Governance spending fixation as a tool for reform to improve institutions

Abstract

There is a consensus in the scientific community that most often the main obstacle to development is the absence of good institutions, however, it is also widely recognised, that this idea is of little use without explanations on how they can be improved and made to self-sustain ‘good equilibrium’. Purpose of this article is to propose such an explanation together with a tool for reform that could potentially address this issue. For this purpose, literature review is conducted to explore links of institutions to economic growth in search for opportunity to have a positive effect to the system through means of reform. Such an opportunity is found exploring governance-corruption-economic growth connection, to exploit which, a simple method is proposed that involves only basic math. A matching exercise between governance spending variables and governance evaluation variables reveals both a best way to judge governance spending and a potential value for orientation that could be used for limiting misallocation of resources in countries with governance overspending. At first glance a modest find may have significant practical implications, as this value could be used as a tool for reform, application of which could be expected to have an effect in two complementary ways: pragmatic – inefficient use of resources could be limited by diverting them to other government sectors; and incentives based, which when looked at through a lens of literature on ‘New institutional Economics’ looks especially promising, as in essence it could potentially enable application of principle of free market in state governance for most of poorly governed post-socialist European countries.

Keywords:

• Governance
• institutions
• corruption
• incentives
• reform

JEL CLASSIFICATION:

• B520
• D720
• D730
• H110

1. Introduction

‘Demonstrating that growth is a function of good institutions is of little help if this is not followed by an analysis of how such institutions emerge and become self-enforcing’ (Avellaneda, 2010, p. 30). Though up until now such analysis ending with reform concept propositions is absent, scientific literature is rich in clues to what needs to be done and how. It is known from leading scientists in the field, that governance reform concept propositions must be based on proper understanding of institutional incentives for progress and prevention (see Kaufmann & World Bank, 2003). ‘…to be more effective, efforts to reduce corruption should be intrinsically woven into policies to develop institutional capacity and optimise the size of government spending’ (Dzhumashev, 2014, p. 12). Or to put it simply: incentives must be built into pay schemes. This article suggests a governance reform concept and a tool for it, that is in accordance with everything as above.

The end result of obtaining usable recommendations is made possible by taking an alternative approach to mainstream economics – that is by avoiding all complex methodologies. Instead, what works is found out in simplest possible ways, so only basic math is used. After theoretically mapping out corruption-growth links, a channel that is easiest to affect is chosen for governance reform. The channel is ‘rent-seeking costs’ that transmit mitigating effect of development to corruption. A way to evaluate state governance¹ spending that is suited for reform is found by a manual matching exercise of extreme simplicity. One concept of evaluating country’s governance spending is found to be superior to others by having values that are predominantly different among groups of good governance and poor governance countries in the country sample, while being predominantly uniform within the groups as mentioned. This discovery allows calculation of a benchmark value of state governance spending of good governance country, that could potentially be used to identify misallocation of resources in the form of governance overspending in severe cases. This leads to a recommendation to lower and fix governance spending at the recommended value in poor governance countries. This act is in principle the proposed reform concept. It necessitates an overall optimisation of governance system by reducing inefficient use of resources in the form of excess financing to governance sector, and the very way that the governance spending is measured creates an incentives-based mechanism for self-sustaining ‘good equilibrium’ or ‘optimal inclusiveness’ when speaking in terms of institutional economics.

This article is organised as follows. A review of relevant literature is in Section 2. Theory is developed in Section 3. Section 4 contains description of methodology and data. Results are presented in Section 5. Implications are explained in Section 6. Conclusion is in Section 7.

2. Literature review

Idea that institutions are a primary determinant of economic performance is a consensus among mainstream economic scientists and, according to some latest works, they may be the most important factor when it comes to welfare, development and economic growth (Acemoglu & Robinson, 2012; Madland, 2015). Incentives created by different institutions channel profit motive differently and this is the main reason for differences in economic prosperity of nations (Acemoglu & Robinson, 2012; Avellaneda, 2010). Strand of analysis established for investigating the relationship existing between institutional framework and economic performance is generally called ‘New Institutional Economics.’² As summarised by Gagliardi (2008), there are three main definitions of institutions: rules of the game (game meaning the economic and political processes); players of the game including organisations along with the rules in force; and self-enforcing equilibrium outcome of the game. In this article, the terms meaning is dependent on context and is used as defined by anyone of the three definitions, whichever best fits circumstances. As each society functions with a set of economic and political rules created and enforced by state and citizens collectively, conflict over scarce resources, income and power, translates into conflict over rules of the game, the economic institutions, which determine economic activities and who benefit from them, making political institutions a key determinant of the outcome of this game (Acemoglu & Robinson, 2012; similar conclusion reached by Basu, 2004). Scientists researching the issue come to conclusions like: countries should promote good governance to accelerate their growth rates (e.g., Aisen & Veiga, 2013; Cooray, 2009; Dar & AmirKhalkhali, 2002), development (e.g., Basu, 2004; Kaufmann & Kraay, 2003; Kaufmann & World Bank, 2003; Shah, 2005), make welfare gains (e.g., Angelopoulos, Philippopoulos, & Vassilatos, 2009; Kaufmann & World Bank, 2005), attract foreign direct investment (e.g., Busse & Hefeker, 2007; Globerman & Shapiro, 2002; Peres, Ameer, & Xu, 2018; Wisniewski & Pathan, 2014), increase happiness (though be it after dispensing with poverty first) (Changbin, 2018) and so on, or simply that governance and institutions matter in one way or another.³ Some conclusions were formulated into useful insights like: ‘…ending endemic corruption requires ensuring that people with a high personal aversion to corruption choose to become government bureaucrats’ (Kahana & Qijun, 2010, p. 7) or: governance improving requires institutional incentives for progress and prevention mechanisms (Kaufmann & World Bank, 2003). Some works provide general principles that can be used for governance innovation to encourage investment and growth (e.g., de Ferranti, Jacinto, Ody, & Ramshaw, 2009; Fisman & Werker, 2011). These and many other authors did a wonderful job at expanding our understanding in the field of governance and institutions with their works that explored the issue and formulated tasks for the next step. An extension of their work would logically be suggesting reform concepts for helping good institutions to emerge and become self-enforcing, which is where main contribution of this article lies, thus putting it on the forefront of current academic effort to give usable suggestions for governance reform and doing it in highly harmonious fashion in regard to other literature that is already present (Bason, 2017; Sørensen, 2017, etc.).

