This paper, available here, takes a sample of takeovers in Belgium over five years, and estimates their impact on employment growth. It finds that the average merger temporarily reduces employment of the combined entity by −1.4%. However, long-term effects are markedly different. Mergers likely to be motivated by acquiring or defending market power show a stronger and permanent employment reduction of −14%, whereas those motivated by efficiency gains lead to employment expansions of +10%.
Section I sets the scene.
Existing research provides a range of estimates for the employment effects of mergers, with no consensus having emerged about the predominant effect of mergers on employment. Studies that find a negative effect outnumber those that find a positive effect, but all come with caveats.
At the same time, the literature has long noted the potential for efficiency gains from mergers. Acquirers often argue that reduced variable costs can offset market power and lead to lower prices, which in turn can raise optimal output and employment. This can be achieved through several channels, including the sharing of assets, the transfer of superior (intangible) assets of an acquirer such as managerial expertise, or the combination of complementary assets.
One possible reason why estimated employment effects are ambiguous is that most studies only consider the effects on acquired firms, whereas effects on acquirers can go in the other direction. The few studies that look at the evolution of total employment for merged entities are inconclusive. To estimate a well-defined employment effect of mergers, we need to answer the following three questions: (i) How can we identify a merger? (ii) What is a proper counterfactual for a newly created firm? (iii) Why did the firms merge?
Section 2 summarises the reasons for firms to merge, and discusses mergers’ likely employment effects.
The literature has considered a range of motivations why firms engage in mergers or acquisitions. They fall in three broad categories: acquiring or reinforcing market power, lowering costs, and other motivations not directly related to short-run profit maximisation. In the industrial organisation literature, market power has received the most attention since the motivation for such mergers and the expected employment effects are straightforward – in short, market power leads to output restrictions that lead to reduced employment.
For mergers to be welfare increasing, they need to entail synergies or cost savings. There are two forces behind cost reductions. On the one hand, mergers necessarily increase firm size, and this can generate scale efficiencies that lower marginal costs. The impact on this on employment depends on the effect that such scale efficiencies have on optimal output, and on which production factor is affected most. In most cases, however, labour demand will be reduced.
On the other hand, a merged firm can realise efficiency gains that lower marginal costs but have limited impact on the input‐mix. For example, merging firms can pool assets, such as knowledge, patents, or skills, and benefit by using them more intensively. Firms can also possess complementary assets that are more productive when used together. Again, the effects of a merger on employment are ambiguous, but, inasmuch as this may lead to increased output, it can also lead to increased employment.
Section 3 outlines the estimation framework.
To evaluate the impact of takeovers on post-merger employment, the authors make three important modelling choices. First, they use the sum of employment of the two merging firms as a dependent variable, rather than looking only at target firms or acquirers. Second, as counterfactual, they use the evolution of employment for simulated (and similar) firm pairs that did not merge.
Third, they identify observationally similar controls in two steps. They begin by pre-matching discrete covariates of both the target and acquirer. The matching variables are chosen to capture the diversity of motivations for mergers. These variables include the industry of the acquirer and the target, controls to predict selection into takeovers as well as future employment growth and employment size, pre-merger growth, labour productivity, capital intensity, firm age, and an indicator for multifirm corporations. All variables are observed separately for targets and acquirers. In a second step, the authors find matched controls (or construct weights) using a propensity score that is estimated separately within each cell defined by the discrete covariates.
Section 4 and 5 introduce and discuss the relevant data.
The analysis is based on the register of Belgian employers maintained by the National Social Security Office. It covers all private firms with at least one employee from 2003 to 2012. The authors then identify takeovers or mergers between two or more firms between 2006 and 2012 using two sources. The first source is a data set compiled by Statistics Belgium based on all official mergers and acquisitions recorded and approved by the Commercial Court. The second source is based on employee-flow linkages between firms using a linked employer-employee data set. Takeovers involving temporary agencies and firms in highly subsidised sectors are excluded to avoid measurement error, as these firms often show incomparable growth patterns. Takeovers involving very small firms are also excluded, since they do not have a meaningful effect on employment.
