Appendix 1 Summary of European Buyout Market
In 2006 the buy-out market value for the whole of Europe
*(including the UK) reached €160.3 billion from 1428 deals. This is almost three times the amount invested in buyouts in 2003. The increase in value was primarily due to activity in Continental Europe where value rose from €92.9 billion in 2005 to €121.5 billion last year. In the UK value rose from €35.4 billion to €38.8 billion over the same period. The average deal value for the whole European market is €112 million. There were 752 buy- outs in Continental Europe compared to 676 in the UK.
Activity at the top end of the market (deals of value greater than €500 million) has increased dramatically from 24 such buy-outs in 2003 (worth €22.3 billion) to 46 buy-outs in 2006 (worth €91.1 billion). The largest buy-out in the whole of Europe in 2006, and the largest deal ever recorded by Centre for Management Buyout Research, was the public to private buy-out of Danish company TDC for €13 billion. (www.cmbor.org)
*Europe = Austria, Belgium, Denmark, France, Finland, Germany, Ireland (Eire), Italy, Netherlands, Norway, Portugal, Spain, Sweden, Switzerland, UK.
Appendix 2 Average Value of Indicators per Year
Table 1a Annual Average Value of Indicators of Buyout Fir
ms
Year -1 Year 1 Year 2 Year 3 Year 4 Background Indicators
EBIT 14.507 12.319 13.057 11.977 11.181
Number of Employees 1.661 1.565 1.499 1.328 1.259
Total Sales 231.852 237.163 222.700 205.988 194.917
Total Assets 166.567 163.045 153.958 146.351 138.687
Operating Performance Indicators
EBIT / Sales 7,20% 7,76% 7,96% 6,76% 5,78%
EBIT / Assets 10,10% 11,13% 11,28% 10,05% 8,39%
Labor Productivity 36,99% 45,36% 44,82% 37,28% 31,68%
Employee Effect Indicators
Assets per Employee 178 183 192 210 208
Wage Rate 40 42 42 44 44
Absolute values are in thousands of €
Table 1b Annual Average Value of Indicator of Control Firms
Year -1 Year 1 Year 2 Year 3 Year 4 Background Indicators
EBIT 9.262 9.045 8.949 7.610 6.795
Number of Employees 991 1.047 1.042 1.021 949
Total Sales 169.936 193.209 195.477 190.653 185.562
Total Assets 136.909 147.432 151.934 148.732 148.700
Operating Performance Indicators
EBIT / Sales 5,34% 5,18% 4,76% 4,09% 3,78%
EBIT / Assets 7,18% 5,97% 6,06% 4,46% 4,14%
Labor Productivity 38,92% 33,02% 25,64% 27,90% 23,42%
Employee Effect Indicators
Assets per Employee 322 333 340 340 331
Wage Rate 44 43 44 46 48
Absolute values are in thousands of €
The results for year zero, the fiscal year that includes both pre- and post-buyout operations, are not taken into account in this research, because they are difficult to interpret as pre- or post-buyout performance. Furthermore, EBIT measured in year zero is biased, because of buyout related fees etc. and therefore not included in this analysis (Kaplan, 1989).