The main difficulty in this topic is that consensus among scientists goes only as far as recognising that institutions and politics that keep them in place are important. There are no accepted single definitions, let alone single accepted measures on these issues. Politics is first defined as activities associated with governance.⁴ While there is a consensus among scientists that governance is important, definitions of it are plentiful⁵ and proposed ways to measure it are in the range of hundreds.⁶ This is so, because international organisations use the term ‘good governance’ and discuss its promotion in the context of their main objectives, just as scientists do the same by adopting definitions and measures according to specific contexts, challenges or end goals desired. Furthermore, as pointed out by Bersch and Botero (2014): ‘Governance, like many other concepts in the social sciences, is a latent variable. It is not directly observed but inferred based on other observable manifestations (indicators) of the latent quantity (governance, or a dimension of governance such as rule of law.)’ (p. 11). The same line of thought is found in work of Malik (2002) who notes that ‘… governance is essentially a qualitative phenomenon, quantification of which would always be subject to considerable empirical limitations’ (p. 3). ‘Governance was never allowed to become a conceptual straight-jacket but was expected to function as a rather loose framework within which each researcher could creatively explore political issues of significance. The problem that we encounter, therefore, is not the limitations stemming from the imposition of a confining concept, but rather the opposite: the challenge of making sense of the wide range of interpretations of governance that the authors bring to the agenda’ (Hyden, Olowu, Hastings, & Ogendo, 2000, p. 6). Consequently, validity of every governance evaluation concept is debatable including the one used in this article and this must be accepted as an inherent property of this topic.

The most widely accepted and used governance measure today is World Governance Indicators (WGI) developed by Kaufmann, Kraay, and Mastruzzi (1999a, 1999b) in World Bank, which comes in form of six separate indexes representing six key aspects of disaggregated concept of governance complete with its definition.⁷ Authors acknowledge that ‘every one of [their] underlying data sources—or any other potential measure for that matter—is at best an imperfect proxy for governance’ (Kaufmann et al., 2007, p. 3). Not surprisingly, WGI has plenty of critics pointing out various possible deficiencies and giving arguments that put its practical use in question. The ones that make author of this article to take a different approach in measuring governance are following: it has been shown that converging and diverging developing countries do not differ in terms of these indicators as well as some others (Khan, 2007); secondly: an opinion exists that ‘the efficacy of [any governance aspect improving] institutional reforms may be conditional on transformations of underlying economic and social structures that themselves determine the degree to which governments can be held to the goals embodied in such reforms or whether they are yet another in a long series of dead letters’ (Kurtz & Schrank, 2007, p. 3). It is worth adding, that creators of WGI as well as most other authors are notorious for avoiding to directly link governance and development indicators. Though understandable from professional standpoint and useful for some scientific purposes, such position is not helpful when it comes to governance reform concept development, as it possibly ignores complex reality of interlinks⁸, as governance and development are sometimes linked by definition.⁹ These interlinks likely make up a complex and dynamic system that is referred to in literature as the underlying structure. Regardless of whether it is called socioeconomic, sociopolitical, institutional or economical – it is one and the same all-encompassing structure, that is a background of every economic action. Understanding interlinks is crucial for reform concept development, therefore established governance indicators such as WGI are disregarded in favor of a different evaluation concept proposed by World Bank.

Obvious simplification of finding links between desired factor and governance is to choose a convenient definition as a starting point. For this reason, definitions given in World Banks ‘Public Sector Governance and Accountability Series. Public Services Delivery’ (2005) are used together with appropriate elements of the framework for governance quality evaluation as offered by the series. In the work mentioned is stated that: ‘Chapter [2] introduces a measurement that allows both theoretical work and policy issues to be discussed in the framework of a concrete definition of governance quality. The definition used could easily be modified to reflect different beliefs about the relevance of the components used in this index. The index could also be narrowed or broadened to reflect differences in beliefs about the role and scope of government’ (Shah, 2005, p. 54). In other words, this framework was designed to be adaptable to specific circumstances, which is the first thing to pay attention to, when it comes to designing institutional reforms, and therefore it will be used in this article. Definition used for governance evaluation criteria selection is as follows: ‘governance is a multifaceted concept encompassing all aspects of the exercise of authority through formal and informal institutions in the management of the resource endowment of a state. The quality of governance is thus determined by the impact of this exercise of power on the quality of life enjoyed by its citizens.’ (Shah, 2005, p. 40). As noted by the series: ‘Although no single index can conceptually capture all aspects of governance, a focus on key observable aspects of the governance dimensions can be helpful…’ (Shah, 2005, p. 40). To this end, present article uses concept of governance evaluation using corruption and social development¹⁰ as offered by the series.

