valley or other tech companies around the world and be able to properly assess the potential impacts of those things. Rather than just the statements that they’re making to the media and putting out through their press releases and things to actually be able to identify the harms or potential harms early on and take measures to try to rectify those things before they actually happen.
RC: Absolutely. This isn’t an issue just for public procurement, or government technological infrastructure, government IT and digital infrastructure. This is also a problem for how governments understand markets, how governments understand these sectors and the technical It’s called developments, more widely, and what they mean for society, Often governments, and I’m talking in very broad terms here, but I can also give some cases. Often governments openly, after a crisis over technology, will openly admit: We didn’t know this was going to happen, or we had no expectation that this was going to happen. This is also just about the loss of government capabilities more generally, and what that means for how much the technology sector and other sectors have within our societies?
PM: Absolutely, and whether government is actually able to deliver the things that we expect from them, as I think many governments have shown, they haven’t been able to do.
RC: Exactly, so there’s a liberal democratic argument around this, which is that if governments change, if government mandates change, and so governments need to be able to adapt the tools at their disposal in order to be able to achieve the mandate that they have, the policy goals that they’re elected on behalf of or that they’re elected to do. If the capabilities don’t exist within them, then it’s very difficult to reconfigure that machine in order to meet these goals. That’s the liberal democratic view of: What is the problem with the loss of capacity, the hollowing out of governments?
There’s also, from a kind of the left politics perspective of: If we want governments to do things that are going to be very different to what governments are currently doing, or ought to have an expanded role, to have infrastructure that increased democratic involvement in the way that decisions are made, and the way things are done in society, then there needs to be some capability or capacity that exists that can be reconfigured. Otherwise, it’s not going to be possible for left governments to achieve anything. We see that actually happening in countries, in South America for example, where left-wing governments are coming in and they’re elected, and they enter and actively find that the government is pretty empty, and there isn’t a lot that they can do. Or it’s very difficult or very expensive for them to do the things that they are elected to do. This becomes then an issue for politics, this hollowing out of capability and capacity. Also, that it’s not just an issue for digital technology and digital infrastructure, but also more widely in governments.
PM: Fantastic point. It goes way beyond that and has much larger implications if we want to demand our governments do larger things, actually address real problems. Maybe that’s part of the reason why — I’m sure many countries but in Canada, as I see it — they’re struggling so much to address the healthcare crises that we’re seeing, because they’ve just been left unable to deal with many of these problems. I want to circle back to something that you mentioned in one of those previous answers, which is the open data. We’ll probably remember, especially American listeners will remember, during the Obama administration when there was a lot of pushing this notion of open data.
People were arguing that we should be providing more government data, putting it online, anyone can access it and do what they want with it. The idea is that this equalizes the playing field that makes government more transparent, anyone can go and make their little tool or whatever that can use the data to make it easier to illustrate things about government to the public or whatever. Then, of course, large organizations, large corporations can have the resources to do much more with that data than the average individual who’s just taking a look at it. In hindsight, how should we look at that movement for open data based on the research that you’ve been doing?
RC: I really liked a term that Timnit Gebru used in your interview with her — this idea of aspirational language in tech. Open government data, or open data more generally, is a really nice example of this. It sounds great. Open data —that’s something for everyone, it’s transparent and veryone can use it. It’s a tool that we can all use, and think about all of these things we can do with this universal resource It sounds great. There’s a few issues with some of those assumptions, so one is that even with nice open government data, and I use open government data in my research. I’m studying the government and government spending and, and contracts. I use this data, it’s still not something that everyone’s going to be able to use. Often it needs wrangling; it needs a lot of cleaning; you need to know how to program to use big sets, because you won’t be able to open them with Excel or something like that. So that’s one issue. This isn’t something that everyone’s going to be able to use, you need to have training to use this data.
That in itself is a challenge to the whole idea of open government data. The other issue with it is that — again a lot of developments in tech that we’ve seen over the past last 20 years — the goals and aspirations of this movement were really important, and we should continue to fight for: freedom, democracy, transparency of all of these tools. Good faith visions of the movements that have come before us is really important. But with a lot of visions, a lot of these aspirational visions in tech and the good people who have pushed them forward, there has been an absence of an understanding of the political economy within which they are being developed. Often, public sector data has been used towards and. I’m not just talking about the data that I was discussing earlier that might be collected.