Appendix 3 Pre-Buyout Test Results
Table 2a Pre-Buyout Test Results Wilcoxon Matched Pairs Signed Ranks Test
P-Value Normality Sum of Negative Ranksa Sum of Positive Ranksb P-Value N Background Indicators
EBIT 0,000 543,50 276,50 0,073* 40
Number of Employees 0,000 677 143 0,000*** 40
Total Sales 0,000 543 277 0,074* 40
Total Assets 0,000 565 255 0,037** 40
Operating Performance Indicators
EBIT / Sales Inconclusive 516 304 0,154 40
EBIT / Assets Inconclusive 513 307 0,166 40
Labor Productivity 0,007 379 441 0,677 40
Employee Effect Indicators
Assets per Employee 0,000 124 696 0,000*** 40
Wage Rate Inconclusive 250 530 0,051* 40
* Significant at 10 percent level
** Significant at 5 percent level
*** Significant at 1 percent level a Control firm < Buyout firm bControl firm > Buyout firm Absolute values are in thousands of €
In the Wilcoxon test, the difference between the variables per indicator background, operating and employee) of the target and control firms are computed. Subsequently, the difference scores are transformed into absolute values, by removing all the positive and negative signs. These absolute values are ranked from lowest to highest value, where the lowest absolute difference score is assigned rank 1, the second lowest rank 2 and so forth. The cases where the difference scores between target and control firm is zero are eliminated, since they provide no useful information. After ranking the absolute values of the difference scores, the sign of each difference score is placed back again in front of its rank. Next ,the sum of the ranks that have a positive sign (W+) and the sum of the ranks that have a negative sign (W-) are calculated. If the sums of the positive and negative ranks are nearly equal, this means that the null-hypothesis is correct and that the sample is derived from a population in which the average difference scores equals zero. However, if the value of W+ is significantly greater than the value of W-, this indicates a high likelihood that the sample is derived from a population with higher scores per indicator than the other population. The null hypothesis is accepted when the smallest sum of ranks is in the rejection region, and the resulting P-value is larger than the ten, five or one percent critical value. (Sheskin, 1997, pp. 83-94; Moore & McCabe, 2001, pp. 665-693).
When the sample is normally distributed (P-value > 0.050), the paired samples t-test is appropriate to test for significant differences between the buyout- and the control firms. In the table below several important characteristics of the t-test are provided. Mean is the average difference of percentage change of buyout minus control firms. The other variables are self explanatory. The null-hypothesis is accept if the P-value is larger than critical value.
Table 2b Pre-Buyout Test Results Paired Samples T-test
P-Value Normality Mean St.Dev Lower$ Upper$$ P-Value Test df Operating Performance Indicators
EBIT / Sales Inconclusive 1,86% 8,83% 0,96% 4,69% 0,190 39
EBIT / Assets Inconclusive 2,92% 11,44% -0,74% 6,58% 0,115 39 Employee Effect Indicators
Wage Rate# Inconclusive -4,17 0,13 -0,08 -2,03 0,049** 39
* Significant at 10 percent level
** Significant at 5 percent level
*** Significant at 1 percent level
$ Lower level 95% Confidence Interval
$$ Upper level 95% Confidence Interval
# Absolute values are in thousands of €
Appendix 4 Pre- and Post Buyout Difference Test Results
Appendix 4a Pre- and Post-Buyout Difference Analysis Year -1 vs 1
Table 3a Pre- and Post Buyout Test Results of the Wilcoxon Matched Pairs Signed Ranks Test P-Value Normality Sum of Neg. Ranksa Sum of Pos. Ranksb P-Value N Background Indicators
Total Sales Inconclusive 291 489 0,167 39
Total Assets Inconclusive 228 592 0,014** 40
Operating Performance Indicators
EBIT / Sales 0,003 287 274 0,908 33
EBIT / Assets Inconclusive 260 268 0,940 32
Employee Effect Indicators
Wage Rate Inconclusive 389 352 0,788 39
* Significant at 10 percent level
** Significant at 5 percent level
*** Significant at 1 percent level a Control firm < Buyout firm bControl firm > Buyout firm
In the Wilcoxon test, the difference between the variables per indicator background, operating and employee) of the target and control firms are computed. Subsequently, the difference scores are transformed into absolute values, by removing all the positive and negative signs. These absolute values are ranked from lowest to highest value, where the lowest absolute difference score is assigned rank 1, the second lowest rank 2 and so forth. The cases where the difference scores between target and control firm is zero are eliminated, since they provide no useful information. After ranking the absolute values of the difference scores, the sign of each difference score is placed back again in front of its rank. Next ,the sum of the ranks that have a positive sign (W+) and the sum of the ranks that have a negative sign (W-) are calculated. If the sums of the positive and negative ranks are nearly equal, this means that the null-hypothesis is correct and that the sample is derived from a population in which the average difference scores equals zero. However, if the value of W+ is significantly greater than the value of W-, this indicates a high likelihood that the sample is derived from a population with higher scores per indicator than the other population. The null hypothesis is accepted when the smallest sum of ranks is in the rejection region, and the resulting P-value is larger than the ten, five or one percent critical value. (Sheskin, 1997, pp. 83-94; Moore & McCabe, 2001, pp. 665-693).