In order to be useful, idea of underlying economic structures has to take shape of a blueprint to work with. For this to happen, identification of relevant factors and mapping of links among them has to be made. It may be proper to begin with realisation made by Yahyaoui, Chatti, and Chtourou (2008) that no clear distinction can be made between governance and institutional factor, however, institutions generally correspond to a broader concept.¹¹ The same line of thought is found in work of Rajkumar and Swaroop (2008) who state: ‘… reality is that public spending, governance and development outcomes are interlinked …’ (p. 3) and there is no shortage of studies to back this up nor is this realisation unique to authors mentioned alone.¹² Greatest threat to good governance comes from corruption, violence and poverty.¹³ When working with countries in which latter two factors are at levels that do not threaten governance, they can be excluded, making corruption the main threat by this UN definition. From the structure of WGI and work of Kaufmann and World Bank (2003) is known that corruption is one among other closely intertwined governance components (p. 5). Another insight by Kaufmann and World Bank (2003) is that, in some settings, a form of corruption can shape political forces and outcomes and therefore can become a force of its own instead of being a symptom of more fundamental political forces. This ‘grander’ form of corruption is referred to as state capture and is described as tendency by some elite firms and conglomerates to illicitly shape the formation of state laws, policies, and regulations. Summing up information above, it can be concluded that corruption is the single biggest threat to good governance while being its component and is capable to become a force of its own at shaping rules of the game for business environment, which all implies that it is a key determinant of nature of institutions in countries where violence and poverty do not present a threat to governance. As noted by Khan (2004) the same ends as in the case of state capture can be achieved by legal processes of lobbying, political contributions and so on, which converts illegal influence-buying in the form of corruption into legal influence-buying of different sorts. Realisation that powerful elites and conglomerates may play an important role in shaping rules of the game for business environment by legal means, as well as illegal ones, implies that what is commonly perceived as corruption, is much more than a problem of bureaucracy. As political institutions are rules that govern incentives in politics and determine the nature of economic institutions, ‘state capture’, together with its legal forms, by their ability to directly affect these rules, constitute both a key channel and a fundamental force of entire underlying system with sole purpose of exerting a direct effect on the equilibrium making institutions more extractive.¹⁴ Though having much in common with corruption as commonly perceived, this force and its effect is not necessarily negative from social standpoint. On the contrary – it is an essential ingredient of dynamic and vibrant capitalist system and is inseparable from process of creative destruction and progress. By making institutions more extractive and thus pushing equilibrium towards more inequality, it increases incentives to put more effort to economic activity as returns from it are also increased. Without it, economy is bound to degrade to stagnant socialist system the likes of which are synonymous with economic failure due to absence of right incentives. Too much of it results in extractive institutions taking over and causing welfare loss by increasing inequality, poverty and with very much the same end result of economic stagnation and social decline. Grander form of corruption referred to as state capture is one of forms of expression of this fundamental force that is balancing out with another fundamental force that pushes equilibrium to the opposite direction – demands of society for general welfare and equality. In such setting, systemic means for keeping ‘good equilibrium’ are needed. In order to design those means, understanding links between factors is key. For theoretical mapping of these links, corruption is a logical starting point due to its importance in the growth–governance relationship.

Analysis on corruption has to begin with two different existing views on the issue. Corruption as a major obstacle to development is the dominating view. The other one is known as ‘greasing of wheels’ hypothesis general idea of which is that corruption can, in some cases, foster development by promoting efficiency, which is done by allowing private sector to correct or circumvent government failures of various sorts, and so, to facilitate economic activities, that otherwise would not be. Real-world instances that support this view likely are happening as judged by positive relationship between FDI and corruption (see Egger & Winner, 2005), yet any positive outcome is always associated with low levels of corruption and/or specific circumstances usually involving governance failure of some sort, that itself can be removed as part of an anti-corruption policy (Khan, 2004). When it comes to governance, corruption could help to overcome government failures in the short term, however, it creates incentives to keep these failures and create more of them in the long term (Aidt, 2009). It is thought that different incentives created by different institutions are the main reason for differences in economic prosperity of nations (Acemoglu & Robinson, 2012). Due to inherent property of corruption to create ‘wrong’ incentives in governance, ‘greasing of wheels’ hypothesis is incompatible with prosperity improving governance reforms and therefore is disregarded. This can be backed up further by research suggesting that corruption reduces substantially the growth rate of genuine wealth per capita—a direct measure of sustainable development, therefore any positive effect on growth is likely to be unsustainable development (Aidt, 2009). For these reasons, corruption is synonymous with governance failure, therefore, it can be possible to improve governance and institutions by proposing reforms, that put in place incentives to resist it or block its effect transmission channels.

Adequate understanding of corruption is necessary for institution improving reform development, therefore analysis of its types is needed. Kaufmann and World Bank (2003) begins this analysis with separating ‘pettier forms of corruption such as administrative bribery and ‘grander’ form – illicit formation of state laws, policies and regulations referred to as ‘state capture’. According to Kaufmann and World Bank (2003), the principal difference between these two forms is that in case of first one, the problem is within bureaucracy, which is generally and wrongly attributable to the whole concept of corruption.¹⁵ In case of ‘state capture’, powerful elites and conglomerates illicitly shape rules of the game, making local bureaucracy more of a means rather than the source. Surveys show that when corruption is unbundled into its components, though types of it are correlated with each other within a country or a region, across some dominating type differs (see Kaufmann & World Bank, 2003), which implies that needed reforms may also differ (see Khan, 2004). Concept of corruption encompasses a wide range of behavior. As summarised by Morris (2011): ‘among the criteria most commonly used to draw distinctions are: the institutional location and function of the public official involved (“political corruption” versus “bureaucratic corruption”), the direction of influence (“bribery” versus “extortion”), and the size and frequency of the transaction (“grand corruption” versus “petty corruption”)’ (p. 13). This list of criteria is in no way complete or the only one available (e.g., Khan, 2004), however, others will be disregarded because in this article corruption is grouped according to its effect transmission channels to economic growth with the aim of mapping their interrelation and identifying any possible opportunity for potential institutional reform.

Main classification of corruption of this article:

Bribery is an offer of money or favors to influence a public official. ‘Kickbacks operate much like a bribe, but illegal payment is made after the service is rendered, usually from a portion of the governmental award itself’ (Morris, 2011, p. 1); speed-money, where money is given to official to shorten the time needed to pass procedures by legal means or not, and extortion, where the public official threatens to use (or abuse) state power to induce payment of a bribe, are all added to the category of ‘bribery’.

Cronyism with its variations and other names of nepotism, clientelism, elitism by their essence all are favoritism shown by public officials to relatives or close friends. As described by Morris (2011): cronyism is the practice of partiality in awarding jobs and other advantages to friends or trusted colleagues, especially in politics and between politicians and supportive organisations, for instance, this includes appointing ‘cronies’ to positions of authority, regardless of their qualifications. Purchase of appointed position can be viewed as the most extreme case of bribery, however, in present work it is grouped as cronyism, because it is hard to imagine it happening without the necessary connections in power. Perhaps most importantly ‘state capture’ in this article is considered an advanced stage of cronyism as in essence it is the same privileged regulation and/or legislation and the scale of it is the only meaningful difference. Privileged regulation and legislation usually reduces competition by trade restrictions or government-created barriers to entry, however, avoidance of regulation is just another form of the same.

Theft sometimes called embezzlement is stealing money or other government property and in its end effect is similar to fraud which by definition is cheating the government through deceit, therefore in this article is also considered as a kind of theft.

