Tomorrow's Economy

Sweden: will history lead the way in the age of robots and platforms?

21 June 2018

To tackle such mammoth challenges, we need a trade union movement resilient to structural change

Fredrik Söderqvist
Giuseppe Milo / Creative Commons

The digitalisation debate in Sweden has become a debate over the future of the ‘Swedish model’. Given the current trajectory in most developed countries, the future of work looks bleak. As large sections of the labour market look set to be displaced by automation, the workforce must either reskill in order to keep up with the quickening pace of structural change, or face a future of precarious work, perhaps by finding that work via low-paid digital labour platforms. For a country with ambitions of inclusive growth and a generous welfare state, much is at risk. If the bleakest scenario materialises, Swedish institutions, developed and guarded by the country’s trade union movement, will have failed to maintain a long-standing tradition of successfully mitigating the adverse effects of creative destruction on the labour market.

The challenges ahead will require action from Sweden’s trade union movement and its counterparts. Job polarisation, as described by Acemoglu and Autor (2011), will need to be met with reforms to enable workers to upskill, so the labour market can keep up with demand for new skillsets in an increasingly digital economy. The growing importance of digital platforms and their algorithms in our societies may become a challenge as platform business models spread across sectors, resulting in the partial automation of the employer. For a country with 90% collective bargaining coverage, if labour platforms are to be regulated successfully it will be necessary to integrate existing collective agreements into these digital platforms.

In this essay I will briefly discuss these two challenges and how they might be met by Swedish trade unions. Given Sweden’s high rate of collective bargaining coverage, and the strength of the country’s unions and labour market institutions, finding solutions to these problems may ‘only’ require tweaks and adaptations of existing systems. The ideas presented below may appear overly optimistic. However, as Sweden’s labour traditions and institutions may be severely disrupted by the challenges ahead, solution-oriented approaches are perhaps the only option.

A trade union movement resilient to structural change1

Swedish (and Nordic) trade unions are possibly the greatest proponents of structural change in the world. This may be explained through the institutional history of Sweden’s labour market, which developed in the wake of the conflict-ridden early 20th century, into the largely codetermined labour market regime of today.

During the ‘golden era’ of Swedish labour relations (or Saltsjöbadsandan stretching from the mid-1930s to the early 1970s) many of the institutions on which union acceptance of Schumpeterian creative destruction depend were developed. These were modelled on the ideas of the economists Gösta Rehn and Rudolf Meidner of the Swedish Confederation of Trade Unions (Landsorganisationen i Sverige) (LO 1951). The Rhen–Meidner model combined active labour market policies, a general welfare state, solidary centralised wage bargaining, and restrictive macroeconomic policy, in order to achieve full employment, price stability, fair wages and high economic growth. In practice sectors dependent on low-cost labour became unprofitable as centralised wage bargaining increased pay across all sectors. Instead of protecting low-paid, labour- intensive industries, the trade unions proposed supporting displaced workers to migrate to more productive sectors through active labour market policies. The embrace of creative destruction, by promoting labour market mobility, was conditional on the safety-net institutions set up in order to mitigate the adverse effects of structural change to their members.

Institutions mitigating structural change were further developed in the 1970s when white-collar trade unions and employers’ associations signed the first structural change mitigation agreements, or omställningsavtal. These agreements formalised the trade unions’ positive stance on technology-driven rationalisations, as long as employers helped finance mitigation efforts. At the time, there was a sense that the state-run active labour market regimes were not sufficiently oriented towards white-collar professionals. The mitigation agreements created the first job security councils – essentially private unemployment offices owned by employers’ associations and trade unions, with a focus on white-collar professionals. Similar mitigation agreements have since been added to cover most of the labour force, and the job security councils have become very successful in helping redundant workers retrain and find new employment (OECD 2015).

Although many aspects of the original Rhen–Meidner model have been altered, abandoned or later revived over the past 60 years, central aspects pertaining to reskilling remain. The ‘Swedish model’ is not a static or stable concept, but one that is constantly evolving and adapting to prevailing conditions in the labour market. As most of the Swedish labour market regime is regulated by contracts, collective bargaining parties are relatively unconstrained to solve challenges deemed important for a well-functioning and competitive labour market.

Incorporating lifelong learning into the Swedish model

Reports on the number of jobs displaced by the ongoing process of digitalisation range from somewhat alarming – 7% of Swedish workers are at high risk of being automated according to Arntz, Gregory and Zierahn (2016) (see also Arnold et al. this volume) – to utterly dismal – half of all jobs in Sweden according to Fölster (2014). The debate seems to have created a consensus that further reforms are needed to reduce the skills gap caused by this structural change. For a trade union movement that has spent decades trying to convince legislators to invest in lifelong learning, the challenge of digitalisation has finally pushed the issue higher up the agenda (see Karjalainen this volume).

As in most countries, institutions that mitigate the effects of structural change for the individual are reactive. Access to unemployment insurance and job security councils is normally granted after redundancy has taken place. Successful lifelong learning reforms will need to push existing institutions to take proactive measures.

