ECONOMIC EFFECTIVENESS OF THE DEVOLUTION OF POWER TO PROVINCIAL COUNCILS IN SRI LANKA: A RESOURCE-PRODUCTIVITY ANALYSIS

The devolution of power through the establishment of Provincial Councils in Sri Lanka is a highly debated subject, but hardly on economic grounds. The present research addressed the subject through the analysis of recurrent and capital expenditure patterns of the Provincial Councils in comparison to those of the National Government. The outcomes of the study clearly shows that the Provincial Council system is still overwhelmingly dependent on Government grants, even for its recurrent expenditure requirements, after 24 years of its existence. The study revealed that the Sri Lankan process of devolution has not been founded on any enabling economic justifications, and has not been able to produce any cost economics, generally expected through economically rational devolution of power. But, the study found that the process has resulted in negative economic implications, particularly by its apparent retarding effect on public investment, to the effect that over half-a-percent additional GDP growth push could be secured annually if the Provincial Council system could be abandoned. The devolution of Power through the Provincial Council system, brought about by the 13 amendment to the Constitution, cannot therefore be concluded as justifiable on economic grounds.


INTRODUCTION
Devolution of power in Sri Lanka was a heavily debated subject ever since the introduction of the Provincial Council system in 1988 based on the 13 th Amendment to the Constitution. 1 Though the advocates of the Provincial Council system (PAFFREL, 2013;Siboti, 2013;Knutsen, 2004;Satyendra, 1988) attempt to justify devolution on the grounds that the centralised system of management has failed to satisfy the aspirations of the people while devolution will guarantee that the developing world economies mitigate the effects of corrupt management in central government and as such will grant these devolved systems the power to ensure their endurance in the competitive global economy (Siboti, 2013). Also, there was growing insistence in Sri Lanka on the necessity of decentralisation of administrative processes in order to achieve rapid economic and social development of the country (thirteenth amendment to the constitution of Sri Lanka, 2013). 2 However, there are divergent opinions expressed regarding its effectiveness and also with respect to future action. Some argue for much deeper devolution of power to Provincial Councils (commonly referred to as 13+) in response to demands for more "autonomy" by the Northern political forces, while others demand the complete abolition of the Provincial Council system out of fear that it would eventually threaten the "unitary status" of the Country's political administration. Several others suggest continuing with the system, but with further constitutional amendments to remove some of its provisions.
It is a fact that Sri Lanka's devolution was politically motivated, and therefore, it is unclear as to whether the economic incentives for devolved management of the affairs of the nation were given due consideration. According to literature, fiscal decentralization may lead to economic efficiency, cost efficiency, accountability and resource mobilization; (Knutsen, 2004) but reaping these benefits, requires the prior existence of significant local administrative capacity, responsive officials, substantial discretionary financial control and geographic incentives (such as large land areas and distantly spread regional activities) which could offer scope for saving on administration and coordination costs (Bird and Vaillancourt, 1998).
Unfortunately, the research that has hitherto been conducted on the Sri Lankan case of devolution have been mainly on the political aspects and does not entail a significant analysis of these economic and developmental aspects of devolution. 3 Economics of provincial management in general, and the impact of Provincial Councils on the 1 Through the Provincial Councils Act No 42 of 1987(Provincial councils, 2014. 2 Government website (http://www.priu.gov.lk/ProvCouncils/ProvicialCouncils.html) 3 One rare exception is the study conducted at the Institute of Policy Studies by Waidyasekara (2004), which point outs several inherent weaknesses and deficiencies of the "fiscal devolution" activity that has assumed importance in Sri Lanka for political reasons in the context of the ethnic crisis. country's overall resource utilisation efficiency in particular, have not yet been adequately appraised.
The present study was conducted with the objective of appraising the economic impact of devolution in Sri Lanka, particularly through a fiscal effectiveness stand-point. It focused particularly on the behavior of the economy's overall fiscal expenditure efficiency, before and after the establishment of Provincial Councils, in order to derive lessons for future. A comparative analysis among the provincial councils was also attempted in order to fathom the differences in relative expenditure intensities. It was also intended that the research would contribute towards enriching the discussion on further action pertaining to devolution of power based on the 13 th Amendment, particularly by adding the economic dimension in to it.