Political graft occurs when funds intended for public projects are intentionally misdirected in order to maximise the benefits to private interests, and in present classification is somewhere between cronyism and theft, and perhaps both at the same time. This includes disguised transfers (where public road may be built to increase value of real estate of cronies etc. as described by Angelopoulos, Philippopoulos, and Vassilatos (2009) and creation of unnecessary government institutions and/or positions to employ relatives, etc. This category of corruption mainly represents altering of public spending composition for rent-seeking purposes, that has been empirically proven to exist (see Hessami, 2014; Mauro, 1998).

Though list of examples of any one type of corruption mentioned above is far from complete, they are enough to see that in real world, boundaries between different types of corruption are blur, yet still make it possible to link them to ‘effect to growth’ transmission channels described in literature.

The first known work on empirically testing corruption transmission channels to economic growth names three such channels: investment reduction, human capital accumulation and political instability with the latter found to be most significant (Mo, 2001). It must be stressed, that proportions of total strength possessed by separate elements of corruption are of little matter to present article. Dynamic nature of corruption in real world means not only that degree of importance varies in time of any one type, but also types themselves appear and disappear, and this makes quantification of little importance. Corruption also affects economic performance through efficiency of public spending (Dzhumashev, 2014; Spinesi, 2009). Adding to this list is a channel of trade openness (Hessami, 2014), misallocation of talent, distorted investment priorities (Malik, 2002). This, in turn, is achieved through distortionary tax, red tape and increased risk. Corruption has also been shown to increase income inequality and poverty through channels of regressive taxing, poor targeting of social programs, unequal access to education, reduced social spending, and higher investment risks for the poor (Malik, 2002). Inequality can affect growth in either positive or negative direction. As summarised by OECD (2015), greater inequality might reduce growth through political instability, reduced human capital accumulation and lack of domestic demand; or increase growth through providing better incentives to work harder, invest and undertake risks, and increased aggregate savings due to rich having lower propensity to consume. Econometric analysis suggests, that income inequality has a sizable and statistically significant negative impact on growth, and further analysis of data reveals that one key channel through which inequality negatively affects economic performance is through lowering investment opportunities (particularly in education) of the poorer segments of the population (OECD, 2015). Level of development may, in turn, affect corruption through rent-seeking costs which rise with wage rate (Dzhumashev, 2014) or through cost-benefit balance if rent-seeking activities get ever more expensive due to sophistication required or harsher punishments with higher chances to get caught, etc. Corruption-growth transmission channels as summarised above are illustrated in Figure 1, which serves as an initial blueprint to work with, so justifying its over simplistic depiction of corruption – growth relationship that ignores much of complex reality with ambiguous interrelations like for example a link through shadow economy (Bayar, Odabas, Sasmaz, & Ozturk, 2018, etc).

Figure 1. Corruption-growth interrelation. Source: Authors own illustration.

As can be seen from Figure 1, corruption has both direct effect and indirect effect on growth through political stability and human capital accumulation channels, as negative effect of inequality is transmitted through them as well. This may explain their comparatively strong overall effect on growth found by OECD, 2015. The most revealing portion of Figure 1 could be the links coming to and from type of corruption called cronyism. Not only that it can be linked to every known negative effect to growth transmission channel, but essential constituent of it ‘privileged regulation and legislation’ can be accessed by legal means of lobbying and political contributions, etc. Such efforts by legal means or not can contribute to or result in ‘state capture’, which in turn can block both known ‘mitigating’ effect channels of growth to corruption as shown in Figure 1. Blocking of mitigating effect through ‘rent-seeking costs’ channel is best revealed in summaries by Berrios (2010) where statement can be found that commonly proposed remedy to strengthen law enforcement will not work if law enforcers are captured by the interests they are supposed to regulate (p. 247). Capability of corruption to block attempts of better redistribution is straight forward and self-evident due to its definition: illicit shaping of rules and regulations. Explanation on how this works is given by Kaufmann and Kraay (2002). After empirically finding that weak and even negative causal effect runs in the direction from per capita incomes to governance, authors name the phenomenon of state capture as the most likely reason. ‘If the fruits of income growth largely accrue to an elite that benefits from misgovernance, then any possible positive impact of income growth on governance could be offset by the effect of the elite’s negative influence’ (Kaufmann & Kraay, 2002, p. 27). ‘To the extent that state capture is important, higher incomes may be appropriated by the monopolistic captors or elites. This, in turn, can lead to additional demand for private purchase of laws and regulations ensuring the continued dominance of the elite. The net effect is to erode overall governance, particularly the public protection of property rights, the incidence of corruption, and more broadly, rule of law.’ (Kaufmann & Kraay, 2002, p. 29). So authors describe the workings of a vicious circle that can be seen in Figure 1 and that cannot be broken by paying more to bureaucrats or hiring more of them. It is a description of an extractive system getting ever more extractive as is described in the work of Acemoglu and Robinson (2012). In this context, it must be stressed that the entire issue of corruption is not straight forward. As warned by Kaufmann and World Bank (2003): ‘a fallacy promoted by some in the field of anti-corruption, and at times also by the international community, is that the best way to fight corruption is by fighting corruption— that is, by means of yet another anti-corruption campaign, the creation of more anti-corruption commissions and ethics agencies, and the incessant drafting of new laws, decrees, and codes of conduct’ (p. 88). The author warns that all this together with instinctive tendency to over-regulate comes at the expense of a focus on prevention and incentives for integrity, and has reduced the effectiveness of anticorruption efforts. Clearly what is needed is incentives based mechanism that would balance out public and private interests and help prevent the system from becoming excessively extractive regardless of whether caused by corruption or by legal means.

3. Theory

The most obvious way to have an effect on the entire system is through rent-seeking costs (ref. Figure 1). Presumably, the larger these are – the less likely should be the incidence of corruption. There are two different approaches to increase rent-seeking costs: through higher salaries to bureaucrats and through cost–benefit balance including higher chances to get caught, severity of expected punishment, level of sophistication required, etc. Both of these approaches are complementary, however, the second one is more heavily reliant on existing institutional background to work properly, therefore is more vulnerable to corruption. When levels of corruption are high enough ‘de juro’ and ‘de facto’ is separated and rules set by law, in essence, lose their purpose. The first approach of salary regulation seems to be more reliable lever to influence the issue and better suited for technocratic approach that is needed to design a governance system reform.