For example, when a new technology is being developed and introduced with a health system, data that is already collected, or has already been collected and is stored and processed internally. So for example, transport data or mutual logical data or health system spending data, this sort of thing. Often, this is only a very small amount of data that companies like Uber would have access to, but it can be very, very important for them. Governments have not recognized when it is important, and then use that as leverage. Like being used to help shape these companies, or help introduce labor agreements that would be in line with existing labor agreements and in the sector. So actually, to use the example of Uber and the Transport for London, which is a public sector body in the UK that manages transport in London, believe it or not, has published this real time open data through this API for about 15 years. In 2019, it was estimated that this data has been used to develop over 675 mobile phone and online apps, mostly within the private sector.
Before its IPO in 2019, which valued the company at $80 billion, or something like that, Uber integrated TfL data into its app. It promised investors that it would become the market leader in journey planning. This became a huge selling point at this time — this was huge for investors. But of course, those working for Uber as a drivers and couriers were unlikely to, or did not see, improvement in their paying conditions after this data that had been developed by the public sector and used in this way and had been introduced. Perhaps one thing that could have been done with this huge data that Uber was able to get access to could have been to use it as a way to say: We want to ensure that the drivers are able to increase their income or something like this, or a decision might have been made that the data was not given to Uber, but was perhaps given to another company entirely that had different principles that were maybe more in line with what most people would have wanted and would expect from a company that is that big.
There are clear uses of open government data where there is no way that you can argue that they have been used in the public interest. I think that the failure to recognize that is, again, just a bigger problem that we have within these aspirational tech narratives — I really like that idea — that just failed to account for the political economy of these developments. Who owns them? Who’s getting the money? Who controls the infrastructure? Whose data is it? These big questions are just so often ignored and that’s been quite clearly in the case of government data.
PM: It’s a real flaw with a lot of the discourse around technology that the political economy of it has not received enough attention. That goes, that goes back a long way, and that shapes a lot of the narratives that we’ve seen around technology for a long time, these more libertarian framings, these framings are on speech and things like that. What you say is important because if we think about the example of Uber, and I’m sure its valuation is worth a lot less now than then when it IPOs, but it was taking advantage of this public data that was available, both from TfL, but I’m sure from many other public organizations in the United States and in other parts of the world. Then was often not sharing its own data back or trying to fight when regulators or local governments were trying to get a holder or get a look at its data, so it could integrate that with its own data and get a good idea of what Uber was doing in cities and in transport systems. You see that lack of reciprocalality there where it’s benefiting from the public sector, and everything that the public sector is doing, and usually not having to pay for those sorts of things.
Then using that in a way that benefits its business rather than having to ensure that the way it uses it is beneficial to the public as well, and that’s a real problem. One piece that really stood out to me, as I was reading In these two papers that you put together, was that you also talked about — I believe in the in the first paper we were talking about about Denmark — that one of the risks as well is that when we allow companies to use public sector data to create their products, there’s also a risk that they are going to create a product that then helps to undermine the welfare state or the public service altogether. Instead, actually, this sets them up to compete with what is being delivered publicly, where it’s going to be beneficial for that company, certainly, but that’s not necessarily the case that it’s going to benefit the public by having this private company start to edge out, or push out the public service.
RC: A really interesting example that I came across in my research was of, I think I mentioned, there had been a triple helix partnership between — this isn’t very common in lots of other countries — but triple helix partnership is a partnership between a university department or University team, an SME, that could be spun out from the University team often.
PM: And that’s a small and medium enterprise, an SME?
RC: Yes, then a public sector body, so health technology is is probably the most kind of ac·ces·si·ble to talk about. In the case of this startup, the technology had been developed by a university computer science research group, that then developed a partnership with the hospital. It was a cardiovascular disease monitoring data collecting technology, essentially. The long term goal and the way that this university group got funding was when a state comes in and state financing comes in. The way that they got kind of Research Council funding for this technology was on the promise that it would then commercialize the technology, and it would then be able to export this as part of the government’s broader export led growth strategy.
The technology was developed in collaboration with one of the hospitals in Copenhagen, and it was a cardiovascular disease monitoring technology. It was essentially just collecting a lot of data. Historically, this had been a task that had been done by nurses, and, interestingly, the nurses who were working in this position on with this new technology, were vehemently against the introduction of this new technology. They didn’t trust it; they didn’t think it was accurate. These are hugely expert people who have been doing this job for a long time. They didn’t think that it was in the interest of patients to be doing this, to be using this technology. But effectively, they probably — this didn’t come up in my interviews with the founder of the SME that had spun out from the university project — they will say recognize that this was something that the goal was for it to replace them in this task collecting this data. Also, the bigger picture was that the technology had started out as something quite simple. It was initially just set up to collect the technology as part of this broader monitoring of patients, but the goal with this data that was then collected would then also be commercialized by this company in the development of its artificial intelligence technology that would then be exported for use in hospitals around Europe.
