Direct Democracy
Back Home Up Next


1 Direct Democracy

1.1 The Instruments of Direct Democracy; the Referendum, the Initiative and the Recall

Traditional instruments of direct democracy

Traditional direct legislation comes in three different vari­eties: the initiative, the referendum and the recall. In modern times, the oldest of these is the referendum, which often lead almost directly to the later introduction of the initiative and the recall.

The Referendum

The referendum is the people's power to approve or reject legislative acts. It comes in several forms depending on the nature of the legis­lation to which it applies etc.. The referendum may be charac­terized along the following four dimensions:

  • form of legislation:
    • constitutions or constitutional amendments 
    • ordinary statutes, and 
    • fiscal issues
  • initiator(s)
    • the citizens
    • the legis­lature or parts of it
    • the executive branch of government (i.e. the president or governor of a state)
    • the states (in a federation or confederation)
  • advisory or binding
  • voluntary or compulsory

This gives many possible combinations. However, some of these combinations are more common and more important than others.

The term referendum when employed alone, usually means a binding compulsory referendum on ordinary statutes. In the United States or Switzerland, the two countries where direct democracy is most widespread, this term is also used in the more specific sense of a citizen initiated referendum. A citizen initiated referendum is also called a petition referendum since it starts as a petition signed by a group of citizens. However, as the number of signatures surpass a certain  number, the petition ceases to be a petition in the ordinary sense and becomes instead obligatory or compulsory on govern­mental authorities. 

If 50,000 Swiss voters (now amended to 100.000 voters (2001)) petition the govern­ment for a referendum on a federal statute, the government may not refuse them. It is obligated by the Constitution to carry out the referen­dum.

Another common category in the United States and in the Swiss cantons, is the binding compulsory fiscal referendum. In its typical form, the state (or canton) is obligated by the state constitution to let the voters approve  governmental debt increases.

Lastly, many countries require popular approval for cons­titutional amendments (normally in the form of a binding com­pulsory referendum.

Within the proposed model constitution the referendum is more narrowly defined as the people's power to approve or reject acts of the legislature.

The Initiative

The initiative is usually defined as the people's power to approve or reject legislation initiated or proposed by someone other than the legislature. It may be characterized along the following dimen­sions:

  • form of legislation:
    • constitutions or constitutional amendments
    • ordinary statutes, and
    • fiscal issues
  • initiator(s)
    • the citizens
    • the states (in a federation or confederation)
  • advisory or binding
  • voluntary or compulsory
  • direct or indirect

The usual form of the initiative is as a citizens' initiative or as a voters' initiative. (The two terms have the same meaning). This is legislation (whether ordinary statutes or fiscal issues) proposed by a group of citizens through a petition. As with the petition referendum, once the petition itself has gathered enough voter support in the form of signatures, it becomes both com­pulsory and binding.

If the initiative is direct, it is placed on the ballot at the next election. If the initiative is indirect, the legislature gets a chance to enact the proposal. Only if it fails in the legislature is it placed on the ballot.

At the federal level in Switzerland, only the constitutional initiative may be found. Many American states on the other hand, allow both the constitutional and the legislative initiative.

The Recall

The recall is the people's power to force a public official out of office. It can be found at the state (and local) level in the U.S. and at the cantonal (and local) level in Switzerland.

As for the other instruments of direct democracy it comes in several forms:

  • application
    • all public officials whether elected or not,
    • all public officials except judges
    • all elected public of­ficials including judges, and
    • all elected public officials except judges
  • direct or indirect

The term "direct recall petition" means a petition that leads directly to a new election. Indirect petitions only lead to a new election if they are successful, i.e., if the official in question is actually recalled at the polls.

Besides these distinctions, there are numerous other variations. Sometimes, the recall may only be employed once during an official's term, in other cases the official gains 6 months' or 1 year's immunity against new recall attempts, and sometimes the defeat of the recall implies reelection for a new term.