From literature analysis in section 2, it is clear that more governance spending can have no positive effect on its quality and harm economy by not putting it to better use, such as investing into public capital (misallocation of resources). Underfinancing governance can also be harmful due to unnecessary limitations of the state to play its role in social transformations needed for development (Khan, 2004) and incentives to engage in rent-seeking that insufficient salary provides to otherwise fair bureaucrats. Theoretically this means that there could be an optimal level of governance spending, maintaining which could strengthen governments capabilities in governance, as insufficient financing could be detected and remedied, and in case of excess financing – loss of funds should incentivise finding optimal structure of government institutions by ditching unnecessary establishments and unnecessary regulations, etc. making overall structure more efficient. Even a governance spending value suited only for approximate orientation could be used for keeping in check needless growth of bureaucracy and misallocation of resources caused by it in countries with severe overspending.

4. Methodology and data

To find a recommendable level of governance spending, it has to be put against evaluation criteria, and nature of relation must be determined.

For the purpose of governance quality evaluation, appropriate elements of framework given in World Banks ‘Public Sector Governance and Accountability Series. Public Services Delivery’ (2005) are used as mentioned in section 2. From work as described above, among other things, present article ‘borrows’ concept of governance quality evaluation using corruption and social development. The main difference of present work and the one described above is that instead of aggregating chosen indexes into one as is in the case of proposed by the series ‘governance quality index’ (GQI), no aggregation is done, and chosen indexes are used as they come from providing sources. This is so, because aggregation as a method is known for its inherent loss of information. As cited in section 2, index given in the series is intended to be modified in any way found necessary together with definitions, and the whole purpose of the series is to provide a framework in which governance could be discussed, therefore it can be argued, that dispensing with aggregation while leaving evaluation components as suggested does not constitute a departure from the framework.

There are three broad categories of social progress indexes that measure welfare: ‘corrected GDP’; indexes that do not use GDP; and composite indexes that include GDP. ‘Corrected GDP’ use national accounts and GDP as the foundation and then add or subtract quantities to mitigate unsuitability of GDP itself to be used as a welfare measure. Indexes that do not use GDP do not measure economic activity; rather, they measure environmental or social activities, well-being, or changes in environmental, social, or human capital (Costanza, Hart, Posner, & Talberth, 2009). This group includes subjective welfare measures that attempt to measure satisfaction with quality of life. This group will be disregarded in this research. Indexes that include GDP are composite indexes that combine several different measures into a single number. Human Development Index (HDI) made by United Nations Development Program is perhaps the best-known of the group and will be used in this research together with best-known corruption index – Corruption Perception Index (CPI) made by Transparency International. To check robustness, research will be done twice – using another welfare measure and another corruption index. To approximately represent category of ‘corrected GDP’ group – productivity will be used for a second social welfare measure together with possibly second best-known corruption index – Control of Corruption index calculated by International Country Risk Guide.

Data of governance spending is taken from Classification of Functions of Government (COFOG) in Eurostat database (see Appendix in Data file for details). Country sample is limited by the database: 31 countries – 28 EU members + Iceland, Norway and Switzerland; however, sample had to be reduced due to insufficient data for Croatia, Cyprus, Malta and Iceland. Period of 2001 to 2016 selected also due to data availability issues.

In order to evaluate governance spending – a comparative statistic must be derived for cross country comparison. It is not known which way of assessing governance spending is best suited for this purpose, therefore several of them are tested. First option is governance spending – gross domestic product (GDP) ratio, that will be referred to as variable A. In order to have governance spending in perspective of living standards of local people in particular country, a second comparative statistic is made by calculating governance spending per worker per year ratio with average brutto monthly salary in that country. Put it bluntly: it shows what portion of monthly income does an average worker in a particular country has to pay for the country’s governance as if it’s one of public services (perhaps the ultimate one). This variable will be referred to as B. It may make sense that governance spending could rise with an increase of overall population of a country, therefore a third variable C is introduced by calculating governance spending per person per year ratio with average brutto monthly salary in that country.

In order to find optimal or at least recommendable level of governance spending to be used as a benchmark, it has to be determined which of the three comparative statistics (variables A, B, or C) is best fit for the purpose. Best-governed countries must have similar governance spending among themselves and worst-governed countries also must have similar governance spending among themselves, and it must be different from the one of best-governed countries. In order to find such pattern, an extremely simple manual matching test is used: in a sample of countries of any given year, top and bottom seven values (a quarter of sample) of evaluation criteria (welfare and corruption measure) are put against each of the comparative governance spending statistics (variables A, B, and C). This is done in order to see if governance spending measured in any comparable way can be matched to better development and/or corruption indexes, which both serve as governance evaluation criteria in this research. Fitness of different governance spending evaluation statistics (variables A, B, and C) for the purpose of determining benchmark governance spending is judged by distribution of their top and bottom seven values in relation to top and bottom seven values of evaluation criteria of governance. To put it simply: if top (or bottom) seven values of governance spending are found to be among seven least developed or seven most corrupt countries in the sample and/or bottom (or top) seven values of governance spending are found to be among seven most developed or seven least corrupt countries in the sample – it would be evidence of existence of optimal governance spending value as measured by that particular governance spending evaluation variable. Since governance is evaluated using two separate statistics – welfare and corruption measure, and a country may or may not have them both belonging to top or bottom seven values of the sample at a given time period, it is considered a match if any of seven top or bottom values of governance spending variable being tested are among any of top and/or bottom seven values of any of the two evaluation criteria. To put it simply: if highest seven values of governance spending are found to be among countries with lowest seven values of welfare measure or highest seven values of corruption or both – it may be interpreted as evidence of highly inefficient governance system due to misallocation of resources. The same goes vice-versa: if lowest seven values of governance spending are found to be among countries with highest seven values of welfare measure or lowest seven values of corruption or both – it may be interpreted as evidence of superior efficiency of the system of those countries and an indication of what governance spending must be.