This is not just the replacement of an existing function or existing task that’s happening within the welfare state by what becomes a private actor. This isn’t even that it was developed using public funding, using public universities. At the time that it was being used in the hospital, it was being provided by an SME. It’s not just that the task is then replaced by a private-sector actor, but also the collection of the data, the processing of the data, what happens to that data, shifts from being the responsibility of the public sector, to becoming something that that private sector actor is responsible for. The ability to then use that data to develop a technology in house — or for it to be used in meaningful evaluations internally — and learning or the development of capabilities internally is also lost as that process, the process of doing something, is outsourced.
PM: It’s so concerning to hear you describe that because it’s exactly what we wouldn’t want to happen. I feel like some of these conversations are a bit more difficult to understand from an American perspective, where the healthcare system is private, this is what they’re used to. If you have a private hospital, and they’re using more private technologies it’s harder to understand the distinction and what’s going on there. Maybe it’s not as big of a worry. When we’re thinking about, say, Europe and Canada and Australia where you have these technologies that are developed by the private sector moving into the public sector health system, and taking away some of the power of that system, taking away some of the things that it used to do. What I describe it as as kind of like a sly privatization. You don’t see it when you go and approach the health service, when you go and see your doctor or get a surgery, but this is happening in the background.
It’s still having impacts on the type of service that you can expect, on what the health service can deliver to you. It potentially has implications, or as you’re describing, it does have implications down the line. One thing I would add to this that maybe you can comment on as well is that during the pandemic we had a lot of tech companies trying to compete for public sector contracts around COVID, and COVID tracking and all these sorts of things. One of the big things that emerged from that moment, if I remember correctly, is Palantir getting a contract with the NHS, the public health service in the UK, so that then they get access to a lot of tha data. As I understand it, the NHS data is seen as world class, really valuable. Private companies really want to get this data, really want to have access to it, and now a company like Palantir is getting access to that data, what does that suggest to you? Because that seems concerning to me.
RC: I think there are two possible responses, or two things that I can say about this that might be relevant. The first is that, as we saw during the pandemic, often when these companies are brought in to do these things, — this is where this kind of thing about them being the Emperor’s New Clothes comes back— they actually aren’t able to do the things that they’re promising that they’re able to do. They often aren’t able to do the things in a way that governments expect them to, and often they can end up spending a lot more money. A really nice example of this would be in the case of the US and the development of healthcare.gov, which was the market exchange platform that was introduced as part of the Affordable Care Act. Now on the day of its launch, this was 2013, on the day of its launch only six people were able to access the platform, which is terrible. That’s terrible result for any project, that on the day that it’s kicking off, no one’s actually able to use it. It was hailed by media and Republicans as this colossal failure. This was supposed to be the landmark reform of the Obama Administration. It completely failed.
Now, there’s been lots of scrutiny of what government did wrong in the development of healthcare.gov, about how there was miscommunication between departments, about how some of the managers weren’t open minded. These really micro-level things. The big elephant in the room is that this was also a program, this is also a technology or public sector technology, that had a development that had been outsourced at huge scale and scope by companies that had never developed a platform like this before. That were subcontracting after subcontracting after subcontracting different parts of the contracts, so the government lost a lot of control over the process, even if it had the capabilities, which that’s not really what I’m interested in this one. Even if the government did have those capabilities, which it probably didn’t because it’s also outsourced a lot, even if it had the capabilities to manage that process properly, it had no oversight over what was happening through the huge supply chains that had been developed.
Some of the big contracts, the CGI group for example, were handed out on a cost plus fixed fee basis, which meant that the overall cost was $24 million in the end. This basically meant that if CGI group made a mistake, or if there were problems, or if something was more expensive, rather than the risk of that and the responsibility for footing the bill for that falling to them, it fell to the government. They were just able to bill the government afterwards. The one issue that these companies aren’t able to deliver on the promises, that they promise and as we’ve discussed a bit already, often the government is not able to assess those promises. After decades of outsourcing, and the infantilizing of governments as we describe it in this book that I’ve written, they just don’t have the capacity or the confidence sometimes to make these decisions challenging these companies. That’s one issue. The other issue, which is related to this second point as well, is that when governments do this over a long period of time, like in the pandemic contracts, they completely lose the ability to do this themselves.