MATERIALS AND METHODS
The research approached the question via appraising the recurrent and capital expenditure patterns of Sri Lanka's economic activity before and after coming into effect of the Provincial Council system. The economic effectiveness of the provincial management structure was examined on the premise that there would be administrative cost efficiencies to be gained, such as advantages of geographic proximity to user communities (Bird et al, 1998) and the possibility of area specific service production without having to bear undue communication, networking or chain-of-command related costs considered to be inherent in centralised or unitary management of public service provision (Samaratunge, Bennington, 2002). Such advantages, if any, would be reflected in a reduced level of recurrent expenditure intensities after the devolution, compared to those before.
When analysing the recurrent and capital expenditures between the national Government and the Provincial Councils, care was taken to include only those comparable expenditure headings. This was because an inclusion of those expenses on activities that are not performed by the Provincial Councils (such as defense, foreign debt service payments or financial allocations to Provincial Councils) would make the analytical bases incomparable.
The analysis was performed using secondary data, sourced from the Annual Reports of the Central Bank of Sri Lanka, the Ministry of Finance and Planning, the National Finance Commission reports and the Financial Statements of the Provincial Councils.
The period from 1981 to 2012 was included in statistical comparisons where the data pertaining to the period after 1990 were considered to be reflecting post-devolution patterns, as separate financial data pertaining to Provincial Councils were not available until 1990, even though the Provincial Council system came into effect in 1988 4 . For the comparative analysis among the Provincial Councils, the authors used the data pertaining to the year 2012 which was the latest available on all Provincial Councils except for the Northern Provincial Council. 5 Graphical observation of comparative trends, analysis of recurrent and capital expenditure ratios, statistical comparison of averages, regression analysis and an investment productivity appraisal were used as analytical methods.

ANALYSIS AND RESULTS
The first and foremost observation that emerged through comparative analysis of expenditure patterns was that the Provincial Councils have been overwhelmingly dependent on Government grants. The highest percentage of total expenditure financed through the "earnings" of the Provincial councils during the period between 1990 and 2012 was 29% (average was 21%) and the share of recurrent expenditure and capital expenditure of the Provincial Councils met through the National Government allocations were never less than 68% and 71% respectively (Table 1).
This implies that the Provincial Councils, even after over two decades of existence, have not yet been able to become "financially self-reliant", and continue to be dependent on the resources pumped in from the parent National Government for their existence.
The information in the Table 1 also reflects the disproportionately high share (nearly 85%) of total expenditure of the Provincial Councils being recurrent. It is also noticeable that the Provincial Councils, in some years, have invested less than the capital grants received from the national coffers, pointing at the possibility of misappropriation of funds, 6 where the possibility of meeting recurrent expenditures through votes granted for development effort could not be excluded.

Recurrent Expenditure Analysis
The question was further examined through a comparative analysis of recurrent expenditure patterns of the National Government and the Provincial Councils after the devolution, the results of which are depicted in the Figure 1.  (1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012) It is evident from the Figure 1 that an overall declining trend in recurrent expenditure intensity of GDP 7 (which is naturally expected through increased per capita GDP and increased economic productivity in a growing economy) can be observed over the years after the devolution. The trend lines however indicate that the National Government was 7 Share of recurrent expenditure as a ratio of GDP managing to reduce its recurrent expenditure as a share of National GDP much faster (tangent of -0.0018) than the Provincial Councils (tangent of -0.0006), both tangents being statistically significant (Table 2)

Figure 1: Evolution of Recurrent Expenditure since Devolution
Source: Author's calculations What can be hypothesized through these results would be either (i) the relative "inefficiency" of Provincial Councils in managing their "consumption" budgets in comparison to national Government during the post-devolution period, or (ii) a functional peculiarity associated with devolution which might have allocated to Provincial Councils those highly "recurrent intensive" activities (such as health services and education) which are difficult to control. It is generally understood that personal emoluments constitute a significant component of the recurrent expenditure of public service, both at the national and provincial levels.
The study attempted to do a comparative examination of these expenditures in order to shed light on the reasons behind the above perceived differential evolution of recurrent expenditure intensities between the central and devolved administrations.

