Anchoring the Portuguese Voter: Panel Dynamics in a Newer Electorate

• DOI: 10.1177/1065912909355715
• Author: Marina Costa Lobo,Michael S. Lewis-Beck
• Date: 01 June 2011
• Published date: 01 June 2011
• Subject Matter: Articles

Political Research Quarterly

64(2) 293 –308

© 2011 University of Utah

Reprints and permission:

sagepub.com/journalsPermissions.nav

DOI: 10.1177/1065912909355715

http://prq.sagepub.com

Anchoring the Portuguese

Voter: Panel Dynamics in

a Newer Electorate

Michael S. Lewis-Beck1 and Marina Costa Lobo2

Abstract

While Portuguese democracy is no longer so new, its national postelection surveys are, with the first in 2002. On the vital

question of what provides the voter a social-psychological anchor, initial evidence gave the nod to party identification

over ideological identification. However, party identification was poorly measured, data were cross-sectional, and the

models single equation. Fortunately, panel studies are now available for the 2005 legislative and the 2006 presidential

elections. Estimating dynamic, multi-equation models with two-stage, instrumental variable regression procedures

establishes the preeminence of ideologically driven voting. Furthermore, ideological identification appears composed of

a unique pre-democratic component, in addition to the more usual social, moral, and economic elements.

Keywords

party identification, ideology, panel studies, voting behavior, Portugal

In democracies everywhere, there is generally some long-

term force that anchors the electorate, giving the individual

voter a “standing decision” to rely on (Key 1966). In the

United States, party identification serves as that behav-

ioral anchor (Campbell et al. 1960; Lewis-Beck et al.

2008). The party identification model has been exported

to Western Europe, with mixed results (Budge, Crewe,

and Farlie 1976; Dalton, Flanagan, and Beck 1984;

Franklin, Mackie, and Valen 1992). The debate in France

has been especially sharp, with some arguing that the

French electorate binds itself through party identification,

others through ideological identification. (On the former

view, see Converse and Pierce 1986; Pierce 1995; Evans

2004. On the latter view, see Fleury and Lewis-Beck

1993; Haegel 1990; Mayer 1997.) Certainly, the power of

ideology has been proposed for other established Western

European democracies as well. (For a general treatment,

see the classic, Inglehart and Klingemann 1976.) The

Dutch electorate, in particular, seems to have ideological

identification as their exclusive anchor (van der Eijk and

Niemöller 1983, 1987, 1994).

But what of the newer democracies of Western

Europe—Spain and Portugal? Are such long-term forces

operative? If so, does ideology count for more than

party, or vice versa? With respect to Spain, a growing

body of work suggests the overwhelming role of left-

right ideology as an electoral anchor (Calvo and Montero

2002; Lancaster and Lewis-Beck 1986; Maravall and

Przeworski 2001; Torcal and Medina 2002). What about

Portugal, the object of study here? Prevailing opinion holds

that party dominates ideology as an electoral anc hor.

Portugal has a semi-presidential Constitution and the

electorate directly elects the President as well as the

Parliament’s deputies. The same four parties have domi-

nated the electoral scene since 1974, together winning

over 90 percent of the votes in all legislative national

elections save one (1985), despite the fact that Portugal

has employed a proportional representation electoral

system (d’Hondt method).

Below this surface pattern of stability, however, there

has been a fair amount of movement. Between 1987 and

2005 there was a shift toward fewer and larger parties, a

trend reversed in 2009. On average between 1987 and

2005 the two main parties polled over 80 percent of the

vote. But in 2009 the main parties—on the left the PS

(Partido Socialista) and on the right the PSD (Partido

Social Democrata)—gathered only 66 percent of the vote.

By way of contrast, in 2009 the two smaller long-standing

parties—on the left the Communists (Partido Comunista

Português) and on the right the CDS (Centro Democrático

Social)—also halted their apparent decline. Furthermore,

a new party of the left, the Bloco de Esquerda, with its

first elected MPs in 1999, has become a fully consolidated

1University of Iowa, Iowa City, IA, USA

2Universidade de Lisboa, Lisboa, Portugal

Corresponding Author:

Marina Costa Lobo, Instituto de Ciências Sociais, Universidade de

Lisboa, Av. Anibal Bettencourt, 9, 1600-189 Lisboa, Portugal

Email: marina.costalobo@ics.ul.pt

294 Political Research Quarterly 64(2)

party on the left, winning 9.9 percent of the voters in the

2009 election. As well, aggregate levels of party identifi-

cation have dropped substantially, while those of

ideological identification have not.