On the one hand, you think: Why didn’t governments just use the money, the huge contracts that they handed out for test and trace programs, why was it not spent internally? There’s an obvious reason for that, which is that actually governments don’t have these teams to do these things. The companies that they went to, like Deloitte in the UK, also didn’t have those capabilities and they weren’t able to assess that they weren’t able to do this. This isn’t necessarily a case of governments doing everything either, that’s not what we argue in the book by any stretch of the imagination. There’s lots of things that governments shouldn’t do, and also can’t do at the moment. In the case where they need to work with other actors, they need the capabilities to assess those actors, there needs to be a good understanding, a thorough understanding, of the interests and incentives of those actors. Also then, how that might shape how they work with governments and that would enable governments to work with them in a way that might actually be in a public interest, or at least the government’s interest.
PM: Just to add to your example — even Palantir got its reputation for scanning all of this military data and saying that it could identify terrorists and all this stuff. There was stories that suggests that it was involved in finding Osama bin Laden, which was never true, but they made sure they didn’t actually say it wasn’t true, and said: People have said that we were maybe involved with this. That helped to build up the reputation of their data analysis capabilities as to get a lot of, not just militaries around the world, but private companies and now increasingly, healthcare systems to be working with them to seek out their services. To build it into like a massive juggernaut. Of course for people who don’t know, Palantir is associated with Peter Thiel.
One other example I will throw out here that maybe backs up your point as well. This is just pulling from my own experience — to give an example of something that I’m concerned about especially when it comes to technology going into healthcare and the things that we’re talking about — is what we saw during the pandemic was a push for more virtual health care, getting doctor’s appointments online, and those sorts of things. Of course, that has been popular in Canada, as well as many other countries. One thing that we’ve seen is that many of those virtual doctor’s appointments, especially when they’re not with your family doctor and you’re just accessing a service that allows you to speak to a doctor, those are private, those are not within the public system.
One thing that has come out in my province in particular is they’re actually paying more for private virtual doctor’s visits, then they would pay to your family doctor, if you went to see them in the public system. The incentives just seem completely turned on their heads, but one way for me that these technologies enable a further creeping of privatization in the system that you might not necessarily notice, and that in some cases allow you to pay for one of these services to get a quicker appointment with the doctors and things like that. These second tier of services if you’re able to pay you can get better service creeps in there. I think it’s a concern, and I think it is an example of how technology and how we associate technology as being something that is just in the private sector, that is not done by the public sector, leads to services being privatized in a way that maybe we wouldn’t want, but it’s actually what ends up happening
RC: Completely! I think you’ve hit the nail on the head of two things. The first thing, which is probably the most important problem to identify is the cost of these things. Both the upfront costs and this is where it relates to broader trends and outsourcing. Outsourcing has been hugely expensive. Sometimes the initial contract will be cheap for different outsourcing functions, including technologies. Companies are allowed to, they’re permitted to, lowball in ways that can be completely egregious sometimes. KPMG famously bid on a 1 million pound contract in the UK in 2013 For one pound. This is allowed. That’s one of the issues, but the long term cost — even though that initial contract was very cheap because that gets in the door then that hollows out the internal capacity — the long term cost of using this mode of governance or this mode of doing things, mode of production, becomes very expensive over time. Then with the loss of that capacity and the absence of the choice of doing something in house that then increases the leverage that the private actor will have, because it knows it’s then the only actor in town that’s able to do this.
This can get really expensive in the case of what you’ve described, sometimes you look at these contracts and think: Why on earth has this happened? On the face of it, even on paper, it’s way more expensive than doing this how it’s currently done. That’s where we have to look not just at how these cost benefit analyses might feed into decisions, but also when narratives about the innovative potential of private firms and video based consulting in medicine, for example — the failures of the public sector where all of these narratives come together and shape decisions also within government and within public sector organizations within local government — how these will feed into the decision to instead turn to the private sector rather than continuing with a public sector option. The second point that you raised around creeping privatization, I completely agree, this is something that I talked about in the paper. I looked at this through the lens of Paul Pierson’s concept of ‘systemic retrenchment.’
