3	About half (between one-third and two-thirds) of the party's support comes from a single institutional sector of society.
4	About half (between one-third and two-thirds) of the party's support comes from two sectors of society.
5	About half (between one-third and two-thirds) of the party's support comes from the government or from more than two sectors of society.
6	Two-thirds or more of the party's support comes from the government or from more than two sectors of society.
7	Two-thirds or more of the party's support comes from party sources, including membership dues and income from party enterprises.

The second code provides an interpretation of the first by indicating the institutional sector or sectors (to a maximum of two) that contribute most to the support of the party. These "sector codes" are detailed below.

Coding Results. It should not be surprising that information concerning a party's source of funds was difficult to obtain. The end results of our search are given in Tables 8.1 a and 8.1 b. While we managed to code almost 70 percent of the parties on BV701, we often did so on the basis of scanty information, as reflected in the relatively low means for AC701. Nevertheless, there was no significant correlation between BV701 and AC701. Consulting the distribution of codes for BV701, we learn that approximately 40 percent of the parties are virtually self-sustaining, drawing two-thirds or more of their funds from internal sources. On the other end of the scale, enjoying far less autonomy of action, about 20 percent of the parties are dependent on a single sector of society for two-thirds or more of their funds.

When a party received the bulk of its support from one or two institutional sectors of society, we identified the sector or sectors in our codes. Table 8.2 tallies all sectors as the first or second most important source of party funds. The influence of business money in party politics across the world shows clearly in the distributions, although labor money lays claim to the allegiance of the next largest group of parties.

Duverger draws a distinction between "direct" and "indirect" party membership, with the distinction hinging on whether membership is taken out as a voluntary act of affiliation with the party itself or whether membership in the party is concomitant with membership in another social organization or organizations. Duverger contends that in indirect parties, "properly speaking, one is not a member of a party but of a social group which belongs as a body to the party" (1963, p. 6). Parties that depend upon other social organizations for their members are considered to be less autonomous then parties that claim members directly, and the extent of autonomy is expected to vary directly with the proportion of direct members among the party faithful.

It might be argued that parties with no membership requirements enjoy conditions that are most favorable to autonomy from other social organizations. On the other hand, one might hold that parties without membership requirements cannot be evaluated according to this dimension of autonomy. This issue has not been discussed much in the literature, and a resolution of the conceptual question will have to await data analysis. In an effort to confront that question with data, we treat the situation of no membership in parties as the most autonomous state.

go to page 93