Given the growing complexity and specialisation of the labour market, efforts will need to be focused on the individual. First, taking time off to retrain or upgrade skills will need to be financially viable for working people. Second, if the financial barrier is overcome, for example through the creation of individual competence accounts or funds, institutions must support individuals to make sound investments in their human capital. Here, the state will need to play a pivotal role in a number of areas, such as giving universities a mandate to include skill upgrading in their educational offerings, and provide skill validation, so that work–life experience is better taken into account in completing academic courses or programmes.

Successful lifelong learning reform would result in a higher proportion of workers making regular investments in their human capital to keep up with demand within the labour market. If the adaptations made by these institutions are successful in changing attitudes and behaviour, upskilling may become more career-oriented, rather than geared towards an individual’s current job. Opportunity-based labour turnover may therefore increase, creating upward pressure on wages, to the benefit of union members, and improving matching in the labour market over time, benefiting firms.

There have already been attempts to incorporate lifelong learning into existing mitigation agreements, but these have failed in negotiations because of disagreements over costs (employers have wanted significantly reduced employment protection in exchange). However, given the increased prominence of the digitalisation debate, there are likely to be further attempts in the near future.

Platform work and the algorithm-based employer

Related to the challenges presented by automation is the sudden rise of digital labour platforms, which may complement or replace the traditional role of the employer in leading and directing work within a firm (see Schor this volume). Examples also show that technology may be used to circumvent regulation, including those related to the labour market.

The economic literature on digital platforms (including labour platforms) mainly focuses on the economic opportunities and efficiency gains made possible with platform-based business models. Platforms are economic agents that facilitate transactions and interactions between its users. In order to be successful in these functions, platforms must gain a critical mass of users. If a platform attains a sufficient base network, this may help attract even more users, so platforms display significant economies of scale. If the users give away data when using the platform, the data may be used to further improve the functions of the platform. As a consequence, inter-platform competition is often oligopolistic in developed platform markets, with high barriers to entry (see Hagiu and Wright 2015; Rochet and Tirole 2003; The Economist 2016).

If platforms are marketplaces “where people and businesses trade under a set of rules set by the owner or operator” (The Economist 2016), what happens when the actions of a platform owner or operator have an adverse impact on its users? If a platform market is characterised by duopolistic competition, for example, there may be few alternatives for labour suppliers. If a handful of platforms become the de facto marketplaces for large parts of the economy, and thus critical infrastructure in market economies, how should regulators act in order to promote fair competition?

Digital platforms play a critical role in an increasingly connected society. When ranking the world’s highest valued publicly traded companies we often see Apple, Alphabet (formerly Google), Microsoft, Facebook and Amazon at the top – firms that own and operate multi-sided platforms. The highest valued privately owned company is Uber, a labour platform.

Since its founding in 2009, Uber has quickly become the largest facilitator of taxi services in the world. As of June 2017, Uber had facilitated over 5 billion trips (Holt, Macdonald and Gore-Coty 2017). Apart from disrupting the taxi industry, the company’s advanced digital platform has shown that the traditional role of the employer to lead and coordinate work can now be undertaken by an algorithm, capable of organising work for a highly decentralised workforce.

Uber has been particularly influential in that it has created a highly imitable model, which entrepreneurs have attempted to apply across various industries, and it may prove to be the first incarnation of an advanced, algorithm-based employer.

However, in the past year, a cascade of scandals has tarnished Uber’s tech-glossy reputation. Among the labour-related scandals, we have seen hire-and-fire practices parallel to day labour (one-sided price setting and driver terminations through an opaque rating system), blacklisting (‘blackballing’), the misclassification of labour, and company stores (dysfunctional car leasing schemes). It appears to be corporate culture built on purposefully breaking rules and regulations in order to gain competitive advantages over its competitors, with the end goal of reaching an international (self-driving) taxi monopoly (Thiel 2014).2

Platforms imitating Uber’s business model often adapt a similar attitude towards their labour force (misclassification). Getting to grips with the undesirable aspects of these platforms involves dealing with the preconception that platform labour must be bogus self-employed, precarious, gig work.

A labour platform may be seen as an organisational tool that employers can use to connect its service providers with customers more efficiently. If the platform dictates the terms and conditions of the transactions it facilitates sufficiently, then the platform is not a free market consisting of independent contractors, and should be classified as an employer. The boundaries considered in Coase’s (1937) firm, regarding whether one is working in a free market or within the confines of a firm’s planned economy, are pertinent not only in discussing the shortcomings of platform economies, but in most labour markets where precarious work and bogus self-employment cause social and economic inequality.

In Söderqvist (2016, 2017), I present a so-called Nordic approach to regulating labour platforms. It first involves integrating platforms into existing sectoral collective bargaining regimes. Here, defining employer status is key; this has not yet been a major issue in Sweden, where most service platforms have accepted employer status within our existing labour standards.3 Second, we are exploring ways to make collective agreements easier to integrate and more compatible with the platform firms’ software, in essence developing digitalised versions of existing collective agreements. This may sound complicated, but needn’t be in practice if the process is carried out at a small scale. Sitting down with platform programmers and discussing how best to incorporate important aspects of collective agreements or labour codes is one way of doing this. From such talks, standards or best practices can be developed that may be used by other firms under the same or similar agreements.