Paul Pierson was writing on changes in the welfare state systems and different types of welfare state models. He was really interested in why Thatcher and Reagan were not able fulfill their ambition of completely rolling back the state through what he termed ‘programmatic retrenchment,’ so that’s the direct cuts. He looked at the importance of political constituencies and how they then create feedback loops for different types of support for different policies and different welfare programs. He suggested that systemic retrenchment is often more critical for how privatization actually happens — how the transfer of resources away from the public sector to the private sector actually develops — most critically is not through these direct cuts to public services, but often through invisible measures that are obfuscated, both very consciously by politicians. An example of systemic retrenchment is increasing the retirement age because it decreases the cost of pensions to the government over time, but it’s not something that most people are going to think about for a long time. So it’s an example of that.
Digitalization is another one, that’s what I argued, because it happens quite slowly. These lock-in effects often happen quite slowly. Towards the end of the paper, I talked about then how this industry, the sector of digital technology, also government digital technology, must also be recognized as a powerful actor in shaping this obfuscation of the effects of its use at scale and scope within the government. There is money invested, there are resources, there are lots of people employed to help promote the idea that the private sector is the best provider of government, digital infrastructure and services. This is not something that develops out of thin air, there is an industry also promoting this idea. So this also then figures into how systemic retrenchment happens, how this kind of creeping privatization, as you put it nicely, occurs.
PM: You’ve explained it perfectly. It also leads really well into the final question that I had for you. You talk a lot about how this process is depoliticized. That is a key piece of allowing it to happen and allowing us to not know that it’s happening and not get angry that it’s happening because it’s something that’s easy to hide away and don’t see. You talk about how it’s depoliticized even though that is the case. It’s an explicit strategy by particular people to promote the private sector as the solution and technology in particular, as a magical fix to all these problems, when really that’s not something that is inherent, that is just the way that things are. Rather, it’s a very political view of how these things can work, and they could work in a very different way if we had a different set of politics guiding these things and if we were more aware of what was happening and able to push back on them.
RC: This feeds into, again, as someone who very much studies production, I’m wary of speaking too much about ideas, but we have to recognize how these ideas are really important to how we understand the role of government and society. So these narratives, not just about technology and about innovation and where that happens, but also what the function of government is. The notion that government, and government administration, is because a lot of these technologies are not being used in frontline services, but public sector digitalization. All that applies to what’s happening behind the doors of ministries and government departments, and places that most people don’t think about.
When they do think about them, they either think of them being stuffy, gray, boring, bureaucratic, Kafka-esque mazes. We have these images in our head, or lots of people do, where workers within them are described as lazy and not doing a lot. In fact, in the UK, there’s this picture of the home working shirking, public sector worker has become quite prominent in recent weeks and months, as the government has been announcing kind of fresh cuts to public sector pay and other areas. We either have that or we have this neoliberal vision of the bureaucrat as being very selfish and only in it for their own interests. Therefore, it’s in everyone’s interest to make sure that the private sector is involved as much as possible. These kind of dual narratives, which are competing in the minds of most people mean that when we think about public sector digitalization, we aren’t thinking about the politics, because we don’t think of this as being a realm of politics.
We either think of it as being stuffy, gray, neutral, boring place that is apolitical, or we think of it as being this self-interested realm, that should not have control over anything. And therefore bring in the private sector. Both of these very kind of deep, apolitical spaces. It’s no real surprise that there isn’t more discussion about public sector digitalization, or public sector infrastructure and outsourcing more widely beyond headlines about fuck ups by individual companies, which are really important. But I think it’s important to dig a bit deeper. This is actually how then my research has become not just looking at IT, and public sector digitalization, and who owns data or who owns these infrastructure, which is what I was initially interested in. The bigger paradigm within which this has emerged in government and the political economy.
I realized, quite surprisingly, that — even though there has been great research looking at outsourcing and some research looking at privatization of public sector infrastructure — often these aren’t placed within these bigger lenses, and what they mean for who owns the economy and who benefits from the economy. Or what the long-term implications are for government capacity looking beyond the programmatic retrenchment, the direct cuts to how this stuff evolves, even when individuals — even when individual bureaucrats, even when individual politicians — have great ambitions and good ambitions. They often do. I like to look at things with good faith a lot of the time. I’m not saying we always should, particularly the UK Government, for example. Absolutely not, but there are people also within these organizations that do have good ambitions and are benevolent and do want to use these things for good, but, often, constrained in doing so, the reasons we’ve been discussing.
PM: Absolutely, I think it’s a really important discussion. It will really be important to us as we think about what our government should do in the future, what our healthcare system should look like in the future, how public services are delivered and to ensure that they can actually serve the public. Rosie, I really appreciate you taking the time to chat with me about this. I wish you the best with the book, as it comes out. Thanks so much.
RC: Thank you very much.