The Referendum, the Initiative and the Recall within the model constitution

The model constitution employs all the traditional instruments of direct democracy; the referendum, the initiative and the recall. The initiative and the referendum is used in order to secure popular sovereignty and in order to foster competition between governmental units. Equally important is a method for citizen-imposed citizen-adjusted taxation more fully described in the chapter on confederate or central taxation. The recall plays a more subor­dinate role. For a description of the referendum, the initiative and the recall see above.

1.2   Principal features of direct democracy

The principal features of direct democracy is its ability: to limit the influence of pressure groups, to unbundle spending and legislative decisions and to secure competition between govern­mental units through proper arbitration.

Limiting the influence of pressure groups

A principal feature of direct democracy is that it limits the influence of numerically small but politically powerful pressure groups.

The direct vote of the citizens reduces the influence of special interest groups by automatically weighing the interest of the smaller group against that of the majority. In order for a proposal to gain a majority in a referendum it has to offer advantages to a much larger cross-section of the population. In all likelihood it has to be a positive sum proposal. A positive sum proposal means that the sum of all benefits exceeds the sum of all costs when measured across the society as a whole. Ordinary legislation influenced by smaller pressure groups, on the other hand, are often negative sum proposals where the cost to the majority are larger than the benefits accruing to the smaller pressure group.  A subsidy for instance, may be highly profitable for the smaller group receiving the subsidy, but costly to the society as a whole due to costs associated with administrating and implementing the proposal.

Referendums and initiatives work by dramatically reducing the direct cash costs and the indirect "time and effort" costs of influencing government decisions. Political influence in most representative systems comes through extensive long-term lobbying. Such lobbying is expensive in terms of time and effort, and it only pays to engage in it if the potential pay-off is substantial.  In most cases it will not be rationally cost effective for individuals to actively oppose governmental programs as the cost of opposition (in terms of time and effort) exceeds the benefits of lower taxation. Direct democracy reduces the cost of opposition to the time and effort required to put a ballot in the ballot-box; an enormous improvement over having to lobby legislators directly.

Furthermore, with the introduction of electronic automated voting either through the internet or phone based or similar systems, the cost in terms of time and effort can be further dramatically reduced.

Direct democracy and direct decision making costs

One objection to widespread use of direct democracy concerns its alleged high direct costs. According to Kendall and Louw (Kendall, 1989, page 135), the Swiss Federal  chancellery estimates the costs of a national initiative combined with a federal counterproposal to about 1 Swiss franc per voter. Even when special ballots have to be held to decide single issues, the costs are modest. In California such a special ballot was held in 1973. It cost the state about USD 20 million, or about 80 cents (USD 0.80) per capita. (Walker, page 93).

In addition to the direct costs incurred by the government, comes the costs associated with launching an initiative. In Switzerland this cost is estimated to at least one franc per petition signature (Junker, page 122). In California initiative campaigns cost several million dollars. In per capita terms however, these costs are still marginal, which is why this method of making decisions is so effective. Even if we assume that the Swiss spend a few million francs (everything included) on national issues every year, this has to be compared with a Swiss federal budget of about 23 billion francs (1985) (Junker, page 40).

The mismatch between the resources allocated by the government, and the input on allocation allowed by each citizen may also be illustrated in another manner. While in the developed world from one third to two thirds of the total resources of the country are allocated through the public sector, ordinary people in most of these countries, are allowed decision making powers only once or twice every four or five years. (I am here disregarding direct lobbying by individual citizens as a practical alternative.)

At the personal level, on the other hand, we are constantly incurring decision making costs as we try to weigh the relative advantages of everything we buy from tooth paste to motor vehicles and homes. Since the cost of each individual decision is much lower in the private market sector, decision making costs including incon­venience and allocated time is much higher for each dollar, franc or € each citizen spends as a private individual than for dollars, francs or €s spent via the public sector. These higher decision making costs in the private sector are, however, more than counterbalanced by the fact that funds are allocated according to each individual's preferences thus contributing to a much higher subjective benefit or utility. Thus there is ample room for governments to increase total direct decision making costs, i.e. direct costs associated with ballots etc., and thus reduce the cost of influencing each individual governmental decision and bring about a much higher total benefit (utility) both on the grounds of better allocational efficiency (giving people what they want) and improved operational efficiency (fewer negative sum proposals).