Thus, despite the apparent resilience of the party

system, it has been experiencing an underlying fluidity.

Still, these are macro findings. With respect to the scope

and strength of party (and ideology) at the individual,

micro level, we have relatively little systematic evi-

dence. While contemporary democratic Portugal held its

first legislative election in 1976, the first comprehensive

scientific national election survey was not conducted

until 2002. (Of course, Portugal and the Portuguese elec-

torate had been the object of social science research prior

to that date. First, a few political surveys were carried

out in 1978, 1985, and 1993; see Freire et al. 2005. Also,

Portugal has been included in the Eurobarometer sur-

veys since 1986. However, that Eurobarometer coverage

has been sketchy, especially with regard to the measure-

ment of party identification. Still, these data, and the

uses to which they have been put, have provided early

valuable insight into Portuguese electoral behavior;

Bacalhau 1994; Bruneau and MacLeod 1986; Gunther

and Montero 2001.)

What is the party versus ideology balance in the psyche

of the mass electorate? In the following, we review the

available literature, focusing on issues of measurement,

model specification, and research design. The question of

measuring ideological identification and party identifica-

tion has been troublesome, as has the placement of these

variables in a properly specified voting model. The heart

of the specification difficulty is the potential of reciprocal

causation between party and ideology. Of course, such

causation issues are difficult to solve in the cross-sectional

analyses carried out thus far. Here we deal with these

difficulties, employing unique panel data from the 2005

legislative election and the 2006 presidential election. Our

multi-equation systems, variously estimated with two-stage,

regression procedures, all point to the strength of ideology

over party.

Ideological Identification and

Party Identification: Meaning

and Measurement

In important ways, left-right ideology appears to play a

critical role in the mass politics of contemporary Portugal.

Voters easily place themselves on the widely agreed upon

left-to-right opinion scale, and that placement correlates

highly with vote choice (Calvo et al. 2007; Gunther and

Montero 2001; Jalali 2004, 2007; Lobo 2006). However,

establishment of the issue contents of that ideology has

been somewhat illusive (Freire 2004, 173-75, 2007; Freire,

Lobo, and Magalhães 2004; Jalali 2004; Lobo 2006). The

first postelection national survey of 2002, upon which

much past work bases itself, included few items useful

for assessing the contents of ideology. In contrast, the

2005 postelection national survey at hand contains a

large number of potentially relevant items, offering a

unique opportunity to record the main components of left-

right placement in Portugal.

It is helpful to begin with consideration of three such

components, commonly considered in the Western Euro-

pean political behavior literature: social factors, moral

values, and economic attitudes (see, respectively, Ingle-

hart and Klingemann 1976; Inglehart 1977; Knutsen

1995; and for a current discussion of the salience of left-

right ideology in such democracies, see Hellwig 2008). In

addition to these established components of left-right

self-placement, we would add one of special relevance to

Portugal—pre-democratic attitudes. The country houses a

relatively new democracy, where the legacy of the author-

itarian regime continues to shape the understanding of

politics. For the Portuguese case, the context of the demo-

cratic founding actually leads to a double legacy, one

derived from the authoritarian regime, the other from the

transitional struggle to democracy, 1974-1976 (Pinto

2006). A large part of the current electorate, not to men-

tion the political elite, was at least fifteen years old (i.e.,

born before 1960) under this double legacy.

The bundle of pre-democratic attitudes acquired during

this decisive political socialization helps separate left from

right in Portugal today. Three characteristics in particular

demand attention. First, the anti-democratic nature of the

New State, especially its distaste for parties, viewed as a

wholly negative phenomenon of political life. There is now

overall support for the democratic regime among the Portu-

guese public (Freire 2003). Nevertheless, it is plausible to

suppose there remain residual differences on the proper

place of political parties, when comparing left to right.

The second pre-democratic characteristic for consid-

eration is the consequences of decolonization. The initial

objective of the military officers in the 1974 coup was to

end the colonial wars. In the aftermath of the Revolution,

war veterans felt shunned by these left-wing military

officers, who signed an agreement for colonial indepen-

dence. Indeed, membership in war veterans groups had

traditionally been promoted by the authoritarian regime,

from the mid-1960s onwards, as a way to forge support

for these increasingly unpopular colonial campaigns.

More recently, the small conservative right-wing party,

the CDS, has defended the war veterans’ interests, call-

ing for added benefits for them. Thus, membership in

veteran associations should be considered an organiza-

tional anchor predicting right-wing leaning.