When could hiring practices give rise to competition concerns?
Competition authorities have recently intensified their interest in firms’ hiring practices, driven by the fierce battle for talent in the generative AI sector. This scrutiny raises questions about whether poaching multiple employees from a rival firm could amount to a ‘backdoor takeover’ and whether such practices should be subjected to the same degree of scrutiny as traditional mergers. But are these comparisons justified? Do labour markets present unique characteristics that complicate the case for intervention? And – most fundamentally – do competition authorities have a clear and consistent set of objectives in addressing these issues?
"OpenAI is nothing without its people”
Last September a defiant message of solidarity spread across social media as dozens of employees of OpenAI – the company behind ChatGPT – rallied against the board's decision to fire CEO Sam Altman. The board cited concerns about Altman not being "consistently candid in his communications". However, this move sparked a public backlash from the staff, many of whom defended Altman. Faced with the threat of a mass exodus, the board reinstated Altman within days.
The boardroom melodrama captured worldwide media attention, and competition authorities also took note for two reasons. First, the incident highlighted the importance – and concentration – of human talent in the development of generative AI: here was a potentially revolutionary technology being spearheaded by a small organisation of (at the time) a few hundred people. Secondly, it served as a reminder of the links between generative AI firms and the established tech giants. During the crisis, Microsoft – a major investor in OpenAI – offered to hire Altman and any other disgruntled employees, prompting speculation that the "best bromance in tech", as Altman once described the relationship, might evolve into a full merger.
Several competition regulators are now probing Microsoft’s relationship with OpenAI. For instance, the European Commission has indicated that it may open an antitrust investigation while the UK’s Competition and Markets Authority (CMA) has sought views on whether it should assess the partnership under merger control rules. Nor has the CMA stopped there: for example, in April 2024 it issued an invitation to comment in relation to Microsoft’s hiring of former employees from Inflection AI (a rival to OpenAI), after Microsoft poached the firm’s CEO, its chief scientist and some of its senior AI engineers. The CMA has now confirmed it is launching a full merger inquiry.
Probing the poaching
These developments prompt several questions. First, is it appropriate to assess firms’ hiring practices through the lens of merger control? On one level the parallels are tempting, since the types of competition concerns that could arise from aggressive poaching from rival firms mirror the familiar ‘theories of harm’ that motivate merger control, such as:
- Foreclosure concerns. Skilled labour is an essential input in many industries, especially in knowledge-intensive sectors. When critical know-how is concentrated in the minds of a few individuals, a well-resourced firm could potentially drive out rivals and stifle competition by poaching all the top talent.
- Killer acquisition concerns. In addition to removing existing competitors, monopolising talent could prevent potential future competitors from emerging. Often, the most credible potential entrants into a market are those that actively supply it with inputs. For example, if a company produces car engines, it is already halfway to producing cars. Similarly, with the potential for generative AI to be embedded in a wide range of digital services in the future (or even to replace some services entirely), established firms may naturally want the AI ‘cool kids’ to join their camp rather than pitch their own tent.
At the same time, the free operation of labour markets can play a beneficial role in promoting effective competition and unlocking innovation: if one firm is better placed to make use of top talent than its rivals – for example, because it has distinctive business practices or resources that can boost its productivity – then it should be able to offer better terms, including higher remuneration, to attract it. This market mechanism stands to benefit employees since a bidding war for their services should enable them to secure a pay package that reflects the value they are creating for the firm. And it should also ultimately benefit consumers, by encouraging workers to go where they can deploy their skills most productively.
Any interventions in labour markets on competition grounds, therefore, need careful consideration. This can be especially challenging in fast-moving sectors characterised by high levels of innovation, since both the benefits and risks to competition may relate to products that have not yet been developed. In these sectors, the tools that competition authorities have traditionally used to evaluate the effects of mergers on competition may not work, since many of these tools rely on empirical evidence relating to prices and patterns of consumer behaviour, which may not yet have reached a settled state in newly forming markets.
Casing the capabilities
In retooling themselves to assess these questions, competition authorities can draw on a rich body of academic literature on the connections between market structure, firm structure and innovation. As discussed in detail in a previous Frontier article published last year, one important insight to come out of this literature is that a firm’s capabilities – rather than its existing suite of products – can play a critical role in determining whether it has a competitive advantage in a dynamic industry. The importance of different capabilities will vary from market to market, but all firms need to combine a range of capabilities to be successful. This is especially important in dynamic markets where firms need to innovate to survive, since, in the modern world, innovation almost invariably springs from bringing together teams of specialists with a range of complementary skills. Long gone are the days of the Renaissance polymath who knows everything about everything.
This ‘capabilities’ framework points to three (cumulative and sequential) questions for competition authorities to consider when assessing firms’ poaching practices:
- How scarce is the talent? If the individuals being lured away have skills that are widely available and/or straightforward to acquire, then the poaching practices are unlikely to raise competition concerns, since it should be straightforward for their previous employers to replace them. By contrast, foreclosure or killer acquisition concerns could be more material where the individuals being poached possess rare knowledge or talents and where there is little prospect of others developing these capabilities in the near future.
- Do the capabilities of the individuals being hired complement or duplicate the capabilities of the hiring firm’s existing workforce? Bringing together talent with complementary skills could create more opportunities for innovation and, thereby, more opportunities to bring new products and services to market. By contrast, where new hires have similar capabilities to existing employees, the chances of them unlocking new insights may be limited.
