Sample selection

The data sample of this thesis comprise sell-off and spin-off divestments from European firms an-nounced and completed in the period from January 2000 until December 2020. European firms are categorized based on the parent firm’s officially registration location, which most often equals the location of the headquarter.²⁷ Transactions are not limited to European subsidiaries. Due to limited available financial information on parent firms from Eastern Europe, the sample is limited to only include parent firms geographically located in Western Europe defined as European Developed Mar-kets in the Capital IQ database. Therefore, the final sample includes transactions completed by par-ent firms based in the Austria, Belgium, Cyprus, Denmark, Finland, France, Germany, Greece, Ire-land, Italy, Luxembourg, Malta, Norway, Portugal Spain, Sweden, SwitzerIre-land, and United Kingdom.

The time span criterion is based on the date of announcement, which is related to the date when the transaction has been announced publicly either through regulatory filings, a company press release or news articles. Transactions that were announced in the sample period but not completed are excluded. The relatively long period of time was necessary to find enough spin-off deals for the data sample. Thereby, the final data sample include transactions from different time periods, which might reduce comparability of specific transactions. However, the long period ensures sufficient data cov-ering full business cycles and transaction waves.

As described in the Section 6, we investigate shareholder value creation of corporate divestments by analysing short-term and long-term stock returns requiring the seller (parent firm) to be publicly listed with accessible share price data. In addition, the analysis of changes in operational perfor-mance requires access to reliable accounting information. Therefore, only transactions of sellers

26 The primary SIC code represents the industry of a firm’s core business. Firms can have up to 5 secondary SIC codes reflecting other industries that the firm is involved in.

27 This is the same definition as used by Kirchmaier (2003).

listed on a stock exchange minimum one year ahead of the divestment announcement were in-cluded. In addition, the sample of corporate divestments were limited to transactions with 100%

change of ownership to ensure comparability across the sample. ²⁸ Simultaneously, only analysing transactions with 100% change in ownership increase robustness of inferences about value creation.

The spin-off subsample includes transactions in the Capital IQ database, where “the parent company distributes a certain number of subsidiary shares to each of its existing shareholders.” The sample only includes spin-offs with 100% of the shares in the subsidiary distributed to shareholders of the parent firm.

The sell-off subsample includes transactions in the Capital IQ database characterized as corporate divestitures, a subcategory of M&A in the database. To ensure comparability of the sell-off and spin-off subsamples, an important focus in the data selection process was placed on discarding sell-spin-off transactions involving asset sales such as single plants, machines or warehouses as those assets are not comparable to a standalone division or unit spun off.

Additionally, a minimum transaction valuation requirement of DKK 50m has been applied in order to screen relevant transactions in the Capital IQ database. Hereby, all transactions without a registered transaction value or with a value below DKK 50m is excluded, in order to avoid potential skewness impact of extremely small divestments. Simultaneously, the idea is to exclude transactions that can-not be listed on a standalone basis. In addition, the availability of accounting data increases with size. However, the risk of the data selection method applied is that not all corporate divestments are included, which can potentially create selection biases.

Based on criteria described above, the initial list of sell-off and spin-off transactions was constructed using the Transaction Screening Report tool in the S&P Capital IQ database. The initial sample of divestments consisted of 4,128 transaction records. Of those transactions, 218 were spin-offs and 3,910 were sell-offs.

Afterwards, additional selection criteria were imposed on the initial sample as the screening filters on capital IQ is not customized enough to ensure comparability of observations and to fulfil the re-quirements of the methodology described Section 6. Therefore, a large initial sample was prioritized requiring manually sorting and screening to reach the final data sample. This process resulted in several transactions being discarded. The criteria applied is highlighted below with the number of transactions disregarded in brackets, followed by a short motivation of each criteria.

28 Both sell-offs and spin-offs can be completed with partial change in ownership, where the parent firm is only selling a stake in the subsidiary.

I) Transactions including parent firms or subsidiaries without SIC code [2157]

Parent firms and subsidiaries without a SIC code in the Capital IQ database was disregarded in the data selection process. It was observed that many of the transactions including firms without SIC codes could be characterised as asset sales causing the missing SIC code. Thereby, this criterion was used to remove transactions defined as asset sales which could not have been divested as single entities.

II) Transactions with parent firms operating in heavily regulated industries [265]

Transactions involving parent firms operating in heavily regulated industries have been removed from the data sample. In relation to methods used in previous research studies including Powers (2001) and Bates (2005), parent firms operating in financial and real estate with SIC codes between 6000-6799 are discarded. Those industries are characterised by heavy regulation and massive pres-sure from legislators resulting in regulatory constraints on the firm structure of large corporations, potentially affecting the market and shareholder reaction of the divestment. This might put pressure on corporations to involuntarily divest business units. Involuntary and compulsory transactions would be noisy in the analysis. Therefore, transactions in those industries were removed.

III) Sell-off transactions involving payment methods other than cash [139]

Sell-off transactions where the payment method has been anything else than cash were removed.

To ensure comparability of the spin-offs and sell-offs included, we investigate corporate divestments with a clear cut between the parent firm and the divested subsidiary. In transactions with alternative payment methods including common equity, the parent firm still maintains its interest in the divested unit. The idea of this criterion is in line with the previously described criterion of 100% change in ownership.

