Private equity and employment effects

Private equity critics claim that buyouts bring job losses. Previous studies have documented that PE-ownership is related to both job creation and job destruction, depending on employment measures and country specific data. However, the majority of these studies are based on non-Danish data e.g. Davis et al. (2014) who document that PE-buyouts in US lead to greater job loss at establishments operated by portfolio companies as of the buyout year (refer to chapter 3). We consequently find this discussion on implications for Danish employment effects very relevant. We strive to discuss whether PE-ownership leads to gross or net increase in employment on a Danish society level.

Former Prime Minister of Denmark, Poul Nyrup Rasmussen, is a critic of PE-firms and their operations. He claims that leveraged buyouts does not create a higher employment increase in portfolio companies compared with the general employment effects in comparable firms created from general cyclical improvements (Nyrup Rasmussen, 2007). His argument is based on i) investigations from other sources than himself, ii) previous studies that violates the truth

107

and validity of the conclusions due to inaccurate assumptions and methods e.g. the time frame analyzed and no isolation of inorganic growth, and iii) Danish cases e.g. TDC.

In chapter 7 our statistical tests of the full data set documented that growth in FTEs is on average 16.54% higher for portfolio companies relatively to industry averages. This finding is statistical significant at a 95% significance level.

This positive median change is also much larger compared to the development in the Danish employment rate, which in the period 2000-2014 actually decreased by 2.19% (Danmarks Statistik, 2017), when using a two-point estimate calculation method of ending value and beginning value, as this is the same calculation method applied in the Wilcoxon Signed Rank test in chapter 7. Note that the labor force survey from Danmarks Statistik was changed in methods of calculations from 2007 and onwards⁶, however, Danmarks Statistik claims that it is possible to compile long data series, hence we evaluate the estimate as valid for our discussion.

We are aware of the potential bias of inorganic growth included in our data set, however, with the median difference being significant on a 95% significance level as well as the large difference between the general employment rate in the event window and the test statistics from table 14 in chapter 7, the result point towards a net employment effect.

Our finding contradict with the claim of Poul Nyrup Rasmussen, as this effect all else equal suggest that there is a positive net employment effect on the Danish buyout market in the event window, however, the result is very difficult to justify 100% taking into consideration the potential biases and limitations in the equation.

The finding of an increase in employment due to PE-ownership is supported by the study of Økonomi- og Erhvervsministeriet (2006) reviewed in chapter 3. Compared to a group of reference firms during 1995-2005 it was found that portfolio companies have significantly higher growth in employment. However, as mentioned in the review of the study in chapter 3, this study should be interpreted with caution, which is also highlighted by Poul Nyrup Rasmussen, because the study apply a less precise control group since they define a reference firm as the ‘median firm’ with 120 employees. Another limitation is that the study does not include extreme observations. One problem with excluding these extreme observations is that especially portfolio companies could be exposed to either a post-buyout expansion or

6 In November 2011 the tables of the Labour Force Survey were revised from 2007 onwards. Please

note that detailed analyses of the development in employment from 2006 to 2007 should be conducted with caution. Equally, caution is needed when detailed analyses of the employment from 2003 to 2004 are conducted. However, it is possible to compile long data series (Danmarks Statistik, 2017).

108

downsizing strategy, this is thus neglected by the study of Økonomi- og Erhvervsministeriet (2006).

Another empirical study suggesting that the claim of Poul Nyrup Rasmussen is not supported by evidence is made by EQT, who finds that all portfolio companies in Sweden owned and sold by EQT had annually growth rates in number of FTEs by 8% five year post exit (EQT 2015). Note that this study does not control for industry effects and there may be a potential source bias as the study is made by EQT themselves.

Further, from the literature review in chapter 3, a study of Holm (2013) finds that significant employment effects can be documented in portfolio companies relative to industry peers.

What should also be taken into consideration with regard to the employment effects is the substantiality of the discussion. In 2015, 26,600 were employed in Danish portfolio companies out of a total Danish working force of 2.18m (DVCA, 2016). Hence, approximately 1.2% of the total work force was employed in companies owned by PE-firms in 2015, which should be considered when evaluating the effects ~ how much can such fraction affect the total employment in DK in positive vs. negative direction?