In cases where multiple values of any variable of any given year are the same and it is impossible to separate top or bottom values so that there would be seven of them in total – all recurrent values are treated as top or bottom seven in the matching exercise, however, if there is more than one recurrent value, total number of them used for matching exercise is kept as close to seven as possible, but not less than seven. To put it simply: if there are nine top or bottom values that are the same of any variable of the given year, then all nine of them are used for matching, however, all other values are ignored. In cases where there are two top (or bottom) values of any variable with four recurrences each, making eight top (or bottom) values in total – all eight are used for matching exercise, etc. This is so in order to see if top or bottom values of any governance spending evaluation (variables A, B, or C) point out best-evaluated governance.

In order to make matching exercise easier and more graphic, top seven governance evaluation criteria variables (welfare and corruption measure) are colored in green, while bottom seven – in red. Similarly, seven highest values of governance spending evaluation are colored in red and seven with lowest values – in green. To both match distribution of values and compare how each of the three governance spending evaluation variables is matching with selected governance evaluation criteria, exercise is carried out manually, by making a separate table for each year with all seven variables of all the countries in the sample. As mentioned above, this is done twice: research 1 uses Human Development Index (HDI) and Corruption Perception Index (CPI) as governance evaluation criteria, values of which are checked for matching with values of governance spending evaluation (variables A, B and C) in Tables 24a–24p (contained in data file); research 2 is done in exactly the same fashion as research 1 only governance evaluation criteria used are different: productivity is used instead of HDI and Control of Corruption index is used instead of CPI (Tables 25a–25p are contained in data appendix file). Results are shown in Table 1.

Table 1. Results of governance spending evaluation variables matching with governance quality evaluation criteria exercise.

	Research 1^a			Research 2^b
Governance spending evaluation variable:	A	B	C	A	B	C
Percentage of matches green to green^c	75.0%	83.2%	81.3%	69.2%	79.6%	81.3%
Percentage of matches red to red^c	61.3%	82.1%	73.0%	53.8%	89.3%	74.8%
Percentage of matches (total)	68.2%	82.7%	77.1%	61.5%	84.4%	78.0%
Percentage of mismatches (total)	17.2%	2.2%	9.3%	47.8%	1.8%	9.7%

Note: ^aResearch 1 uses HDI and CPI as governance evaluation criteria for matching exercise with each of governance spending evaluation variables (A, B, and C).

^b Research 2 uses productivity and corruption control indexes as governance evaluation criteria for matching exercise with each of governance spending evaluation variables (A, B, and C) and is carried out in exactly the same fashion as research 1.

^c Green and red refers to top five governance evaluation criteria values (of both welfare and corruption measure) being colored in green, while bottom five – in red. Similarly, highest five values of governance spending evaluation are colored in red and five with lowest values – in green. This is done to all three governance spending evaluation variables (A, B, and C) used for the matching exercise as described in the methodology section.

Source: Authors’ calculation. See Results section in Data file.

5. Results

Manual matching exercise reveals the best way and the worst way for measuring governance spending in relation to governance quality. Robustness of the find is proven by the fact that both researches give similar conclusions in the sense that best way for measuring governance spending is found best at both researches; second best is second best at both researches and one is found to be useless also at both researches.

From result table, it can be seen that best conception for judging governance spending in relation to governance quality is variable B, which is governance spending per worker per year ratio with average brutto monthly salary of the country. This way of measuring governance spending surpasses others on nearly every statistic used to describe accuracy of it’s matching with selected governance quality evaluation indicators, and shows satisfactory overall results. In this article from this point onward ‘governance spending’ refers exclusively to measurement by variable B unless otherwise stated. Its seven lowest governance spending values have 83.2% matching in research 1 and 79.6% matching in research 2 with one or both top seven values of governance quality evaluation. Seven highest values of governance spending (red) have 82.1% matching in research 1 and 89.3% matching in research 2 with one or both bottom seven values of governance quality evaluation.

Percentage of mismatches 2.2% and 1.8%, respectively, could have at least two possible explanations. First, one must allow for a possibility, that these figures may show some kind of developments in governance institutions. Increased governance spending in high development and low corruption country may signal corruption of political institutions, however, temporary misallocation of resources and possible acts of corruption are not capable of instantaneously undermining development and, likely, do not affect corruption perception, especially if these acts stay undiscovered, therefore method used in this research is unfit to prove or disprove this. Such suspicion is suggested by the fact that most mismatches that were found are rare occurrences of highest values of governance spending among low corruption and high development countries, that are exceptions to otherwise a rule of low governance spending relating to good governance performance. An opposite occurrence – if high corruption and low development country would decrease its governance spending to the level of at least that of an average of good governance country sample, could suggest that an effort is taking place to combat excess bureaucracy and misallocation of resources, and this might be the case for Bulgaria. Unfortunately, there is another way to explain mismatches: there might be differences in how statistics used in research were calculated within different providing countries and at different time periods. Possibility of biased data cannot be excluded.

A surprising find is that governance spending ratio with country’s GDP (variable A) is the worst way to judge governance spending. It has every worst statistic from ones used to describe accuracy of it’s matching with selected governance quality evaluation indicators, with number of mismatches especially impressive when compared to other two tested variables: 17.2% and 20.5% against 2.2% and 1.8% of variable B and 9.3% and 9.7% of variable C. For this reason, judging governance spending by its ratio with GDP must be pronounced useless and potentially misleading.

The objective of finding optimal governance spending value is not achieved due to limitations of the method used. Since match between lowest values of variable B and top values of governance quality evaluation is not perfect, it cannot be concluded that lowest values are optimal. This is all that this research is capable of uncovering in regard to optimal governance spending value, not least because of probability that none of the countries in the sample may have it.

Benchmark value of governance spending, that can be used for determining inefficiencies in severe cases, could potentially be suggested using obtained data. It has been uncovered that while lowest governance spending values as measured by variable B are commonly found among best-governed countries, highest values are mostly found among countries with worst governance. This relationship gives reason to believe that optimal value is somewhere at the lower end of the scale of governance spending as measured by variable B and allows to make a claim, that an average value of good governance countries can be used for orientation by countries with poor governance performance. To obtain such value, several averages are calculated. First, an average value of variable B is calculated for countries with highest development values for every year (Average 1), then an average value of the same variable is calculated for least corrupt countries in the sample for every given year (Average 2). Last, average value is calculated for countries that have at least one of top governance quality evaluation figures used in research (Average 3) for every year. Table of average values of governance spending of best governance countries in the country sample is given in Table 27 in Data file. Statistics describing resulting collection of average values as described above are given in Table 2.