This concept can be taken further by making similar adaptations outside the regulatory reach of traditional collective bargaining. The idea is that the social partners take a more active role in adapting and creating new standards in areas of relevant regulation, in order to make them better able to integrate into digital business models. To accomplish this, we propose the creation of a social-partner-owned institution where digital regulatory standards can be developed with a more holistic perspective, closer to market forces, so platforms, unions and national regulators develop novel digital regulatory standards in cooperation. The hope is to provide better digital regulatory standards, closer to market forces than would be possible if regulators take a piecemeal approach to adapting to the rise of labour platforms.

Recognising that such platforms could make a significant positive contribution to the Swedish economy if regulated properly, these proposals would have unions take a proactive approach, focusing on migrating existing regulation rather than attempting to stall the implementation of productivity-enhancing technology. If playing by the rules is made easy for firms with platform-based business models, then the often-heard arguments lambasting the sclerotic regulations of ‘the old economy’ will ring hollow. If, as Lawrence Lessig (1999) eloquently stated, “code is law”, regulators will need to gain influence over the code. To do so efficiently firms may need to devote more resources towards building sustainable and competitive business models.


I have presented solution-oriented approaches to dealing with some of the challenges faced by the Swedish trade union movement in the years ahead. Although the proposals may seem overly optimistic, our tradition, and the long-term survival of our institutions, leave Swedish trade unions with little choice other than to act. The approaches proposed here depend on the ability of capital, labour and politics to engage in constructive dialogue over common challenges. Such a commonsense approach may be difficult in a political discourse where trade unions are not seen as legitimate representatives of the labour force. Getting to grips with the rise in income inequality, democratic polarisation and capital gains capturing an ever-increasing share of economic surpluses, involves fixing fundamental shortcomings of capitalism pertaining to bargaining power. The Swedish example shows that complicated issues within the labour market may be solved at the bargaining table, given that bargaining power between labour and capital is relatively equal. If more countries recognise the key role of trade unionism in solving the challenges ahead, then the transition to the new digital economy may prove less turbulent.


1 For more on the history and development of Swedish labour market institutions see for example Erixon (2010) and Swenson (2002).

2 Continual updates about Uber controversies can be found on the website http://www.uberscandals.org.

3 However, there are unresolved complications as to who should be seen as the employer.


Acemoglu, D. and D. Autor (2011), ‘Skills, Tasks and Technologies: Implications for Employment and Earnings’, in Handbook of Labor Economics, vol. 4, 1043–171, Elsevier.

Arntz, M., T. Gregory and U. Zierahn (2016), ‘The Risk of Automation for Jobs in OECD Countries: a Comparative Analysis’, OECD Social, Employment and Migration Working Paper 189, Paris: OECD Publishing.

Coase, R. (1937), ‘The Nature of the Firm’, Economica, 4(16), 386–405.

Erixon, L. (2010), ‘The Rehn-Meidner Model in Sweden: its Rise, Challenges and Survival’, Journal of Economic Issues, 44(3), 677–715.

Fölster, S. (2014), Vartannatt Jobb Automatiseras, Stockholm: Stiftelsen för Strategisk Forskning.

Hagiu, A. and J. Wright (2015), ‘Multi-sided Platforms’, International Journal of Industrial Organization, 43, 162–74.

Holt, R., A. Macdonald and P.-D. Gore-Coty (2017), ‘5 Billion Trips’, press release, Uber, 29 June, https://www.uber.com/newsroom/5billion-2/.

Lessig, L. (1999), ‘Code is Law’, in L. Lessing, Code and Other Laws of Cyberspace, 3–8, New York: Basic Books.

LO (1951), Fackföreningsrörelsen och den Fulla Sysselsättningen, report to the Landsorganisationen i Sverige Congress, Malmö: translated by A.-B. Framtiden.

OECD (2015), Back to Work: Sweden – Improving the Re-employment Prospects of Displaced Workers, report by the Directorate for Employment, Labour and Social Affairs, Paris: OECD Publishing.

Rochet, J.-C. and J. Tirole (2003), ‘Platform Competition in Two‐Sided Markets’, Journal of the European Economic Association, 1(4), 990–1029.

Söderqvist, F. (2016), Plattformsekonomin och den Svenska Partsmodellen, Stockholm: Unionen.

Söderqvist, F. (2017), ‘A Nordic Approach to Regulating Intermediary Online Labour Platforms’, Transfer: European Review of Labour and Research, 23(3), 349–52.

Swenson, P. (2002), Capitalists Against Markets: the Making of Labor Markets and Welfare States in the United States and Sweden, Oxford: Oxford University Press.

The Economist (2016), ‘The Emporium Strikes Back: Platforms Are the Future – but Not for Everyone’, The Economist, 21 May, https://www.economist.com/news/business/21699103-platforms-are-futurebut-not-everyone-emporium-strikes-back.

Thiel, P. (2014), ‘Competition is for Losers’, The Wall Street Journal, 12 September, https://www.wsj.com/articles/peter-thiel-competition-is-for-losers-1410535536.