When the sample is normally distributed (P-value > 0.050), the paired samples t-test is appropriate to test for significant differences between the buyout- and the control firms. In the table below several important characteristics of the t-test are provided. Mean is the average difference of percentage change of buyout minus control firms. The other variables are self explanatory. The null-hypothesis is accept if the P-value is larger than critical value.
Table 3b Pre- and Post-Buyout Test Results of the Paired Samples T-test
P-Value Normality Mean St.Dev Lower$ Upper$$ P-Value Test df Background Indicators
EBIT 0,180 5,08% 98,04% -29,12% 39,29% 0,764 34
Number of Employees 0,200 -11,73% 23,99% -19,61% -3,84% 0,005*** 37 Total Sales Inconclusive -5,70% 31,51% -15,92% 4,51% 0,265 38 Total Assets Inconclusive -15,32% 37,87% -27,44% -3,21% 0,014** 39 Operating Performance Indicators
EBIT / Assets Inconclusive -2,52% 79,37% -31,13% 26,10% 0,859 31 Labor Productivity 0,160 0,62% 97,02% -31,73% 32,97% 0,969 37 Employee Effect Indicators
Assets per Employee 0,112 -3,15% 43,31% -17,00% 10,70% 0,648 38
Wage Rate Inconclusive 0,79% 20,73% -5,93% 7,51% 0,812 38
* Significant at 10 percent level
** Significant at 5 percent level
*** Significant at 1 percent level
$ Lower level 95% Confidence Interval
$$ Upper level 95% Confidence Interval
Appendix 4b Pre- and Post-Buyout Difference Analysis Year -1 vs 2
Table 4a Pre- and Post Buyout Test Results of the Wilcoxon Matched Pairs Signed Ranks Test P-Value Normality Sum of Neg. RanksaSum of Pos. Ranksb P-Value N Background Indicators
Total Assets Inconclusive 219 601 0,010** 40
Operating Performance Indicators
EBIT / Sales 0,008 264 232 0,754 31
Employee Effect Indicators
Wage Rate Inconclusive 392 388 0,978 39
* Significant at 10 percent level
** Significant at 5 percent level
*** Significant at 1 percent level a Control firm < Buyout firm bControl firm > Buyout firm
In the Wilcoxon test, the difference between the variables per indicator background, operating and employee) of the target and control firms are computed. Subsequently, the difference scores are transformed into absolute values, by removing all the positive and negative signs. These absolute values are ranked from lowest to highest value, where the lowest absolute difference score is assigned rank 1, the second lowest rank 2 and so forth. The cases where the difference scores between target and control firm is zero are eliminated, since they provide no useful information. After ranking the absolute values of the difference scores, the sign of each difference score is placed back again in front of its rank. Next ,the sum of the ranks that have a positive sign (W+) and the sum of the ranks that have a negative sign (W-) are calculated. If the sums of the positive and negative ranks are nearly equal, this means that the null-hypothesis is correct and that the sample is derived from a population in which the average difference scores equals zero. However, if the value of W+ is significantly greater than the value of W-, this indicates a high likelihood that the sample is derived from a population with higher scores per indicator than the other population. The null hypothesis is accepted when the smallest sum of ranks is in the rejection region, and the resulting P-value is larger than the ten, five or one percent critical value. (Sheskin, 1997, pp. 83-94; Moore & McCabe, 2001, pp. 665-693).
When the sample is normally distributed (P-value > 0.050), the paired samples t-test is appropriate to test for significant differences between the buyout- and the control firms. In the table below several important characteristics of the t-test are provided. Mean is the average difference of percentage change of buyout minus control firms. The other variables are self explanatory. The null-hypothesis is accept if the P-value is larger than critical value.
