In Portugal, the venture capital is a financial product relatively recent, which sector had a great increment between 1989/90 both in the number of firms in activity and in the available funds to invest.

Nowadays, this sector is on a critical state because the existent funds are mostly stately held, showing a significant lack of interest from the private investors, on the contrary to what happens with the rest of the European countries, not to mention the U.S.

However, when we analyze the legal and fiscal background to which the Venture Capital Societies are subjected, we see that these entities are not contemplated with any special advantage in fiscal terms, contributing, according to those who are connected to the sector and to the examples of other countries, to the weak "apport" of funds, from the economical, financial and private Groups, to this activity sector.

2. THE BACKGROUND

The statistical study of the factors conditioning the attitudes of Venture Capital managers to the suggested proposal of modification, in our inquiry, becomes a vital factor to the understanding of the presented answers.

We considered as possible determinants to the answers of the inquiry three levels:

So, we’ll try to find cases where firms with similarities at the level of one or several of the dimensions studied here have presented similar attitudes to the modification of the present legal and/or fiscal framework, and, at the same time, showing a distance from the answers of the other firms. That means, it will be tried to find groupings of firms similar among each other that have answered to the inquiry on a similar way.

This way, it will be possible to establish relations between the dimension, profitability or kind of the firm with its attitude to the fiscal and legal areas.

3. STATISTICAL METHODOLOGY

Before entering the statistical analyses, we must refer the base of the inquiry formulation, according to the following point.

3.1 Inquiry Construction

The inquiry has been built having as a basis legal and fiscal limitations detected during the research, the interviews made as well as the claims presented by the APCRI (Portuguese Association of Venture Capital Corporations) on the last few years, with the objective of perceiving the sensibility of the Venture Capital Managers to each of the presented factors.

The statistical analysis was made with the resource to the factorial analysis technique in which we started by identifying the more important attributes of the fiscal (F1,... F7) and legal (L1,... L6) changes which are presented on the schemes (I - II).

The group of Venture Capital Managers evaluated those attributes by answering to a group of closed questions. The codification of the answers consisted on the attribution of values to each one of the possibilities, to permit the computer treatment of the information, according to the following ordinal scale.

more important = 1

important = 2

less important = 3

little importance = 4

On these scales, what can be measured are the attitudes and preferences of the inquiries in face of the uni-dimensional variables presented in the inquiry.

Next to the transformation of the information in measurable data, we counted the answers and crossed the variables obtaining statistical indicators, namely, medians, which allowed us to find the central trend of the named managers in face of each one of the alterations to propose.

3.2. Results of the Inquiry

We’ve obtained a high number of answers, 13 in 15 obtainable, allowing us to reduce the magnitude of the bias and, consequently, increasing the results precision.

We now present the answers on the following Figures I-II.

Figure I - Answers Table - Fiscal

Inscription:

We can say, through the present study that the Venture Capital managers consider as more important to the sector the reinvestment tax credit, the extension of the fiscal benefits, previewed in the investment tax credit regime to VC corporations, and the creation of a 100% tax credit to avoid the dividends double taxation. They also find important the extension of the time limit of capital gains tax exemption prescribed in the Article 23 of EBF, as well as the attribution of direct benefits to the Stockholders.

The managers don’t consider specially relevant the existence of a fiscal transparency regime on capital gains tax to pay, nor the readjustment of the Pro-rata on VAT.

In what concerns to the questions put in the inquiry to the modification of the present legal framework, we can summarize the following results according to the answers Table we now present:

Figure II - Answers Table - Legislation

We can see that the Venture Capital managers consider as very important to the sector the flexibility of Services Supply, the elimination of the temporal restrictions to the realization of investments and the enlargement of the time limit to the reposition of the limit on the active operations.

It shows to be important that the Accounting of the Venture Capital firms begins to follow the Accounting Official Plan rules and not the Accounting Plan to the Banking Sector. They also consider as less relevant to the sector the permission for Venture Capital firms to participate in the holdings that participate in societies which target covers the mediation over real estate and real estate societies as well the increase of the minimal parameters of admission in the main Stock Exchange in order to improve the secondary capital market.

3.3. Association Tests

The description of the sample using central trend measures can offer us a very clear idea of the population which is pretended to know, specially if the sample is representative of the universe, as this one is.

A simple bi-variacional analysis can provide us the most useful means to understand the spread of manager’s opinions, specially if we investigate the relations between the type, dimension and profitability of Venture Capital firms and the opinions of those company’s managers. The bi-variacional methods are applied in an early stage of the data exploration loop, taking us, later, to an investigation of the relations among a sequence of variables using a multivariacional methods.

Thus, we used central trend measures to try to find relations between the analyzable dimensions (type, dimension, and profitability) and the opinions of the Venture Capital managers

3.4. Cluster Analysis Technique

Beyond the simple statistical indicators mentioned above we can follow more sophisticated techniques of quantitative analysis of information, as it is the case of the Cluster Analysis Technique. This technique uses a number of methods and algorithms which group a multivariated sample in clusters, that means, homogeneous groups.