Table 2. Statistics describing collection of averages of variable B for every year of both researches of best governance countries.

Min value	Max value	Median	Average	Coef. of variation	SD
0.86	1.09	0.99	0.99	5.26%	0.052

Source: Authors’ calculation. See Results section in Data file.

Data in Table 2 shows that an average governance spending value of best-governed countries in the sample is remarkably consistent regardless of which way the average is calculated with only 5.3% coefficient of variation in the time period even if considering values of all three options (average 1,2 and 3) simultaneously. This leads to a conclusion that average value of governance spending per worker per year ratio with average brutto monthly salary of best governed countries could potentially be used as a benchmark to determine misallocation of resources in comparatively poor governance countries that have larger governance spending than suggested value. Namely those countries are: Romania, Hungary, Latvia, Portugal, Slovakia, Czechia, Lithuania, Greece and Italy. In the well-governed countries in the sample that includes Norway, Switzerland, Germany, Ireland, Denmark, Netherlands, United Kingdom, Luxembourg, Finland, Belgium, and Sweden the latter four could potentially also find the value useful.

6. Implications. Changing the rules of the game

Up until this point in history, governments themselves determined governance spending with no guidance of any kind. To some countries it has worked well. Government members and people working as bureaucrats and possessing fundamental willingness to do the right thing for their nations have played their crucial part in building and maintaining countries of yet unseen prosperity in human history. Problem is that not all countries have enough of people with ‘right mindset’ that can resist rent-seeking temptations and successfully maintain ‘good equilibrium’ by preventing their economic institutions becoming excessively extractive. This failure requires systemic changes to make things right. The idea is to use governance spending fixation as a tool for governance reform, which mainly consists of one act: lowering and fixing governance spending by legal acts, using governance spending per worker per year ratio with average brutto monthly salary of the country. Proposed reform would exert its effect on governance institutions via two channels: checking of resource misallocation and incentivising institutional progress.

After its implementation, all government sector workers beginning with top government members and ending in lowest rank bureaucrats are in a position where they all share a common salary pool that is fixed by proportion and not in total size. This implies that a government sector worker has three options how to increase personal income in the form of salary: (1) through increasing efficiency of governance sector and/or (2) through increasing numbers of non-governance workers and/or (3) through increasing salaries of local population.

The first option of salary increase gives incentives to increase efficiency of governance sector. This can be achieved by joining institutions, getting rid of redundant procedures, functions, positions, closing down unnecessary establishments, reducing number of government workers etc. Useless compliance work is likely to become the first victim of the proposed system once it is in place.

The second option of salary increase implies that there is a strong incentive to solve unemployment problem, as every additional worker automatically means bigger overall governance sector salary pool. The reform provides strong incentives to solve social problems, as this means more and better workers and more economic growth, all of which translates to bigger governance spending/salary pool.

The third option of salary increase is to increase salaries of local population. Raising salaries in government sector would also increase governance sector salary pool. Though proposed reform cannot directly increase salaries in private sector, it can affect them indirectly, by providing incentives to government sector to find best possible use to public resources so to best aid economy and people, which is the job of government anyway. It is known that government corruption alters public spending composition for rent-seeking purposes (see Hessami, 2014; Mauro, 1998). Making salaries directly dependent on carrying out governance functions in best possible manner not only gives incentives to do so, but more importantly, it creates an environment in government sector that is intolerant to rent-seeking activities, beginning with those that visibly waste resources and therefore do not make or prevent proper contribution of public funds to economy, like already mentioned alteration of public spending for rent-seeking purposes, disguised transfers, fraud, expropriation of public property, cronyism and theft of public funds, etc. It is unlikely that incentives to foster economic development and make best use of public resources and at the same time discouraging certain forms of corruption can fail to positively contribute to the overall economy, however, true reach of proposed reform concept goes beyond this.

The most important contribution to state governance of proposed reform concept is that by rewarding willingness to do the ‘right thing’ and encouraging intolerance to corruption and incompetence it creates environment in which good institutions and true professionals with knowledge, capability, desire to work for their people and full dedication to their profession can emerge and thrive. This is due to large salaries that are likely to appear with time when governance system is sufficiently optimised.

7. Conclusion

By now there is plenty of literature suggesting that institutions play a central role in development and it is thought that they work by creating different incentives. Literature study on the topic reveals that from the start of the twenty-first century scientific community largely converged at a consensus that can best be described as a call to take ‘the next step’ (Kaufmann & World Bank, 2003) that is to make use of large body of literature that is now present and to suggest reform concepts intended for application. At present, explanations on how institutions can be improved are rare and usually very general, therefore of limited use. In this context, present article takes full account of widely accepted wisdom that institutional factors come down to governance issues which ultimately sum up into ‘good/bad equilibrium’, and suggests that further progress can be made through study of interrelations of governance components with economic growth. The suggestion is complete with an example of how reform concepts can be developed using only basic math and theoretical framework that is available. Specifically, this article suggests how governance spending fixation could be used as a tool to improve institutions in nearly all comparatively poorly governed countries of the sample excluding Bulgaria. Namely those are: Romania, Hungary, Latvia, Portugal, Slovakia, Czechia, Lithuania, Greece and Italy. In the well governed countries in the sample that includes Norway, Switzerland, Germany, Ireland, Denmark, Netherlands, United Kingdom, Luxembourg, Finland, Belgium, and Sweden the latter four countries could potentially also find the suggestion useful. This is made possible by using elements of framework given in World Banks ‘Public Sector Governance and Accountability Series. Public Services Delivery’ (2005). From it, present article takes concept of governance quality evaluation using corruption and social development and some definitions that are then used together with insights found in other literature on the subject, to expose and map links of governance to other factors (see Figure 1). Mapping these links allows to identify ‘rent-seeking costs’ channel as the simplest way to make a positive effect to growth in corruption–growth relationship. Theory of existence of optimal level of governance spending is proposed. Radically simple manual matching exercise is used to reveal that best suited for obtaining a recommended value is governance spending per worker per year ratio with average brutto monthly salary of that country (variable B). The same method shows that while lowest governance spending values as measured by variable B are commonly found among best-governed countries, highest values are mostly found among countries with worst governance. This relationship gives reason to believe that average value of variable B of best-governed countries could potentially be used as a benchmark value for poorly governed countries to keep in check both misallocation of resources and excessive growth of bureaucracy. Optimal governance spending value was not found due to limitations of the method used and therefore remains an objective for future research. Finding it would enable application of the proposed reform concept to every country instead of only the worst-governed ones that benchmark value suggested by this article could potentially enable. Making governance sector salaries dependent on salaries of local population could ensure that institutions cannot become excessively extractive, thus be helpful in keeping ‘good equilibrium’. By setting rules so that government workers at all levels contribute to public good simply by acting in their own personal best interest to increase their own personal gain, proposed tool for reform could potentially enable greeting governance sector with invisible hand of Adam Smith, therefore creation of a dataset dedicated for tracking governance spending should be considered by Eurostat.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Governance spending is not a synonym to government spending. Rather it is a constituent of it. See Appendix in Data file for description.