Table 4b Pre- and Post-Buyout Test Results of the Paired Samples T-test
P-Value Normality Mean St.Dev Lower$ Upper$$ P-Value Test df Background Indicators
EBIT 0,135 0,52% 82,81% -32,02% 30,98% 0,973 28
Number of Employees 0,200 -15,09% 28,97% -24,62% -5,57% 0,003*** 37
Total Sales 0,200 -6,99% 36,21% -18,89% 4,91% 0,242 37
Total Assets Inconclusive -16,99% 47,81% -32,28% -1,70% 0,030** 39 Operating Performance Indicators
EBIT / Assets 0,200 22,82% 67,62% -2,43% 48,06% 0,075* 29
Labor Productivity 0,200 11,68% 68,82% -13,56% 36,92% 0,352 31 Employee Effect Indicators
Assets per Employee 0,200 -0,30% 49,31% -16,07% 15,47% 0,970 39
Wage Rate Inconclusive -0,08% 24,11% -7,90% 7,74% 0,984 38
* Significant at 10 percent level
** Significant at 5 percent level
*** Significant at 1 percent level
$ Lower level 95% Confidence Interval
$$ Upper level 95% Confidence Interval
Appendix 4c Pre- and Post-Buyout Difference Analysis Year -1 vs 3
Table 5a Pre- and Post Buyout Test Results of the Wilcoxon Matched Pairs Signed Ranks Test P-Value Normality Sum of Neg. Ranksa Sum of Pos. Ranksb P-Value N Background Indicators
EBIT Inconclusive 235 200 0,705 29
Employee Effect Indicators
Assets per Employee Inconclusive 376 444 0,648 40
* Significant at 10 percent level
** Significant at 5 percent level
*** Significant at 1 percent level a Control firm < Buyout firm bControl firm > Buyout firm
In the Wilcoxon test, the difference between the variables per indicator background, operating and employee) of the target and control firms are computed. Subsequently, the difference scores are transformed into absolute values, by removing all the positive and negative signs. These absolute values are ranked from lowest to highest value, where the lowest absolute difference score is assigned rank 1, the second lowest rank 2 and so forth. The cases where the difference scores between target and control firm is zero are eliminated, since they provide no useful information. After ranking the absolute values of the difference scores, the sign of each difference score is placed back again in front of its rank. Next ,the sum of the ranks that have a positive sign (W+) and the sum of the ranks that have a negative sign (W-) are calculated. If the sums of the positive and negative ranks are nearly equal, this means that the null-hypothesis is correct and that the sample is derived from a population in which the average difference scores equals zero. However, if the value of W+ is significantly greater than the value of W-, this indicates a high likelihood that the sample is derived from a population with higher scores per indicator than the other population. The null hypothesis is accepted when the smallest sum of ranks is in the rejection region, and the resulting P-value is larger than the ten, five or one percent critical value. (Sheskin, 1997, pp. 83-94; Moore & McCabe, 2001, pp. 665-693).
When the sample is normally distributed (P-value > 0.050), the paired samples t-test is appropriate to test for significant differences between the buyout- and the control firms. In the table below several important characteristics of the t-test are provided. Mean is the average difference of percentage change of buyout minus control firms. The other variables are self explanatory. The null-hypothesis is accept if the P-value is larger than critical value.