Figure 2: Evolution of Personal Emoluments after Devolution
Source: Author's calculation It is notable (in the Figure 2) that the curve corresponding to the national Government is positioned below that of the Provincial Councils in the case of personal emoluments, whereas the positioning was the opposite with regard to curves pertaining to recurrent expenditure ( Figure 1). This indicates that the recurrent expenditure of the national Government, in the early years of devolution, has been much less "bureaucratic intensive" 8 than that of the Provincial Councils, bringing further evidence to support the hypothesis made earlier that the pattern of the allocation of functions at the time of devolution would have been "skewed".
It is interesting, however, to note that the gap between the two curves has been narrowing over the years, and they are much closer to each other by 2012, reflecting an increasing trend of "bureaucratic intensity" of the functioning of the national Government while the Provincial Councils, having started from a much higher level, have managed to better their bureaucratic intensity over the years. This could reflect a few possibilities. First, it could be an indication of the provincial administrations being more successful than the National Government in managing their human resources with increasing productivity over the years. Starting with a much lower intensity level than that of the Provincial Councils, the National Government, over the years, has not been able to manage its bureaucratic intensity gap compared to Provincial Councils. This could possibly be a reflection of the increased presence of wasteful duplication of basic functions, which figures among the possible "adverse effects of devolution" discussed by Andres Rodr Guez-Pose and Nicholas Gill (2005). 9 Taking over several provincial hospitals and schools under the central administration, constructing provincial and local roads by the central authorities and duplicated bus passenger transport administration could be among examples of increased central Government intervention into constitutionally devolved subjects, which would have prevented the central bureaucracy from keeping pace with Provincial Councils in reducing bureaucratic intensity. Second, the National Government being able to reduce its overall recurrent intensity much faster than the Provincial Councils (as depicted in the Figure 1), in spite of it being unable to maintain the gap against the Provincial Councils with regard to bureaucratic intensity, could mean more and more emoluments being spent by the national Government to manage less and less "non-emolument oriented" recurrent expenditures. This possibly reflects deteriorating human resource productivity at the central administration. Third, this could also be read as a reflection of increasingly more "non-emolument type" expenditures incurred per rupee of personnel emoluments paid by the Provincial Councils compared to the National Government. While this could be argued as a reflection of relatively high human resource productivity of Provincial Councils as against the central Government, the possibility of this being a reflection of the relative presence of wasteful spending in "non-emolument based" recurrent expenditures at Provincial level cannot be excluded.
This prompted the authors to examine the bureaucratic expenditure patterns of each of the Provincial Councils in detail, and in comparison with those of the National Government, the results of which analysis are summarised in Table 3. This analysis mirrors the overwhelmingly dominant share of personal emoluments in recurrent expenditures of the Provincial Councils (over 80%, except with regard to the Western Provincial Council) compared to the national Government (Table 2).
While it is quite possible that this is a result of the above hypothesised "skewed" devolution of functions by the 13 th Amendment to the constitution, the question pertaining to the overwhelming "consumption orientation" of the Provincial Councils cannot be ignored. The very high bureaucratic intensity of Provincial Councils (possibly excluding the Western Provincial Council, to some extent) prompts us to question the economic rationale of their existence as barely any development orientation is seen through these expenditure patterns. The question of re-unification of these functions naturally becomes pertinent (particularly if these high bureaucratic intensities are owing to any other reason than a mere "skewness" in functional devolution) as any "economies of scale" that could be secured, and any bureaucratic expenditure savings that could be realised, through such re-unification would amount to an "efficiency gain", sparing more investable resources, thus enabling greater growth potential. This observation prompted the study to research into the apex provincial administrative burden, which could possibly be avoided if the Provincial Council system were not to be. In such a scenario, the costs of the service wings of the public service (such as health, education, etc.), now under the Provincial System, would persist under a reunified management of affairs by the national administration, but the apex provincial administration and political management would be done away with. The recurrent expenditure heading of each Provincial Council was thus categorised into their individual sub-headings to identify those cost items which would be "specific" to Provincial Council system, and would not be incurred should the Provincial Council system be abolished. The Table 4 summarises the results of this analysis.
As per the results of this analysis, Rs 12 Bn. per year of recurrent expenditure (nearly 9% of the recurrent expenditure of the Provincial Councils or over 1% of the country's overall public spending in 2012) would become "unnecessary" if the Provincial Council system were to be abolished. Unlike in the case of the "public service functions" delivered by the Provincial Councils, this apex political and administrative over-burden of the devolved power units will be a direct saving to the economy if the Provincial Council based devolution of power is to be reversed. Such saving could be added on to the Gross Domestic Capital Formation of the economy under the hypothesis that the entire amount savable would be invested by the national Government, thus generating an additional growth impetus.
The recurrent expenditure analysis thereby enables a few interesting inferences. First, the devolution in Sri Lanka does not appear to have produced any economies of scale advantage as no significant gains in recurrent expenditure intensity of economic activity, over and above any naturally expected trend in a growing economy, could be observed at the national or provincial levels after the implementation of devolution.

