In this section we try to assess whether and to what extent M&As among EMU financial intermediaries have had any effect on shareholders’ market risk. We have carried out our analyses over an eight-year period – from 2000
graph; methodology and results are discussed in the second paragraph.
Description of the M&A sample
To analyse the effect of M&A operations on the market risk borne by the equity holders of the acquiring companies, we selected a sample from the list of deals commented on in section 4.3. Because our analyses have shown that most opportunities for risk reduction appeared after the changeover to the euro in 1999, we focused on operations concluded from 2000 to 2007 that involved only companies domiciled in EMU countries.
As a first criterion for selection of the sample, we excluded all transactions where the acquired stake was lower than 10 per cent unless complete con- trol over the target was gained; the motivation for applying this filter to the list is that we do not expect to find any meaningful changes in market risk when the share acquired in the target company was minor.
Our preliminary selection comprises 73 listed acquirers that concluded 217 deals over the period of interest; each company was thus involved in three M&A operations on average, while the actual number of deals per company spans a range from one to nine.
A feature of this data subset, which is inherited from the parent list, is that target companies are seldom listed on stock exchanges; this is so in less than 8 per cent of cases. This is unfortunate, since we are de facto unable to per- form any meaningful comparison of the acquirer and the target market risk on these data, as time-series of market price relatives do not exist for most target companies. Therefore, our analyses are based on market risk estimates of the acquirers only.
One quarter of the 217 deals that we initially selected (55 cases) were excluded because of the absence of sufficient time-series data for estimating risk before and after the conclusion of the deal. This left 162 M&A opera- tions performed by 67 companies available for the analyses.
Deals are evenly distributed over the eight years of the sampling period, with a single peak in 2006 (17 per cent) preceded and followed by years of lower levels of activity (9 per cent in 2005 and 2007).
In the sample 75 per cent of cases are in-sector deals, of which two-thirds are domestic (Table 4.8). In contrast to our finding for the complete M&A list, several cross-sector deals in the sample are domestic (15 per cent) and less than 10 per cent are cross-border. Furthermore, there are relatively more financial conglomerates (18 per cent) and fewer banks (53 per cent) among acquirers in the sample than in the list; the former are the only intermediaries engaged in cross-border and cross-sector M&As.
The distributions of target companies by sector for each type of acquirer are very close to those shown in Table 4.2. The similarity to the M&A list also holds with respect to the domicile country of acquirers, except for
Banks Insurance companies
Financial conglomerates
Total
Domestic / In-sector 65 26 1 92
(%) 75.6 55.3 3.4 56.8
Domestic / Cross-sector 7 4 13 24
(%) 8.1 8.5 44.8 14.8
Cross-border / In-sector 14 17
– 31
(%) 16.3 36.2 19.1
Cross-border / Cross-sector
– – 15 15
(%) 51.7 9.3
Total 86 47 29 162
(%) 100.0 100.0 100.0 100.0
Note: A dash indicates the absence of any cases.
Source: Own processing of Zephyr Database.
Italy and Greece, which are under-represented, and Belgium, Germany and Luxembourg, which have a larger share in the sample than in the list.
For all the sample cases we have been able to retrieve or calculate the per- centage equity stake acquired – except for two deals by banks; descriptive statistics by acquirer sectors are displayed in Table 4.9.
The average stake, in the range of 60 to 70 per cent depending on the sector, is well above the majority-holding threshold; also, in at least three- quarters of cases the deal was an acquisition of all the target’s equity.
The dispersion of data, either measured by the standard deviation or assessed by the interquartile range, is not high and is almost the same across sectors;
most cases concentrate towards the upper values. Stakes acquired by banks and financial conglomerates have very similar distributions, with half of cases above 50 per cent and a minimum value of 10 per cent. Insurance companies’
acquisitions were of larger stakes, 10 or 20 per cent above the values recorded in the other sectors; in a few cases, the acquired stake was lower than 4 per cent.6
M&As and changes in market risk
We have examined variations in market risk that could have been induced by M&As on the acquirers’ equity by comparing the individual beta values before and after operations were completed. Since a large body of empirical research provides evidence that beta fluctuate over time, so is not safe to assume that beta estimates taken over long periods are unbiased, we have computed beta over three month spans using daily data.7
Table 4.10 Beta estimates before M&A conclusion Statistics Acquirer sector
Banks Insurance companies
Financial conglomerates
Total
Mean 0.63 0.67 1.02 0.71
Standard deviation 0.49 0.51 0.45 0.51
Minimum ⫺0.36 ⫺0.1 0.17 ⫺0.36
25th percentile 0.28 0.27 0.70 0.32
Median 0.54 0.57 1.08 0.65
75th percentile 0.97 0.95 1.32 1.20
Maximum 1.97 1.83 2.05 2.05
Number of cases 86 47 29 162
Source: Own processing of Datastream Database.
Banks Insurance companies
Financial conglomerates
Total
Mean 63.7 70.6 63.5 65.7
Standard deviation
32.4 32.1 34.2 32.6
Minimum 10.0 4.0 10.2 4.0
25th percentile 39.8 49.0 30.1 41.0
Median 51.0 78.6 50.0 66.6
75th percentile 100.0 100.0 100.0 100.0
Maximum 100.0 100.0 100.0 100.0
Number of cases 82 47 29 158
Note: Two values are missing among banks’ deals.