Remaining takeovers satisfy the following conditions. The acquirer has at least 10 employees and is at least four years old in the year before the transaction (t−1); the target has at least two employees and is at least one year old in t−1; the target represents at least 1% of employment of the combined entity; the merged entity survives in t. The authors also control carefully for other changes in the firm structure that may occur before or after the takeover, such as divestitures in the year of the transaction or in the post-merger period, and subsequent mergers or spin-offs.
Section 6 discusses the employment effects of mergers.
Takeover activity is a more dynamic and frequent process than most studies of mergers would lead one to expect. Even for the relatively small Belgian economy, over a seven-year period there is a comprehensive set of 2,411 domestic takeovers, or an average of 344 takeovers each year, involving 3,060 target firms. In an average year, more than 4% of all employees in the Belgian private sector work for a company that is involved in a takeover: 3.5% are employed by an acquiring firm, and 0.8% by a target firm. Over an average five-year period, firms that engage in takeover activity cover 17% of total private sector employment.
Immediately following the merger, takeovers have a small but significantly negative impact on the employment of the combined entity. Employment growth is 1.4 percentage points lower than for control firm pairs. However, this adverse effect does not persist for very long. In the first full year after the merger, employment growth is higher for merging firms; the difference is almost of the same magnitude as in the first year, but in the opposite direction. This result withstands a number of controls, and holds constant for different firm-types and across industry sectors.
However, once one controls for transactions where one can expect market power to play a role –e.g. takeovers where the acquirer and target are in the same 3-digit NACE industry, the acquirer shows lower than average growth in period t−3 to t−1, and the target is of above average size in t−1 – one finds strongly negative effects on employment over time. Employment declines by 3.3% upon impact, and this grows to a 6% decline by t+3, always relative to the evolution for the control group. Mergers involving market power amount to one-fifth of all takeovers. The employment evolution for remaining takeovers is now significantly positive from t+1 onward, and even shows an employment expansion of 3.5% by t+3.
Mergers where market power is likely to play a role show a much stronger employment decline right away, lowering employment by 2.5% more than other mergers, and the difference rises to 9.5% over the full four-year period. The point estimates are even more striking if the subsample is defined more narrowly. In industries where HHI is higher than average, mergers lead to an employment contraction of 13.8% (t−1 to t+3), which is 16.8% lower than the employment evolution over all other takeovers. Interestingly, this distinct pattern of strong employment destruction only occurs if two criteria are met: a slow-growing acquirer and large target. Changing the attributes of either of the merging firms has a large impact on the result – sufficing to negate any detrimental effect on employment relative to the average.
Section 7 concludes.
Recent literature has documented rising concentration and aggregate mark ups in the United States. That mergers often serve to increase market power is well understood and underpins merger control.
In a static context, the exercise of market power naturally increases prices and lowers output. The paper shows, however, that mergers are more likely than not to increase employment and, presumably, output. This could further strengthen their impact on concentration, and lead to market power in subsequent years. This potential evolution is an important topic for further study.
This paper provides a different perspective on the ongoing debate about whether and how merger control should assess the impact of a transaction on employment. It contains a very interesting (and to my eye comprehensive) overview of the literature on the topic, with some important insights.
First, it seems that the impact of a merger on employment is negative, at least on the short run, even when the merger leads to significant efficiencies. However, positive effects may cancel the impact of such mergers on employment in the long run. Second, the impact of most mergers on employment will usually be ambiguous, even if there are some discernible trends. Comfortingly for competition practitioners, mergers leading to market power seem to have detrimental effects on employment, with mergers with other underlying motivations having potentially positive effects.
At the same time, one must be cautious. These results are relative to an individual country, and concern individual firms, not market-wide impact. More important, to my mind, this paper tells us very little on how we should address mergers that may have a detrimental impact on employment but may be otherwise pro-competitive – which are those mergers where, by definition, employment considerations will have to be taken into account independently of other competitive effects, and on their own merits.