However, it is important to say that the clusters suggested by these methods are highly dependent of the sample variations and the measure errors in the observations, for what the choice of a certain number of clusters may not result directly from the algorithm, compelling to the introduction of subjective factors.

For these factors, the cluster analysis (according to Morrison - 1990), cannot be considered as a precise statistical tool, and should be applied with care and with the support of other informations about the existent samples.

However, we applied this technique in the quantitative analyses of our sample considering the following variables:

These variables were analyzed considering the following dimensions:

The clusters creation process could have been made on a rudimentary way, looking for in a matrix the minor distances and grouping the observations in clusters according to this procedure.

However, we followed in our work a formal algorithm - Average Linkage (between Groups) - that combines the original set of N points hierarchically in clusters according to the following steps:

The distances addicted in line and column, for the generic point (G,H) are determined according to the expression d(g,h),i=min(dig,dih) i¹ g,h.

The second cluster is constituted by two variables including the cluster determined on the first step, which have the minimal distance.

Next we form a new matrix of distances (n - 2)X(n - 2) and we locate the minimal distance value to form the third cluster. The algorithm continues, on this way, until all the observations have been grouped on a hierarchy of clusters.

The hierarchy of clusters may then be shown by a graphic called "Dendrogram".

4. RESULTS OF THE STATISTICS TREATMENTS

We will now present some of the results originated from the application of the statistical method to the answers of the inquiry.

4.1. Descriptive Statistics

The answers have been divided in two groups, which were analyzed separately. One of the groups is connected to the questions about the fiscal background while the other is connected to the legal background.

We proceeded to an analysis of the medians for each firm, one by one, in order to discriminate those that are more favorable to change.

So, we concluded that, relatively to the fiscal area, the companies which are more favorable to the change are the firms 1, 4, 10, 15, while the less favorable are 2, 7 and 9.

On what concerns to the legal area, the companies more favorable to change are the 4, 11, 13 and 15, while the less favorable is the 5.

Considering the two aspects together, we see that the firms more favorable to change are the 1, 4, 10 and 15, while the less favorable are the 2, 5, 7 and 9.

4.2. Bivariacional Analyses

We can build an association matrix by analyzing the firms two-by-two and verifying the degree of similarity on the given answers.

We found several associations of companies that answered on the same way to some of the questions, so, we will only present the more relevant results:

Concerning to the fiscal area:

• the publicly held companies with negative net incomes (1, 3, 4) had the same answer to the questions F1, F2, F3 and F7
• the companies 2, 11, 13 and 15 which have a minor dimension relatively to the average had the same answer to the questions F5, F6 and F7;

Concerning to the legal area:

• the companies 5 and 9 which have positive results, had the same answer to all the answers except for the L1
• the companies 2, 5 and 6 (publicly held) had the same answer to the questions L3, L5 and L6
• the companies 13 and 15, which are two of the smallest and are privately held had the same answer to the questions L1, L2, L5 and L6.

This analysis allows us to investigate relationships between the managers attitudes in face of the legal and fiscal background change and the considered dimensions (profitability, type, and dimension), but it doesn’t allow us to check if there are relations of cause-effect between the kind of firm in question and the answers given by the managers of these societies. For that, we have to use the cluster analysis technique.

4.3. Cluster Analysis

The results obtained from the statistical treatment of data suggest that, in the case of the questions about the fiscal area, we can find four clusters containing the following entities (see appendix 1):

• Cluster 11, 3, 4, 10, 12, 15;
• Cluster 22, 11, 13;
• Cluster 35, 6, 9;
• Cluster 47.

On the first cluster we see that, in its majority, the companies consider as more important the F1, F2 and F5, and consider important the F7.

On the second cluster we see that the companies give little importance to F2, and consider as more important the F5, F6 and F7.

On the third cluster we see that, in its majority, the companies give little importance to F1, F2 and F4, and consider F7 important.

On the fourth cluster comes the company 7, isolated, because it considered as more important the F3 and F4, and gave little important to F5 and F6.

In none of these groupings can be found relevant correlations concerning to the type, dimension or profitability of the companies, except for the cluster 2, where all the companies are in the second half of a table sorted by a decreasing order of dimension, that means, among the smallest ones.

Considering the proposed questions at the legislation level, the results indicate the existence of five clusters containing the following entities:

• Cluster 11, 4, 11, 12, 13, 15;
• Cluster 25, 6, 9, 10;
• Cluster 37;
• Cluster 42;
• Cluster 53.

On the first cluster we can see that, in its majority, the companies consider L3 and L6 more important.

On the second cluster we can see that the companies consider L2 as more important and L5 as of little importance.

On the other clusters appear isolated companies, because they gave answers very different from the answers given by the other companies; the company 2 gives few importance to L2 and L6; the company 3 gives more importance to L1 and L5 and gives little importance to L4; finally, the company 7 gives little importance to L4.

It is also possible to understand a certain gathering of the companies with negative net incomes on cluster 1, while in the cluster 2 all the companies have positive results.

All the companies that stayed isolated in clusters are public.

4.4. Analysis of the results obtained