- When new hires have complementary capabilities, does the firm have a convincing plan to integrate their talents? As any jaded supporter of an underperforming football team will attest, bringing together a group of talented ‘superstars’ is no guarantee of success. Instead, to boost productivity and unlock innovation, the organisation must create an environment in which the individuals have the possibility to work together as a high-functioning team. At the most basic level, the new workforce will need opportunities to mix: locking the new staff away in separate subsidiaries or divisions that do not interact with the existing workforce is unlikely to unlock much innovation. Restructuring an organisation to integrate a new workforce may take time and involve some cost and disruption, but if firms are not willing to commit to these costs, then competition authorities may infer that the benefits of integration are limited.
There are several practical indicators that competition authorities could consider to inform their thinking about these questions. For example:
- To assess the scarcity of entry- or mid-level talent, they could track trends in the number of mentions of the requisite skills in job adverts and compare them to trends in the number of university study programmes teaching the relevant skills and how many people are completing these courses. (Both growing rapidly in the case of AI-related skills, according to a recent study by Stanford University.) Data on the average length of time it is taking for firms in the industry to fill their vacant posts may also provide a useful insight.
- When assessing the replaceability of senior talent – and especially industry ‘superstars’ – the most revealing indicator may be remuneration. Individuals possessing unique talents that are capable of affecting the competitive landscape and transforming the profitability of a company should be able to command a compensation package that reflects this. Pioneering companies may now need to prepare for competition authorities to scrutinise their payrolls, bonuses and share schemes with as much interest as their own employees.
- Testing the second and third questions is likely to require a more forensic analysis of the hiring firm’s business plans and organisational structure. If the new hires are bringing genuinely important rare and complementary skills to the organisation, competition authorities would expect to see a paper trail of supporting evidence. In other words, there could be internal documentary evidence of the firm identifying a specific shortfall in its capabilities and systematically looking to fill the gap, either internally through training programmes or externally by identifying potential recruits with the required skills.
In practice, poaching top talent could simultaneously have both positive and negative effects: it could help the hiring firm improve its own offer to customers, while at the same time reducing the ability of rivals to compete. Weighing up these costs and benefits from a consumer standpoint would not be easy, but there are sound reasons for authorities to be cautious about intervening ex ante in such a scenario. After all, if pooling talent can unlock innovations that are so powerful that rivals can no longer hold onto their customers, those customers will not thank competition authorities for standing in the way of such innovation. In these scenarios the better option may be to allow competition in the labour market to function and then, if necessary, to regulate the ‘winning’ firm ex post. This would be consistent with the principles that guide economic regulation in a wide range of industries: for example, in the case of network utilities where market forces naturally tend towards monopoly because it is not efficient to have more than one provider, regulators typically control the behaviour of the monopoly rather than trying to manufacture artificial competition.
Are we human...?
The capabilities framework treats employees as inputs into production processes. While this is valid, employees are also profoundly different from other factors of production such as machinery or technology in more than one respect. For starters, they have their own hopes and dreams, and may resist being moved around like pawns on a regulatory chessboard. This can matter when assessing what might have happened if the workers had not switched firms (the ‘counterfactual’ scenario). For example, if the employees had grown unhappy or unsettled at their previous organisation, they may not have made a constructive contribution if they had stayed on.
More fundamentally, firms do not ‘own’ their workers in the way that they own other assets. Employees have their own rights and cannot be forced to continue working for a particular organisation. Indeed competition authorities themselves are beginning to consider workers’ interests alongside those of consumers in their assessments. For example, in April the US Federal Trade Commission announced rules prohibiting non-compete clauses nationwide, citing higher earnings for workers as one of the benefits it believed the ban could deliver. Earlier this year, the CMA’s microeconomics unit published a report to better understand trends in “employer market power” in the UK and the impact this was having on competition for workers. And in a recent policy brief, the European Commission reiterated its guidelines warning that wage-fixing and no-poach agreements are likely to “infringe Article 101 TFEU by object”, i.e. they are likely to violate competition law on the basis that they are capable of damaging competition. The Commission specifically cited as a concern the risk that such agreements could depress employees’ wages.
It clearly makes sense for policymakers, tasked with maximising social welfare, to take a holistic view of the interests of both workers and consumers. After all, they are the same thing – human beings. However, from the perspective of a competition authority whose specialist remit is to make markets work well for consumers, actively promoting the interests of workers alongside those of consumers risks opening a hornet’s nest of complexity, which the ongoing probes into poaching practices may expose. Imagine, for example, a case where a competition authority concluded that poaching of highly skilled employees was akin to a merger and, further, that it would reduce competition in the relevant product market. It is unclear in this case what a possible remedy to restore competition would be:
- Competition authorities could impose restrictions on employees’ ability to move from one employer to another. While this could conceivably increase consumer welfare if it created a more level playing field between firms and boosted competition, it would also reduce the welfare of employees by limiting their freedom to switch jobs and weakening their bargaining position in wage negotiations.
- Any behavioural remedy also seems infeasible. For example, in vertical mergers where competition authorities identify concerns that the merged entity could foreclose rivals by withholding access to important inputs they control, the merging parties can sometimes assuage these concerns by promising to continue providing these inputs to rival firms on fair and reasonable terms. However, it is hard to imagine that, after changing job (quite possibly in acrimonious circumstances if they have switched to a direct competitor), an employee could be required to continue providing services to their former employer as if nothing had happened.
As things stand, competition authorities seem to be pursuing two incompatible goals. On the one hand, they are seeking to promote competition in labour markets, for example by treating agreements to prevent poaching as violations of antitrust law. Yet in other cases they are investigating firms’ poaching practices and considering whether to block or constrain them in the interests of consumers. Each goal could be justified in isolation, but viewed together they appear contradictory. This tension needs to be resolved. Competition authorities must provide more clarity on how they will weigh up the interests of workers and consumers when assessing firms’ hiring practices – and they must do so as a matter of urgency.