The results show that there is some relationship between the measures of risk and syndication in the direction expected by theory, but it is not particularly strong. The management share of equity is the only variable significant at the 1% level in both models. This has the expected negative sign indicating that the more committed and tied the management is to company performance, the less likely the deal is to be syndicated. In the model using imputed values, the amount of debt has a negative influence (significant at the 10% level) on syndication, whilst the total amount of equity funding has a positive effect (significant at the 10% level). For the weighted complete case analysis, the amount of debt similarly has a negative coefficient, which is significant at the 5% level, whilst gearing is positive and significant at the 1% level indicating that the higher the gearing the more likely the deal is to be syndicated. The coefficients on buyout, divestment and receivership are all insignificant at the 10% level. Only three of the industry dummies and four of the regional dummies are significant at the 10% level or lower in either of the models. Thus, deals in hotels and transport are less likely to be syndicated as are those in the North, South West and West Midlands, whereas deals in Scotland are more likely to be syndicated.
Venture Capital Firms
As regards the characteristics of the venture capital firm, the coefficient on the deals variable is negative and significant at the 1% level in both models. This result indicates that, holding all other factors constant, the more experienced the venture capital firm is the less likely a deal will be syndicated. For the model with imputed values the volume of funds managed and the portfolio size are also negatively related to syndication (significant at the 10% level). These results are consistent with the hypothesis which sees smaller and less experienced venture capital firms as more likely to syndicate deals given that the larger firms are separately controlled for by their own dummies.
The individual dummies for the venture capital firms are all generally significant at the 1% level in both models. These results suggest that not only are there significant differences between venture capital firms in their propensity to syndicate, but the signs are all positive indicating that the major players are all more likely to syndicate deals than their smaller counterparts.
There seems to be a common view in the industry that syndication has changed over time (Wright and Robbie, 1996). The results in Table 4 provide very little evidence in support of this hypothesis since all the coefficients except one are insignificant at the 10 level or lower. It is, however, worth investigating this matter a little more closely. Both the median and modal number of syndicate participants is 2 in each year of the period (Table 6), but the mean has fallen somewhat since its peak at 3.33 in 1989. Correspondingly, the percentage of syndicate deals involving only 2 members has risen from 51% in 1989 to 71% in 1995. What has virtually disappeared off the scene in the last three years is the syndicate involving large numbers of members. Thus, between 1989 and 1992 inclusive there were 15 deals involving 7 or more members, but in the period 1993 to 1995 there has only been one such deal.
Although it must be recognised that a načve model would classify 68.9 % of the deals correctly (syndication = 0), the models proposed here do make a significant improvement and the success in predicting the syndicated deals is notable. Of the two main hypotheses for the existence of syndication it seems that the greater support is given to the view that syndication is a means of improving the selection of projects through joint decisiontaking. However, syndication involves costs and it is clear that most syndicates are small with 81% of them over the period involving 3 or fewer members with 60 % having only 2 members. Syndicates with 2 partners are more likely to involve joint decisionmaking than those with a greater number where there is more likely to be a lead investor who underwrites the deal and then syndicates it down.
There is weak support for the view that higher risk as measured through financial variables increases the likelihood of a deal being syndicated.TABLE 6 Number of Equity Syndicate Members 19891995
|Year||Mean||Median||Mode||2 members (%)||2 or 3 members (%)|
Networking Between Venture Capital Firms
The pioneering work of Bygrave (1987, 1988) analysed the structure of networks amongst venture capital firms in the United States. A similar analysis can be conducted for the venture capital firms in the dataset used here. There are some 190 venture capital firms involved in management buyouts and buyins during 19891995, although the distribution of investments is highly skewed.. To make the analysis manageable the details were restricted to the 18 firms with the largest number of syndicated deals, but all other venture capital firms and the vendor acting as an equity provider have both been included as separate categories.
Table 7 shows the number of syndicated deals involving each pair of companies. The data for all other companies are shown by case 16 and that for vendors by case 20. An obvious feature of the table is the degree to which the firms are interconnected. The density or connectedness of a network can be measured by examining the percentage of pairs with one or more syndications to the maximum number of all such possible pairs. As noted earlier, we follow the approach of Bygrave (1988) in analysing the interconnectedness of the UK management buyout and buyin networks. For Table 6 this figure is 74.2%. Further, if other venture capital firms and vendors are excluded, the value remains roughly the same at 72. %. These values are substantially higher than calculated by Bygrave (1988) for the US reflecting the smaller and more geographically concentrated nature of the UK market.
The results for the measures of centrality and intensity are shown in Table 8, which is sorted by the value of the sum of the weights. The figures show clearly the intensity of the network in that, for example, there are only 3 firms which have less than 10 out of a possible 19 links and leaving aside the other venture capital firms (20), three firms have links with every other firm in this leading group. Further, syndication activities are clearly not limited to the major players and all the firms shown here have a substantial number of links with other venture capital firms. Thus, the data reveal substantial interconnectedness in the British venture capital industry and that a number of players are central to this network.TABLE 7 Networking Within Venture Capital Firms Number of Syndicated Deals by Pair
The interviews covered all venture capital investments not just those in management buyouts and buyins. Respondents to our detailed interviews took the view that syndication of venture capital investments had changed radically over the past decade with changes to historic syndication 'clubs', personality movements in the industry, reactions to restructuring problems in the early 1990s and the weight of money now available for investment. This was seen as being most evident in the recent trend to smaller syndicates at the larger buyout end of the market which was specifically seen to be driven by a number of factors. These included greater funds availability to individual institutions, the increasing need to differentiate oneself from other venture capital firms in order to attract funding, the difficulties experienced in restructuring problem deals where there were larger syndicates, fee income and greater experience in managing risks. However, a number of respondents pointed out that much of the earlier predeal syndication had been replaced currently by the deal leader underwriting the equity and subsequently syndicating a significant proportion postdeal to a limited number of takers.TABLE 8 Centrality and Intensity of Venture Capital Networks
Sum of links
Sum of syndications
Sum of syndications with other VCs
Sum of weights
The type of institution which the lead would be prepared to syndicate to could be seen as driven by forces such as fund requirements and more general considerations. Increasingly the providers of funds insisted that the fund managers should be leaders and initiators of the transactions they were not prepared to support managers who relied on others for deal flow and wanted to see the venture capital firm's investments differentiated from other players in the market and that they should have coinvestment rights themselves. For those that felt the need to syndicate, more general considerations included the size of the transaction and the associated underwriting need, the feeling that a particular partner could bring added value to the company, and that partners should be of similar type and aims.
There was some difference in views about whether the trend to smaller syndications would continue over the next five years. For the majority, syndicates would remain smaller and more homogeneous. However, respondents differed in their perceptions about whether or not there would be pressures to change. Whilst a big failure may provide the impetus to change, some took the view that this was unlikely to happen as leverage was not generally as high now as in the 1980s and that the economic environment was unlikely to change as dramatically as before. Others were less sanguine, expressing concern about the difficulties in picking good deals and adding sufficient value when high prices were being paid.
In the smaller earlier stage segment of the market there was general agreement that the market had developed such that there were very few other players with the appropriate skills with whom one could syndicate. Nevertheless the nature of this segment of the market involves a number of financing rounds over a period with high levels of risk and financial commitment traditional reasons for needing syndication. There was some evidence that syndication patterns in this area could become European orientated.
ł 1997 Babson College All Rights Reserved
Last Updated 03/19/98