Anecdotal evidence; taxes

The most potent form of direct democracy,  voters' initiatives, is today practiced primarily in Switzerland and at the state level in the United States.

It  is well known that Switzerland has one of the lowest tax rates in Europe. Among the European members of the Organization for Economic Cooperation and Develop­ment (OECD), for instance, only Portugal and Turkey have lower total taxation as a percentage of GNP. Other European OECD members at a comparable level of income and development with Switzerland have much higher levels of taxation. In fact, comparable in this respect is not so easy to define as the Swiss also have the highest incomes. The Swiss position can pretty much be summed up by saying that the Swiss enjoy the highest income and the lowest taxes of Europe. 

Table I Swiss and European taxes


Total tax receipts as % of GDP

Highest rate central gmnt. income tax




United Kingdom



European Com­munity average



OECD average



            Source: OECD in Figures June/July 1991

In America, the most famous of all initiatives was Proposition 13 in California in 1978. In spite of opposition from both political parties and most public officials, the voters decided to cut property taxes by 57%, from USD 12 billion to USD 5 billion and restrict increases in the future.  Similar proposals were later enacted in other American states. (Kendall, Frances: Let the People Govern page 139)

The following table of state tax rates gives further evidence of the correlation between democracy and taxes.


Table II Direct democracy and tax rates

Y = yes, state has provisions for legislative citizens' initiative

N = no, state does not have provisions for legislative citizens' initiative

Highest in­dividual income tax rates

Highest sales tax rates


New York



West Virginia










New Jersey


Rhode Island









No individual income tax

No sales tax




South Dakota














New Hampshire







Sources:         tax rate information from Wilson, James Q., page 644

             information about state initiatives from Butler, David, pages 71-72

Remarkably, only one of the high tax states (Washing­ton) has a provision for voters' initiative.

Similar evidence is found in the listing of states without income taxes. Six of seven states without an  individual income tax are "initiative" states, while only one (Texas) is not. The results for states without sales tax (5) are divided as close to the mat­hematically correct distribution as possible, with 2 "initiative" states and 3 non-initiative states. (24 U.S. states or territories have implemented the initiative.)

Even though more comprehensive statistics are less clear-cut,  the above evidence taken together with the history of the American tax revolt in the late 1970s and early 1980s and the Swiss ex­perience, substantiates a correlation between the more potent forms of direct democracy and  lower relative taxes/tax rates and less governmental interference.

Other evidence seem to suggest that there may also be a correlation between the degree of democracy and centralization, i.e. the less democratic, the more centralized.

Unbundling spending and legislative coalitions

By its very nature direct democracy and especially the initiative, tends to unbundle decisions since it enables individual citizens to vote separately on each issue.

Rep­resenta­tive political systems on the other hand tend to bundle political decisions. As an individual you may like a particular candidate's views regarding taxes, but be opposed to the same candidate's views on abortion. In a representative system, however, you have to take the whole bundle. As an individual voter you are not able to separate the two issues.

Bundling tends to reduce the general welfare. Any system that enables unbundling is thus a superior system. Direct democracy allows unbundling and is thus superior to a purely representative system.

Parliamentia and Democratia, an example

Let's suppose we have two identical states: Parliamentia and Democratia representing respectively a parliamentary (rep­resentative) system and a semi-direct system. ( Semi-direct democracy: democracy encompassing both a representative system (an elected legislature) and direct legislation by the citizens. Semi-direct democracy can be found in Switzerland, both at the federal and the local level, and in many American states.) In both countries a proposal to spend 100 million currency units (CU) on a national tiddlywinks team (proposal A) and another proposal (proposal B) to spend 100 million currency units on bubble gum for school children is introduced.


