3.3 Analysis of entry mode
3.3.2 Testing for contingency by region, for greenfield ventures only
Without the 194 cases of acquisition for the whole Europe the sample size reduces to 567 new venture cases. Table 5, Panel B, shows their distribution across regions.
Although the exclusion of the acquisition cases somewhat weakened the results of the chi-squared test and 2-sample t-tests, this could be due to the reduced sample size and in any way seems too weak to confer to the greenfield – acquired venture factor any explanatory power. Therefore another alternative explanation is proposed, for example different sector distribution.
Before proceeding to test various possible reasons for regional WOS/JV proportion discrepancy, however, a caution is necessary. All tests above use the count of the subsidiaries as unit of measurement. It contains supposedly the fewest measurement errors; yet, firms differ along size of equity, number of employees, etc. The most informative alternative unit of measurement is the size of the equity - how many dollars was the equity in WOS vs. JV. This new measurement unit is used with the same methods as those above to test for regional WOS/JV proportion discrepancies. Its magnitude makes the chi-square test useless, but the 2-sample t-tests results are valid (the only difference is that the proportion this times equals equity in WOS divided to total equity per region). According to them the equity distribution in WOS vs.
JV in the five regions (755 firms) is unequal both for unadjusted serendipity26 (the UK has higher proportion than the rest of EU) and for adjusted one (the UK lead becomes even higher). The UK has larger WOS/JV equity proportion than the rest of EU, while region 5 stays between those two extremes and is insignificantly different from either of them (see Table 5, Panel C).
When acquired ventures are omitted (leaving 562 new firms in the five regions) the 2-sample t-tests results become economically insignificant for unadjusted serendipity. The only statistically significant difference is 0.07 between regions 1 and 2&3, which means 1% to 7% difference in the ratios. For adjusted serendipity the same statistically significant difference now is of the same magnitude (4% to 16%) as were the differences calculated with all the cases.
If we do not consider serendipity important, this result means that the greenfield – acquired venture factor explains successfully the regional WOS/JV proportion differences, because the loss of significance is so big. The most obvious explanation is that the WOS/JV proportion tends to be lower for M&A than for new ventures, because the former often leaves a share of the capital to previous/general shareholders. Therefore, when to the equal regional
26 Serendipity refers to accidental fluctuations in the equity distribution. “Accident” is
defined as any investment in any sector of one country, which accounts for more than 90% of all the investments in that sector in the same country (therefore it is sector and country specific). For the number of local subsidiaries as measurement unit, serendipity
distortion is negligible, but for the money value of the investment it may be quite
substantial. For the manufacturing sphere, there are two major occasions of serendipity, one in Switzerland (region 2) in the food sector (new WOS for $744 million) and one in Finland (region 3) in the chemical sector (acquired WOS for $735 million). Adjustment for serendipity means exclusion of these two cases from the manufacturing sample.
distribution of the new venture WOS/JV proportion that of acquired ventures’ is added, the overall WOS/JV proportion is reduced for all regions, but for some (UK) less so than for others (the EU peripheral four). The latter effect is due to regional differences in acquired-venture-WOS/JV proportions, with those of UK and core EU being 4.8 and 3, and those of peripheral EU and region 5 being 0.6 and 1.3 respectively. Therefore the different M&A approaches, regulations and (in case of privatization) government policies account for the regional WOS/JV ratio difference. Obviously, these factors are most restrictive in region 4 and less so in regions 1 and 2.
The earlier finding of regional difference, based on number of firms as measurement unit, did not lose its significance when only the new ventures remained in the sample. This happened because the regional WOS/JV ratios in terms of numbers are about the same for greenfield, acquired and all cases, and significantly different in each of them. Now we found that in terms of equity the significant difference in all cases is due to significant difference in acquired cases only, because there is no difference in greenfield cases. Therefore, acquired firms are different, both in terms of number and equity, with M&As in regions 1 and 2 having greater WOS/JV ratio than those in region 4. This is consistent with the proposed explanation of different regional M&A policies. In addition, the results of different “new” ventures’ ratios in terms of number, but not in terms of equity, suggest that in spite of the bigger JV number (relative to the total number in the region) in region 4 the resource commitments to JV are lower than the respective ones for regions 1and 2. More precisely, the average WOS in region 1 and 2 has two times bigger equity than the average JV there, while for region 4 the commitments to
the average WOS are three times bigger than those to the average JV. This may reflect the higher new venture contract risks and enforcement costs in the latter regions. Yet, in terms of overall equity there is no difference in the WOS/JV regional ratios.
In conclusion, at the level of combined sectors and investment time, the WOS-JV distribution appears quite independent from investment regions. Dependencies exist mainly due to risk differentials (with region 4 having smaller-equity, bigger-number JV to share the higher risk) and acquisition rules differentials (with region 4 having more ownership-restrictive rules). If serendipity and measurement in terms of numbers is really important, then the UK appears to have significantly higher WOS/JV ratio than the other regions. The assumption of non-interfering government policy seems realistic with the important qualification for the acquisition policy effect (on the acquired venture WOS/JV ratio) and the “UK-effect” (on all ventures). Whether the latter one really pertains to policy of the government (in UK or, oppositely, in the non-UK regions) remains to be investigated. Analysis of sector and time distributions will be especially helpful, because the government policy is usually sector and time specific.