2 See Gagliardi (2008) for literature analysis and references regarding the term ‘New Institutional Economics’.

3 See Gradstein (2008), Makuta and O’Hare (2015), Glaeser, La Porta, Lopez-de-Silanes, and Shleifer (2004), Pushak, Tiongson, & Varoudakis (2007), Huynh and Jacho-Chavez (2008), Vijayaraghavan and Ward (2001), Efendic, Pugh, and Adnett (2011), Nelson and Sampat (2001), Jetter (2014), Klomp and Haan (2009) among many others.

4 In all three main dictionaries (Merriam-Webster, Oxford and Cambridge) term ‘politics’ first is defined as activities associated with governance.

5 Gisselquist (2012) summarises working definitions of good governance from selected multilaterals in Table 1, p. 8. Author notes that ‘not only do definitions vary across organizations, they also vary within organizations’. Gisselquist (2012, p. 3) and provides five different governance definitions used by World Bank.

6 Malik (2002) gives a rough estimate of over 150 quantitative indicators on governance (p. 3). The World Bank Institute data portal on Actionable Governance Indicators is said to have 1182 of them (Wetzel, 2009, p. 10).

7 Kaufmann, Kraay, and Mastruzzi (1999b, p. 1) ‘…define governance broadly as the traditions and institutions by which authority in a country is exercised. This includes (1) the process by which governments are selected, monitored and replaced, (2) the capacity of the government to effectively formulate and implement sound policies, and (3) the respect of citizens and the state for the institutions that govern economic and social interactions among them.’ Definition then is used to organise a subset of governance indicators into six clusters: key aspects of the process by which those in authority are selected and replaced are summarised with clusters ‘Voice and Accountability’ and ‘Political Instability and Violence’; capacity of the state to implement sound policies is captured by two clusters ‘Government Effectiveness’ and ‘Regulatory Burden’; respect of citizens and the state for the rules which govern their interactions is captured by ‘Rule of Law’ and ‘Graft’.

8 There is no doubt that governance and development are linked, however, the direction of causality is debated at present point in time. Gisselquist (2012) presents six of the simplest causal possibilities illustrated by figure: ‘… first, good governance may ‘promote’ or ‘cause’ development (path A). Second, development may cause good governance (B). Third, another factor may cause both (C). Thinking in the disaggregated terms sketched above, it is also possible that some component of good governance may cause development (D); development may cause some component of good governance (E); or a third factor may cause both (F). To complicate the story still further, it could also be that some component(s) of good governance causes development (or some component(s) of development), while others contribute to economic stagnation, but that the effect of those that cause development is stronger. Alternatively, it could be that the interaction of several components of good governance causes development. Or, these ‘paths’ could operate sequentially, such that ‘B’ leads to ‘A’ or ‘C’ leads to ‘E’, and so on’ (p. 19).

9 Gisselquist (2012) highlights that six countries by their definitions tie governance specifically to development objectives (p. 9). The same link is made in some UN documents (Gisselquist, 2012, p. 8). The most explicit link by definition is made by World Bank (Gisselquist, 2012, p. 6).

10 In this article, terms ‘social development’, ‘welfare’ and ‘social progress’ are used interchangeably.

11 Yahyaoui, Chatti, and Chtourou (2008) state: ‘the question of the governance takes us then to that of the institutions and the study of the first one must necessarily pass by the study of the question of the ability of institutions to support the growth. Indeed, no clear distinction can be made between the concept of governance and institutional factor. Institutions generally correspond to a broader concept which includes the formal and informal constraints, rules and laws which are associated not only with the state's functioning but also with that of the private entities’ (p. 2).

12 To list of authors in Kaufmann and World Bank (2003; p. 16) it can be added: Kaufmann and Kraay (2008); Afonso and Furceri (2010); Karim, Zouhaier, and Adel (2013) and many others.

13 Definition found in Gisselquist (2012, p. 6) Table 1 is said to originally come from United Nations website.

14 As defined by Acemoglu and Robinson (2012), to be inclusive, economic institutions must feature secure private property rights, an unbiased system of law, a provision of public services that provides a level playing field in which people can exchange and contract; it also must allow people to choose their careers and permit the entry of new businesses, thus presenting economic opportunities not just for the elite, but for a broad cross-section of society so reducing the importance of monopolies and creating a dynamic and inclusive market economy and paving the way for two other engines of prosperity: technology and education, all of which reduces the economic benefits that can be secured in the short run by usurping political power. Economic institutions are referred to as extractive when they have opposite properties to those of inclusive and are called so because they are designed to extract incomes and wealth from one subset of society to benefit a different subset.

15 The notion that corruption is not solely a problem of bureaucracy is not limited to Kaufmann and World Bank (2003). de Ferranti, Jacinto, Ody, and Ramshaw (2009, p. 27) devote two chapters of their book to explaining oversimplifications of principal-agent model and give detailed description of multiple mismatches with reality of view that citizen-principals are public- spirited and like-minded. The same view is found in work of Kahana and Qijun (2010).