Even if we assume that the critics are right that PE-ownership does not lead to positive net employment effects in Denmark, one can argue that effects of PE-ownership of Danish portfolio companies on employment can be found in foreign countries.

“When politicians are worried about the employment effects of private equity operations, they often refer to the effects on a Danish perspective. But is this relevant? It depend whether you measure the effects on a global or national scale. Portfolio companies owned

by EQT have on average experienced a growth in FTEs on 8-10%. Much of this growth is attributed to foreign countries, and it is often in countries other than DK where a net employment effect can be measured, e.g. in the case of ISS.” (Nicholas Hooge, EQT).

Another aspect to be included in the equation of employment effects is the indirect effect of increased activity in the portfolio companies.

“Along with increasing activity in portfolio companies with regard to sales and purchase, there are derived effects in e.g. the activity of suppliers. If increased activity of suppliers

lead to employment effects in the supplier firms, this effect should be considered in the overall equation.” (Nikolaj Vejlsgaard, Axcel).

It is also very difficult to speculate on the effects on employment if a portfolio company was not acquired by a PE-firm, but kept in its current ownership structure.

109

“When we acquired the majority of the shares in F. Junckers Industrier A/S, it was necessary for us to close down 150 jobs. Of course, this was not a great thing for the employment on society level, however, in case we had not acquired the firm, one can argue

that maybe even more jobs would have been lost?” (Nikolaj Vejlsgaard, Axcel).

In other words, you also have to include the trade-off effect between PE-firms acquiring a company vs. not acquiring. In some cases, when a PE-firm initially acquire a company, it is already in a bad situation with regard to profitability and in general inefficient operations (refer to e.g. case example of F. Junckers Industrier A/S in chapter 9). So what would be the case if the PE-firm did not enter the business and made the necessary initiatives to improve numbers? Well in the case of F. Junckers Industrier A/S, this company had suspended its payments when Axcel did their entry. Thus, a possible bankruptcy case could have occurred, which is definitely destroying jobs and employment. So in some cases it seems that there is evidence that PE-firms are not just ‘bad guys’ destroying value, but is often a catalyst for minimizing trade-off effects.

Sometimes, it also seems that outsourcing of jobs from Denmark to foreign countries is just a need.

“In Royal Copenhagen, we had to move a lot of jobs from Denmark to Thailand due to the fact that the required competencies simply were not in place in Denmark as the amount of

educated plate painters in Denmark is decreasing.” (Nikolaj Vejlsgaard, Axcel).

Another claim from the former Prime Minister with respect to employment effects is that employee satisfaction is violated by PE-firms. His arguments are among other things the fear of firings and less favorable working conditions as a result of PE-firms using employees as just a source for increased earnings (Nyrup Rasmussen, 2007).

From table 16 in chapter 7, we documented that labor productivity, measured in terms of Revenue/FTEs, is 7.09 percentage points higher in portfolio companies’ relative to industry averages, however, this result is not statistically significant. A study from 2014 conducted by economists from the University of Warwick, document that an increase in labor productivity is directly related to an increase in job satisfaction. More concretely, it was documented that job satisfaction leads to a 12% increase in labor productivity (Oswald, Proto, & Sgroi, 2014).

Note that this study is made on UK employment data, hence a direct parallel to the Denmark should be made with caution, however, as we documented in table 4 in chapter 2, the Danish and UK buyout market is rather similar when it comes to PE investments and divestments when taking into account the sizes of the respective economies. Hence, this may support the

110

argument that the finding of the study of Oswald, Proto, & Sgroi (2014) can be used in the discussion on the Danish implications.

With this study in mind, the findings of this thesis that portfolio companies are 7.09 percentage points more productive than industry averages, does not support the claim of Poul Nyrup Rasmussen that PE-operations violate employee satisfaction. However, note that our result is not statistically significant. Another angle to consider in the discussion is the result of Lichtenberg & Siegel (1990) who find that plants involved in LBOs during 1981-1986 had significantly (about 14 percent) higher rates of productivity growth over that five-year span than other plants in the same industry due to increased intensity of effort by labor, which may be created from employee satisfaction, Oswald, Proto, & Sgroi (2014) suggest.