Political fragmentation is regularly considered to be a relevant factor affecting the outcomes of fiscal policy. It is assumed that a fragmented fiscal policy process leads to higher expenditures and subsequently in a higher deficit. But “hard evidence” in support of this hypothesis is rare, and conclusions can only be drawn very carefully. The present paper gives an overview of the existing literature on the role of (size) fragmentation and its effects on fiscal policy outcomes. The main focus is on the work contributed by Yianos Kontopoulos and Roberto Perotti, who used a panel of 20 OECD countries over 1960-1995 for their analysis of government fragmentation.
With relatively high state expenditures and shrinking revenues, many governments throughout the world are facing public debates about fiscal policy adjustments. Cutting subsidies and lowering taxes have become common demands. In countries where political power is shared among high numbers of parties, politicians are often heard saying that such steps would be easier to take if they were in a single-party government and would not be forced to find compromises with their respective political partners. It is a commonly held view that decision-making processes become less efficient if the number of participants increases1. This is commonly known as fragmentation:
“ [...] fragmentation arises when several agents or groups participate in the fiscal decision-making process, each with its own interests and constituency to satisfy ” 2
Fragmentation is generally considered to consist of two components3: size fragmentation (the number of decision-makers) and procedural fragmentation (the structure of the budgetary process). The research carried out on the effects of procedural fragmentation did not deliver strong significant results. Implications, if any, can hardly be drawn. An explanation for this lack of significance of procedural fragmentation could be that changes to the structure of budgetary procedures are implemented rather rarely over time. For this reason, the research discussed in this paper mainly explores the role of size fragmentation.
As more interests are represented within a fragmented fiscal policy process, one would expect to see a rise in expenditures and a higher budget deficit:
“ To participate in the majority, each group demands a share in the budget; as all groups do this, the end result is a high level of expenditure or a large deficit. ” 4
The assumption that the number of decision-makers is positively related to the size of the budget deficit is expressed by the common pool model5, in which government assets are regarded as common property. According to this theory, the incentive for fiscal discipline decreases for higher numbers of participants. The rationale behind this is that if one decision-maker limits expenditures for his department, the other decision-makers will appropriate the resulting savings. As a result, aggregate expenditure is expected to be higher than it would be in a less fragmented budgetary process.
Another idea associated with fragmentation is the internalization of excess costs. The rationale is that the taxation associated with aggregate expenditure will be distributed among the policymakers. Hence, the anticipated cost of additional expenditure falls for the individual decision-maker and the process becomes more fragmented:
“ Fragmentation of fiscal policy-making is the degree to which the costs of a dollar of aggregate expenditure are internalized by individual fiscal decision-makers. ” 6
Accordingly, if the costs of aggregate expenditure are not being internalized, the fiscal policy process is fragmented and vice versa. Lack of internalization therefore leads to fiscal profligacy.
The reasoning behind the fragmentation hypothesis outlined so far would probably appear to most people as something they would instantly agree on. But is it possible to find empirical support for this view? Does fragmentation have an impact on fiscal outcome? By examining research on the topic, the present paper attempts to give an answer to this question. In doing so, it will mainly focus on the work of Kontopoulos and Perotti.
The paper is structured as follows. First, a brief introduction to the relevant literature will be given. The second part will then focus on the research conducted by Kontopoulos and Perotti and explain their model. The third chapter offers a summary of the basic results. After that, the robustness of these results will be examined in order to determine to which extent they are sensitive to changes in both modeling and the data used.
2 Research Overview
The first to investigate the relationship between fragmentation and fiscal outcome were Roubini and Sachs7. They assumed that the type of government influences fiscal policy outcomes. This assumption is usually referred to as “weak government” hypothesis in research literature. To test their hypothesis, Roubini and Sachs created a variable measuring the level of political cohesion. Depending on the type of a particular country's government, the corresponding power dispersion index (POL) assigns a score from 0 to 3. A single-party majority government would be assigned a score of zero and a majority coalition government of two or three parties would be assigned a score of one. Scores of two and three would be assigned to majority coalitions of four or more parties and minority parliamentary governments, respectively. By doing so, they rank governments according to what is supposed to be their “political strength” - hence the term “government weakness”. Using data for 20 OECD countries over the time period from 1960 to 1985, Roubini and Sachs found a significantly positive correlation of the POL index and budget deficits. Disagreeing with this conclusion, Edin and Ohlsson8 soon criticized that, due to the ranking method employed, minority governments would have a greater impact on the indicator than majority governments. The effect would therefore be biased toward minority governments, they argued. However, de Haan and Sturm9 did not find evidence in support of even this view and demonstrated that it does not withstand a robustness and sensitivity analysis. Accordingly, the weak government hypothesis by Roubini and Sachs has been rejected by these and most other scientists in subsequent literature, including Kontopoulos and Perotti10. With regard to the entire preceding literature which only focused on government strength and/or coalition size, they criticize that fragmentation would have been perceived solely as what they term legislative fragmentation. Enhancing research on fractionalization, they investigate executive fragmentation as well by introducing cabinet size as a measure. Kontopoulos and Perotti argue that the effects of legislative fragmentation are less significant and robust than those of executive fragmentation. Since they do not require such fundamental changes as, for instance, reelections, adjustments to the executive side of the fiscal decision-making process (changes in the cabinet setup) can obviously be implemented more easily than changes to the legislative side. Therefore the political relevance of executive fragmentation would be enough of a reason to investigate its effects11.
Kontopoulos and Perotti (1999) conduct a panel analysis for 20 OECD countries over the time period of 1960 to 1995 with regard to their respective levels of size fragmentation. In addition, they focus on the varying effects of fragmentation depending on the particular economic scenario by differentiating between “good” times (positive economic growth) and “bad” times (periods of economic contraction).
1 see e.g. game theory, where the level of cooperation is inversely related to the number of players
2 Kontopoulos, Y.; Perotti, R. (1999), p.81
3 Kontopoulos, Y.; Perotti, R. (1998), p.1-2
4 Kontopoulos, Y.; Perotti, R. (1999), p.82
5 see e.g. Velasco (1995)
6 Kontopoulos, Y.; Perotti, R. (1999), pp.81/82
7 Roubini, N.; Sachs, J. (1989a,b)
8 Edin, P.; Ohlsson, H. (1991)
9 De Haan, J.; Sturm, J.-E. (1994, 1997)
10 Kontopoulos, Y.; Perotti, R. (1999), p.83