IV) Insufficient publicly available data to confirm announcement date [176]

In the process of validating the announcement, several transactions were disregarded due to insuf-ficient public information about the transactions, i.e., the announcement date could not be confirmed.

Often, access to firms’ press releases is restricted to a recent time period making it challenging to find older press releases. If the parent company have been acquired after the transaction, the access to historical press releases is often limited.

Similarly, this criterion comprises disregarded transactions where the divestment was rumoured prior to the official sell-off or spin-off announcement, which made it difficult to determine the exact event date. The same problem was present for transactions where the parent had previously announced the intend to divest its subsidiary. Particularly, the sell-off transactions in which the subsidiaries were officially put for sale or transactions resulting in bidding processes have been excluded. In accord-ance with the EMH elaborated in Section 3.1, the markets incorporate new information to stock prices when the information is released. If a company has previously announced the intend to divest

a subsidiary, the announcement of a sell-off or spin-off was partially expected reducing the share-holder wealth effects at the announcement date. Therefore, divestments of subsidiaries rumoured and partly expected before the actual deal announcement have been excluded.

V) Parent firms announcing multiple transactions on the same day [28]

In the data selection process, substantial considerations were put into how overlapping event should be handled. In some of the existing studies, transactions of firms engaged in multiple divestments within a specified period up to a five-year window centred on the completion day were excluded from the sample (e.g. Daley, et al. (1997)). Thereby, firms continuously divesting business units as part of their corporate strategy would be disregarded. However, Brauer, Mammen, and Luger (2017) have provided evidence that previous sell-off experience has substantial importance on subsequent firm performance after a divestment. This is consistent with basic learning curve theory demonstrated in other academic areas. If a firms’ ability to conduct value increasing divestitures increase with the number divestments completed, this methodology would risk excluding some of the most value cre-ating divestments. Therefore, we have only excluded firms announcing several transactions on the same day or in the event windows applied in the analysis of the short-term stock return.²⁹

VI) Transactions were misclassified [48]

The initial screening from Capital IQ included a number of transactions that were misclassified. A limited number of transactions classified as spin-offs was disregarded as the transactions proved to be either listings of tracking stocks or transactions as part of a merger between two or more firms.

Specifically for the sample of sell-offs, some transactions were actually complete takeovers or part of a larger merger. In addition, a limited number of transactions involved buyback agreements or other agreements regarding parent firm’s access to assets of the business unit following the divest-ment. Including the transactions above would complicate inferences about the underlying value cre-ation of corporate divestments.

VII) Involuntary transactions [2]

Involuntary transactions were removed from the final data sample as including transactions due to restructuring, bankruptcy or competitive authorities requiring the divestment. An example of a disre-garded transaction was EnBW AG announcing a sell-off of the subsidiary GESO Beteiligungs in 2010. However, the divestment was requested by Germany anti-trust authorities in order to proceed with an acquisition in VNG-Verbundnetz Gas. Thereby, the transaction might not have been driven by the objective of maximising shareholder wealth. The other disregarded transaction was Konecranes Plc’s divestment of STAHL as the sell-off was required by the European Union anti-trust review related to Konecranes pending acquisition of Terex Corporation's Material Handling and Port Solutions business.

29 For further information please see Section 6.1.

VIII) Insufficient firm data on Capital IQ or insufficient accounting data [69]

Analysing and comparing transactions across a twenty-year time period involves risk of insufficient data from the time of the transactions. Missing or insufficient stock price information on parent or spin-off subsidiaries in the Capital IQ database has resulted in several transactions being disre-garded. As described in Section 6, the methodology applied in this thesis requires reliable and con-sistent financial and stock price data for firms included in the final sample. Several transactions from the initial screening have been removed due to missing or insufficient stock prices, market capitali-zations, accounting data etc.

We acknowledge that removing observations due to specified criteria may reduce representativity of the final data sample. The dilemma of using the data that survived the screening criteria are defined in the academic literature as survivorship bias which is previously documented to have measurable influence on results (Brown, et al., 1992). However, the applied criteria are based on extensive search of existing literature and previous empirical studies on corporate divestments increasing re-liability and, thus, the results of the analysis.

Figure 8: Summation of data selection process

The first three criteria above were implemented as filters in excel to sort the dataset. Subsequently, the sample of transactions were manually validated using other M&A databases including Merger-market, Zephyr and ThomsonOne. The reliability and validity of particularly the event study method-ology depends highly on the accuracy of the announcement date and the closing date. Therefore, considerable time have been used on validating those dates for each transaction to ensure that the event study of abnormal returns captures the actual event. The registered announcement and clos-ing date in capital IQ of each transaction included in the final data sample have been manually cross-checked with related publicly available official statements and press releases or with other databases mentioned above. The purpose of cross-checking several databases is to increase the validity of the final sample.

Consequently, the final sample consists of 1244 transactions divided on 98 spin-offs and 1146 sell-offs.³⁰ The purpose of the thorough screening process is to ensure that the transactions included were either spin-offs or sell-offs by nature. In addition, the criteria were to ensure comparability of sample firms increasing ability to capture the true underlying value creation effects. It should be noted that the 1,244 transactions are our final base sample. However, various analyses may require additional specific data, why even further transactions may have been discarded for the specific analyses. We have highlighted the number of transactions included in the various analyses for trans-parency purposes.