Mona M. Lyne, "Parties as Programmatic Agents: A Test of Institutional Theory in Brazil," Party Politics, 11 (March, 2005), 193-216.

First Paragraph:
Scholars have long condemned Brazilian parties as undisciplined and feckless - little more than clusters of particularistic legislators and lacking collective identity (Ames, 1995a, b, 2001; Geddes, 1994; Geddes and Ribeiro Neto, 1992; Hagopian, 1996; Mainwaring, 1992, 1995, 1999; Mainwaring and Perez-Liñán, 1997; Sartori, 1994; Shugart and Carey, 1992). Additionally, it is commonly argued that Brazil's electoral institutions are the chief cause of its weak parties. Recently, however, new evidence has emerged that calls this unanimously held view into question: systematic empirical studies have shown that Brazilian parties' unity in legislative voting has increased significantly in the current period (1989 to the present), as compared to the earlier period of democracy (1945-64).1 These findings have sparked a lively debate centered on some of the most enduring questions in the comparative study of parties: how can we most accurately assess the degree to which parties are meaningful collective actors that offer voters alternative national policy programs? And, what are the key causal forces driving parties to organize around national legislative programs?

Figures and Tables:
Table 1. Seat shares per party (%) in the Chamber of Deputies (1946-1963)
Table 2. Seat shares per party (%) in the Chamber of Deputies (1991-2003)
Table 3. Roll-call voting in Brazilian Chamber of Deputies, 1946-64
Table 4. Roll-call voting in Brazilian Chamber of Deputies, 1990-2002
Figure 1. Roll call voting in Brazilian Chamber of Deputies, Rice Indices (weighted averages)
Figure 2. Roll call voting in Brazilian Chamber of Deputies, index of likeness
Figure 3. Weighted and unweighted Rice Indices roll call votes in Brazilian Chamber of Deputies, 1945-63
Figure 4. Weighted and unweighted Rice Indices roll call voting in Brazilian Chamber of Deputies, 1989&endash;2002
Table 5. Party composition of executive cabinet in Brazil, 1946-64
Table 6. Party composition of executive cabinet in Brazil, 1990-2002
Table 7. Electoral alliances: federal legislative elections for the Chamber of Deputies, Brazil, 1950-62
Table 8. Electoral alliances: federal legislative elections for the Chamber of Deputies, Brazil, 1986-2002

Last paragraph of Conclusion:
In sum, the data suggest that parties' incentive to organize around programs is powerfully influenced by other variables in addition to electoral law. Further study of the transformation from direct clientelist exchange to the indirect delivery of particularistic/programmatic goods appears promising in the quest for a fully general theory of party behavior. While it is beyond the scope of this article, investigation of whether and why politicians are abandoning clientelist, quid pro quo, links to voters for more indirect links through local and national collective goods is a promising approach to explaining these changes in party behavior.