Source: Own processing of Zephyr Database.
Table 4.10 displays statistics for beta values by sector for all deal types.
Beta have been estimated on closing-price relatives recorded over the quarter before the M&A operation was concluded (for example, the first quarter if the acquisition was completed any day from the beginning of April to the end of May) and retrieved from Thomson-Reuters Datastream.
A general feature that is apparent from the table is that beta values have a wide range of variability, from slightly less than zero to about two. Market risk was remarkably similar for banks and insurance companies: average beta are in the 0.60 range for both sectors; also, the median and quartiles values
be expected from their typical diversification of business activities, financial conglomerates’ beta are closer to unity than those of banks and insurers. The overall proportion of statistically significant beta values is about 65 per cent, and the average R-squared level is 70 per cent.
In order to assess the impact of M&As on risk, we have estimated quarterly beta after the deals were concluded using the same method. We have then computed the difference between the post- and the pre-operation beta for each individual M&A (that is, price data recorded over the quarter when operations were concluded were not considered);8 the distributions of esti- mated beta changes by geographical and sector characteristics are described in Table 4.11.
Domestic M&As have a similar – and generally low – impact on market risk;
average and median values of changes in beta are very close to zero and the central half of the distribution lies in the ⫾0.20 range for both in- and cross- sector operations. Cross-border M&As are quite diverse. In-sector cases have impacts similar to those of domestic M&As, albeit with values that are shifted slightly upwards. Cross-sector operations are the ones that seem to have had the largest effect on market risk, with an average drop in beta of ⫺0.10;
compared to all other deal types, the distribution is also shifted downwards.
While these features may be suggestive of market risk reduction deriving from cross-border and cross-sector operations, there are some reasons for exercising caution. Firstly, only financial conglomerates performed this type of operation in our sample, and they were already quite diversified (Table 4.8 and Table 4.10); thus, further diversification is likely to have been less important to shareholders than in the banking and insurance sectors.
Table 4.11 Beta changes after domestic M&As by type of deal Type of deal Statistics
Minimum 25th perc.
Median 75th perc. Maximum Mean
Domestic/
In-sector
⫺0.90 ⫺0.19 0.02 0.19 1.57 0.03 Domestic/
Cross-sector
⫺0.74 ⫺0.23 0.02 0.22 2.00 0.05 Cross-bor-
der/In-sector
⫺0.60 ⫺0.19 0.07 0.32 0.60 0.05 Cross-bor-
der/Cross- sector
⫺0.75 –0.32 ⫺0.11 0.11 0.37 ⫺0.10
Source: Own processing of Datastream Database.
in the sample.
The picture emerging from the general examination of our estimates is that beta changes are not evidently oriented towards either risk reduction or enhancement, and that if any clear influence is to be found then it will be among outlying cases. Indeed, when variations in beta are tested for statistical significance at customary levels, the proportions of rejections are higher than they would be by chance alone: at the 5 per cent level, the number of significant changes is 22 out of 162 (13.5 per cent) and at the 10 per cent level the figure is 51 (31.5 per cent); this pattern also holds at the more conservative one per cent level and for each deal type.9 In this regard, in Table 4.12 we display the average and count of 10 per cent significant beta variations by deal type, split into positive and negative cases.
The magnitude of significant changes is not negligible in both instances;
positive variations averaged 0.52, negative ones –0.43. In line with the descriptive statistics in Table 4.11, the largest mean risk reduction is in cross- border and cross-sector operations (–0.59), while risk increases by 0.69 on average in domestic and in-sector operations. These types of deals are also the ones for which the absolute differences in (mean) positive and negative changes are the largest. However, the most striking aspect of these results is the balance between the number of positive and negative effects of M&As on beta for each type of deal; such that, if we average positive and negative values for each type, the results are in a range very close to zero (from –0.06 to 0.15).
The analyses of individual operations presented in this section lead towards two main conclusions. First of all, merger and acquisitions among financial intermediaries of EMU countries from 2000 to 2007 were not neutral with respect to the equity market risk exposure of the acquirers’
Table 4.12 Statistically significant beta changes by type of deal and sign Type of deal Sign of Beta change
Positive Negative
Cases Mean Cases Mean
Domestic/In-sector 12 0.69 15 ⫺0.42
Domestic/Cross-sector 5 0.32 5 ⫺0.44
Cross-border/In-sector 5 0.44 3 ⫺0.34
Cross-border/Cross-sector 3 0.35 2 ⫺0.59
All operations 25 0.52 26 ⫺0.43
Note: The table is based on cases significant at the 10 per cent level.
Source: Own processing of Datastream Database.
not only for stock shares of banks and insurance companies, but also for those of financial conglomerates. This is consistent with the potential for market risk variations in EMU countries discussed in section 4.4. However, it is not possible to clearly trace the effects on beta to any specific type of M&A, since negative and positive changes are evenly distributed within each type and, in general, of comparable magnitude. While we would not dispose of the evidence that domestic and in-sector M&As were more likely to increase risk, while cross-border and cross-sector ones might have reduced it, we are not strongly confident in these claims either.
Although difficult to explain, variations in market risk did happen in asso- ciation with M&As. This is consistent with what predicates most of the existing literature on the topic. From a methodological perspective, this should be considered in any empirical research that uses CAPM-like approaches to assess the effects of M&As on shareholders’ market value.