Horizontal Accountability in New Democracies

O’Donnell starts out by stating that his interest in horizontal accountability stems from its absence and he will mainly focus on those countries that are polyarchies, but have weak horizontal accountability, which in his view are almost all the Latin American countries, some new Asian polyarchies, such as the Phillipines, South Korea and Taiwan, as well as India, and some of the post communist countries – Russia, Belorussia, Croatia, Slovakia, and Ukraine and perhaps also Hungary, Poland and the Czech Republic.
In these countries, most of the aspects of vertical accountability exist, i.e. reasonably fair and free elections, social demands that can be articulated without suffering state coercion, and regular coverage by the media of these demands and apparently wrongful acts of the public authorities. So these polyarchies choose who is going to rule them for some time and they can freely express their opinions and demands. But what is missing, according to O’Donnell, is the liberal and republican components, which is what makes the horizontal accountability weak. This view is based on O’Donnell’s assertion that polyarchies are the complex synthesis of three historical currents or traditions: democracy, liberalism, and republicanism.
He then discusses each of these three traditions before getting to horizontal accountability. Liberal component is the idea that there are some rights that should not be encroached upon by any power, including the state. The republican component is the idea that the discharge of public duties demands careful subjection to the law and devoted service to the public interest, even at the expense of sacrificing the private interests of the officials. For both traditions, the distinction between private and public spheres is crucial, but the implications are different. For liberalism, the area of proper and fuller development of human life is the private one; whereas for republicanism it is the public one, meaning that it is there that the demands of dedication to the public good require, and nurture, the highest virtues. The democratic tradition ignores these distinctions: there may be private activities, but those who participate in the collective decisions are not a virtuous elite but the very same people who may undertake an active private life, and the demos can rule over any matter – it has the full right to decide on whatever issue it deems appropriate.
The values attached to the public and to the private spheres by liberalism and republicanism and democracy lead to diverging conclusions about the directionality of rights and obligations: liberalism attaches defensive rights to individuals placed in the private sphere; republicanism attaches obligations to individuals placed in the public sphere; and democracy asserts the positive right of participation in the decisions of the demos. But there is an important convergence: Democracy (in its equalizing impulses), liberalism (in its commitment to protect freedoms in society), and republicanism (in its severe view of the obligations of those who govern) each in its way supports another fundamental aspect of polyarchy and of the constitutional state that is supposed to coexist with it: the rule of law. The actual effectiveness of the rule of law registers important variations across different polyarchies. 
O’Donnell argues that these three currents have combined in complex and changing ways in the formally institutionalized polyarchies and consequently it is mistaken to postulate a certain dimension as the basic or more fundamental foundation of polyarchy. He is interested in looking within the set of polyarchies and asking what differences exist among these cases and what accounts for such differences. The reason he discusses these three currents or traditions in such depth is because he believes that the relative weight of the liberal, the republican, and the democratic currents exhibits important variations across time and space in polyarchies. One further element is that these currents converged in the constitutions and legislation of the territorially based state. This is important because 1) most of whatever law exists is law issued and/or backed by the state and 2) state bureaucracies are crucial seats of the power resources that come into play when matters of horizontal accountability are at stake. The kind of state we find in a given country will also significantly influence the kind of polyarchy the country is likely to have as well as the ways it is likely to change. Polyarchies are complex and unstable mixes of the four elements – the three traditions and the state. 
In the historical memory of many populations and in the expectations of many actors, the idea of democracy (i.e., polyarchy) has become closely identified with the process of elections, blurring the no less constitutive role that liberalism and republicanism play in polyarchy. O’Donnell argues that this creates problems when, among other things, horizontal accountability is discussed.
O’Donnell defines horizontal accountability as the existence of state agencies that are legally enabled and empowered, and willing and able, to take actions that span from routine oversight to criminal sanctions or impeachment in relation to actions or omissions by other agents or agencies of the state that may be qualified as unlawful. Such actions can affect three main spheres, namely democracy, liberal sphere, and republicanism (another reason why he discusses these three traditions in great length up to this point in the article). Democracy is impinged, for example, by decisions that cancel freedom of association or introduce fraud in elections. Liberal is infringed when state agents violate or allow private actors to violate liberal freedoms and guarantees, such as the inviolability of the domicile, the prohibition of domestic violence and of torture, the right of everyone to a reasonably fair trial, and the like. The third sphere, republicanism, refers to actions by officials most of whom are highly placed in the state or the regime. These actions entail serious disregard for the demands placed on these officials by the republican tradition – that is, to subject themselves to the law and/or give determinate priority to public interests, not their private ones.
For delegative or authoritarian conceptions of political authority, it is this republican dimension of restraint that is the more counterintuitive: Why recognize powers other than one’s own when one is trying to achieve goals toward the public good? And why not benefit personally and bestow favors on one’s family, friends, and business associates while in public office if at the same time one is aiming at some aspect of public good?
Many deficiencies of horizontal accountability arise because republican injections are seen as something which one should at most pay lip service and that one should take into account to prevent damaging consequences. 
For horizontal accountability to be effective there must exist state agencies that are authorized and willing to oversee, control, redress, and/or sanction unlawful actions of other state agencies. These agencies include the classical institutions of the executive, the legislative, and the judiciary; it also extends to various overseeing agencies, ombudsmen, accounting offices and the like.
Effective horizontal accountability is not the product of isolated agencies but of networks of agencies that include at their top courts committed to such accountability.
There are two main directions in which horizontal accountability can be violated. One is the unlawful encroachment by one state agency upon the proper authority of another –encroachment - ; the second consists of unlawful advantages that public officials obtain for themselves and/or their associates – corruption. O’Donnell contends that in the long run, encroachment is more dangerous than corruption for the survival of polyarchy as its systematic utilization will simply liquidate polyarchy. Furthermore, encroachment places a stronger obstacle than corruption to the emergence of the relatively autonomous state agencies acting according to properly defined authority.
In the final section of the article, O’Donnell discusses what can be done to enhance horizontal accountability. He is not very optimistic but offers some “modest and scarcely original suggestions”. These are:
 Give opposition parties that have reached some reasonable level of electoral support, an important role in the directing agencies that are in charge of investigating alleged cases of corruption (caution: there is no guarantee that the opposition is any better or that the government will not ignore and/or deprive them of necessary resources);
- Ensure that agencies that perform an essentially preemptive role, such as general accounting offices, are highly professionalized, endowed with sufficient resources, and independent of the whim of the executive;
- Ensure that the judiciary is highly professionalized, well endowed with resources, independent from the executive and congress, and autonomous in its decisions with respect to both;
- Reliable and timely information is essential;
- Lively and persistent participation of the domestic actors – the media and various social organizations of vertical accountability. Transnational organizations and networks are helpful, too. The effectiveness of horizontal accountability is to a significant extent contingent on the kinds of vertical accountability that only polyarchy provides;
- Individuals, especially political and other institutional leaders, do matter.