Setting parameters for a realistic debate about direct democracy
Direct policy voting is on the increase. In the latest survey, LeDuc (2003: 21-2, 152) estimates that its use extended from around 250 referendums and initiatives from 1961-80, to nearly 350 from 1981-2000 in the countries of the world, excluding Switzerland. In both the American states and Switzerland, policy votes doubled in the last twenty years compared to the preceding period. In many jurisdictions, such as the German Lander, the UK and New Zealand, direct policy voting has now been introduced for the first time.
There is little to surprise us in this trend. In a world where the majority of citizens are better educated, better off and increasingly self-confident, it is natural that they should take the promise of democracy seriously and seek to get their preferences directly enacted into public policy. The ability of democracy to make a ‘necessary connection’ between the two through elections is, as we have seen, its core characteristic. This is what gives direct democracy its driving force and wide appeal in the modern world: there is no better way of enforcing the link between popular preferences and public policy than by voting directly on each policy.
Of course, the groups pressing for direct voting have other motivations too. They feel that their causes - whether to reduce taxes or protect the environment - are so obviously correct that they will get majority support if they can only get them on the ballot and sweep self-serving parties away. So far, analysts have failed to find any clear evidence that direct policy voting favours particular outcomes, either in terms of its immediate outcomes or indirect influence on legislatures from the threat of an initiative. There is some evidence, however, that its presence does bring policy closer to median (majority) voter preferences - which vary of course over time and between jurisdictions (Gerber and Hug 2001: 106; see the further analyses reported in Kriesi in this volume).
As critics have pointed out, sweeping away parties and other mediating institutions brings many undesirable consequences that may lead, in the end, to popular majorities voting against their own preferences and interests. This may result from lack of the essential, if minimal, information about the wider policy implications of referendum proposals that party endorsements provide, or from shifting majorities voting against taxes in one consultation, and for public services in another.
Despite the aspirations of many of its advocates, however, direct democracy does not generally take on an anti-party or non-partisan form. It can be argued that even in the US states, established parties fought back successfully against policy proposals that threatened their central interests, as with tax cuts (Cronin 1989: 205-6). The minority Republicans also built up to their present dominance by exploiting popular initiatives, among other tactics. Elsewhere, established parties dominate referendums, and opposition and emergent parties exploit policy votes to embarrass the government and force their own recognition. Of course, the best way to fight parties is to form an anti-party party, which many proponents of extended participation and popular voting have done (for example, the German Greens and Danish Progress Party).
In terms of actual practice, therefore, direct democracy tends towards either strongly mediated or moderately mediated, rather than unmediated forms. This is hardly surprising, as it tends to take place in party-run representative democracies with a plethora of institutions - governments, parliaments, bureaucracy and courts - overseeing their processes and codifying them along the lines of fair play embodied in general elections. The American experience should not be allowed to dominate discussion, especially since weak regulation of representative as well as direct elections is the norm there. The fact that California is so often cited as an awful warning of the perils of direct democracy is down to its lax regulation, which extends even to its general elections and legislative voting. Switzerland is a much better example of its potential for improving the quality of democracy, which is explored in the next chapter.
Convergence between specific policy consultations and general election practice should not be surprising, since in the modern world they are both about policy. An essential starting point for informed academic debates on the merits of direct democracy should be that so-called representative democracy (really party democracy) is mainly about putting policy packages to electors and following through on them in government.
Our choice between direct democracy and representative democracy should not, therefore, continue to base itself on outdated contrasts between popular policy decision and representative deliberation. Rather, it should characterize itself as being between voting on individual policies and package policy voting. Put this way, it seems much less apocalyptic than has been portrayed. The two procedures cannot be 100 per cent guaranteed against producing different outcomes, but this is far from saying that they will generally do so.
In any case, decisions on the issues involved are probably best arrived at using the different procedures. Where issues are linked to each other, generally through forming part of overarching left-right divisions, decisions on one may well have consequences for the others and so are best voted on as a package to be effected over four to five years. Where issues are more discrete and have fewer mutually interactive effects, they are probably best voted on separately, especially when they do not ‘fit’ in left-right terms and get ignored or totally excluded in general election debate.
Happily, this division of labour seems to be evolving in actual democratic practice. In this sense, the modern expansion in specific policy voting enhances and extends the ‘necessary democratic connection’ between popular preferences and public policy, much more so than threatening and undermining it. For convincing proof of this we now examine the current Swiss experience in the next chapter.