In Parliamentia the leader of the Leftright party (the sponsor of bill A) confers with the leader of the Rightleft party (the sponsor of bill B). The Leftrights hold 25 parliamentary seats, while the Rightlefts hold 26 seats. Together they hold 51 of the total 100 seats. They quickly agree on a quid pro quo, if the Leftrights vote in favour of B, the Rightlefts in turn will vote in favour of A. This is done, and both measures pass though each counted separately is opposed by about 3/4 of the MPs.

Peter Individual, a parliamentarian of the Rightleft party briefly considers voting according to his convictions, but realizing the risk of being censored by or expelled from the party (by that eliminating his chances of reelection to Par­liament), he decides to toe the party line

Though the leader of the Leftrights could theore­tically renege on his promises after the approval of A, he doesn't do so as he realizes that he needs a similar gentlemen's agreement with the Rightlefts for the approval of proposal F, to be decided in Parliament the following week. (Here he is prepared to offer his support for G). -

In other words as the order will be reversed the following week he has to keep to his part of the current agree­ment.


In Democratia, on the other hand, the proposals are to be decided directly by the people. The leaders of the Leftrights and the Rightlefts announce that they have made a deal. They urge their supporters to back both measures. However, since voting is secret, neither leader can identify party members or citizens that don't heed their admonitions, and they certainly cannot expel anyone from the country. Thus the citizens are free to vote according to their individual convictions. The citizens are also a little bit piqued about the way the politicians have taken it upon themselves to tell the people what to do. Thus both measures are heavily defeated.

The following year, however, the leaders of the Leftrights and the Rightlefts decide to submit a single proposal (A&B), in­corporating both measures, to the public. By simple cal­culation  they figure such a joint proposal will be favored by 51% of the voters.

Unfortunately, Truespeak, an independent citizens' action group sponsors a petition drive for a proposal to withhold funds from the tiddlywinks team (A-negative). Another group, People­power, sponsors a drive to withhold funds from the bubble ­gum project (B-negative).

At the polls, A&B receives 51 % of the votes, A-negative receives 75%, while B-negative receives 74% of the vote. Since conflicting measures are decided according to the number of votes; the 100 million currency units support for the tiddlywinks team (A) is neutralized by A-negative (51% of the vote versus 75%), while the bubble gum project is neutralized by B-negative (51% versus. 74%).


As the example shows, in representative assemblies, proposals may pass even though they do not represent majority opinions. This happens when the 25 representatives favoring proposal A combines with the group of 26 representatives favoring proposal B to form a majority of 51. This tendency is most obvious when it comes to ap­propriations, but it also takes place when it comes to legis­lation. The net result is a steady increase in the size of government despite the opposition of most voters and politicians.

Such alliances to push up spending or enact legislation are not possible however, when decisions are returned to individual citizens. Since votes are secret,  the leaders of each faction have no means of enforcing agreements. There is no stick that prevents individuals from voting according to their own convictions. 

Popular votes as the arbiter between governmen­tal units

As the proposed model constitution creates a multi-governmental entity, it has to provide a mechanism for resolving conflicts between subunits. I am proposing that this power to arbitrate between governmental units is devolved directly on the people. This method of ultimate arbitration is the only one that ensures competi­tion between governmen­tal units in fulfilling the needs of ordinary citizens.

1.3   Summary

There are 3 forms of direct legislation:

  • the referendum, often referred to as the people's veto power,
  • the initiative, which gives the people the power to propose and enact legislation, and
  • the recall,  which gives the people the right to call a new election

The referendum and the initiative limit the influence of pressure groups by dramatically reducing the cost of opposing narrow pressure groups. The referendum and initiative also improves the allocation of resources by un­bundling spending and legislative decisions.

Popular votes also may be used as the ultimate arbiter in cases of conflict between representative bodies and/or branches of government.

Evidence from both the United States and Switzerland suggest that a greater say for the people may lead to lower taxes and a smaller more efficient public sector.

Revised: 2003-12-14

Copyright © 1991-2003 John F. Knutsen

All rights reserved