Table 5b Pre- and Post-Buyout Test Results of the Paired Samples T-test
P-Value Normality Mean St.Dev Lower$ Upper$$ P-Value Test df Background Indicators
EBIT Inconclusive 5,40% 81,19% -25,48% 36,29% 0,723 28
Number of Employees 0,200 -13,44% 34,37% -24,73% -2,14% 0,021** 37
Total Sales 0,200 -3,05% 46,39% -18,30% 12,20% 0,688 37
Total Assets 0,072 -18,54% 59,34% -37,78% 0,69% 0,058* 38
Operating Performance Indicators
EBIT / Sales 0,200 12,31% 76,25% -15,66% 40,28% 0,376 30
EBIT / Assets 0,068 11,37% 73,34% -16,53% 39,27% 0,411 29
Labor Productivity 0,120 10,93% 75,06% -16,13% 37,99% 0,417 31 Employee Effect Indicators
Assets per Employee Inconclusive -3,86% 56,45% -21,92% 14,19% 0,667 39
Wage Rate 0,200 0,66% 27,82% -8,35% 9,68% 0,882 38
* Significant at 10 percent level
** Significant at 5 percent level
*** Significant at 1 percent level
$ Lower level 95% Confidence Interval
$$ Upper level 95% Confidence Interval
Appendix 4d Pre- and Post-Buyout Difference Analysis Year -1 vs 4
Table 6a Pre- and Post Buyout Test Results of the Wilcoxon Matched Pairs Signed Ranks Test P-Value Normality Sum of Negative Ranksa Sum of Positive Ranksb P-Value N Operating Performance Indicators
Labor Productivity Inconclusive 235 200 0,705 29
Employee Effect Indicators
Wage Rate Inconclusive 328 452 0,387 39
* Significant at 10 percent level
** Significant at 5 percent level
*** Significant at 1 percent level a Control firm < Buyout firm bControl firm > Buyout firm
In the Wilcoxon test, the difference between the variables per indicator background, operating and employee) of the target and control firms are computed. Subsequently, the difference scores are transformed into absolute values, by removing all the positive and negative signs. These absolute values are ranked from lowest to highest value, where the lowest absolute difference score is assigned rank 1, the second lowest rank 2 and so forth. The cases where the difference scores between target and control firm is zero are eliminated, since they provide no useful information. After ranking the absolute values of the difference scores, the sign of each difference score is placed back again in front of its rank. Next ,the sum of the ranks that have a positive sign (W+) and the sum of the ranks that have a negative sign (W-) are calculated. If the sums of the positive and negative ranks are nearly equal, this means that the null-hypothesis is correct and that the sample is derived from a population in which the average difference scores equals zero. However, if the value of W+ is significantly greater than the value of W-, this indicates a high likelihood that the sample is derived from a population with higher scores per indicator than the other population. The null hypothesis is accepted when the smallest sum of ranks is in the rejection region, and the resulting P-value is larger than the ten, five or one percent critical value. (Sheskin, 1997, pp. 83-94; Moore & McCabe, 2001, pp. 665-693).
When the sample is normally distributed (P-value > 0.050), the paired samples t-test is appropriate to test for significant differences between the buyout- and the control firms. In the table below several important characteristics of the t-test are provided. Mean is the average difference of percentage change of buyout minus control firms. The other variables are self explanatory. The null-hypothesis is accept if the P-value is larger than critical value.
Table 6b Pre- and Post-Buyout Test Results of the Paired Samples T-test
P-Value Normality Mean St.Dev Lower$ Upper$$ P-Value Test df Background Indicators
EBIT 0,158 4,68% 71,83% -25,65% 35,01% 0,752 28
Number of Employees 0,070 -13,70% 36,87% -25,81% -1,58% 0,028** 37
Total Sales 0,200 -9,73% 46,94% -25,16% 5,70% 0,209 37
Total Assets 0,056 -21,84% 64,34% -42,70% -0,99% 0,041** 38
Operating Performance Indicators
EBIT / Sales 0,200 0,83% 57,27% -21,83% 23,49% 0,940 26
EBIT / Assets 0,200 20,47% 73,49% -10,56% 51,50% 0,186 23
Labor Productivity Inconclusive -4,12% 86,42% -36,99% 28,76% 0,799 28 Employee Effect Indicators
Assets per Employee 0,200 -8,65% 53,65% -26,04% 8,74% 0,321 38
Wage Rate Inconclusive -2,00% 29,39% -11,53% 7,53% 0,673 38
* Significant at 10 percent level
** Significant at 5 percent level
*** Significant at 1 percent level
$ Lower level 95% Confidence Interval
$$ Upper level 95% Confidence Interval