Table 4: Recurrent Expenditure on Apex Political and Administrative System of PCs
Second, the devolution, as it has been undertaken under the 13 th amendment, appears to be "skewed" with regard to its functional allocation where more "bureaucratic intensive" functions have been devolved. This is likely to be the cause behind the relatively lower bureaucratic intensity of the National Government compared to that of the Provincial Councils in the early phases of devolution. Duplication by the National Government of such "highly personal emolument intensive" devolved functions as well as taking back of some elements of such functions under the central administration, could possibly explain why the National Government has not been able to keep pace with the Provincial Councils in reducing bureaucratic intensity. Third, the devolution in Sri Lanka appears nothing but "handing over" to Provincial Councils of several functions hitherto managed by the National Government, where the individual "regional branches" of the formerly national hierarchical management structures (such as operational wings of the Education and Health departments) became simply detached and brought under Provincial Councils. The devolution thus appears to have not produced any administrative economies. On the other hand, significant savings could be envisaged through re-unification of structures, particularly by doing away with the overarching political and administrative bureaucracies in the Provincial Councils (which would be made redundant if devolution under the 13 th amendment were to be abolished); which, if invested, could fuel growth.

Analysis of Capital Expenditure Patterns
The capital expenditure patterns of the Provincial Councils were examined in order to perceive the effects of devolution on investment. When examined in proportion to total expenditures incurred, a breakdown of the "investment share" of overall public spending, since the coming into effect of Provincial Councils, could be observed (Figure 3). While the apparent relatively high investment intensity 10 of national Government's expenditure compared to that of the Provincial Councils, particularly in the earlier years after devolution, could be explained as an outcome of an unbalanced or skewed structure of functional devolution (in which highly service intensive activities were devolved while retaining the development oriented functions at the centre), such reasoning could not explain the overall reduction of investment intensity in the economy's public expenditure. In fact, the average investment share of 46% of the overall public expenditure of the economy prior to devolution has come down to 31% after devolution, the difference being statistically significant (t-value: 6.1). Even though an increasing trend in the investment intensity of overall public expenditure since devolution could be observed over the years, that also appears to be driven essentially by the capital expenditure intensity of the national Government spending. 11 A preliminary hypothesis could therefore be advanced that the devolution would have caused an adverse effect on the overall capital formation drive of the economy's public investment. This overall effect was further studied through an Ordinary Least Square regression analysis where the investment intensity of the overall public spending (Provincial Councils and the national Government together) was examined. In this analysis, the national political orientation towards "State interventionism" was represented by the fiscal deficit ratio the successive Governments chose to run, 12 while the effects of "internal insurgency" and "devolution" conjunctures were examined by introducing two dummy variables.  A national political regime willing to run high budget deficits could spend more on public investment than a political regime with a neo-liberal ideology. This effect was captured by the introduction of budget deficit ratio as an explanatory variable, which was also proven significant in the regression analysis.
As per the results of the regression analysis summarised in Table 5, the R 2 and tstatistics confirm that nearly 87% of capital to total public expenditure ratio is explained by the selected independent variables and all three variables have statistically significant effects on the overall investment intensity of public spending. Also, the results bring further evidence to strengthen the hypothesis (advanced through the graphical examination) that devolution has had a significant negative impact on the investment orientation of the country's overall public spending. Quantitatively, this revelation estimates that approximately 18.4 cents more could have been invested per each rupee publicly spent, if the Provincial Council system did not exist.
This enables estimation of the additional growth impetus under the hypothetical condition of the absence of Provincial Councils, where the public investment would increase with no augmentation of the overall fiscal burden. For instance, an 18.4% growth of investment intensity of public expenditure (or approximately Rs 160 Bn in 2012, of which nearly 8% would correspond to the direct recurrent expenditure savings that would be secured by getting rid of the provincial apex political and administrative over-burden) would mean a supplementary public investment ratio in the economy of around 2.3%; the possible augmenting effect of which on the country's GDP would be around half-a-percent under the prevailing average level of Sri Lanka's Incremental Capital Output Ratio (ICOR) 13 of 4.4.

CONCLUSIONS
The study leads to the inferences that (a) the provincial management in Sri Lanka has been overwhelmingly dependent on the grants received from the national Government indicating that one of the widely recognised conditions for successful devolution, namely the control over its own finances, has been quasi-absent, (b) the process has not generated any significant administrative cost savings that are generally expected through economically rational devolution, (c) inefficiencies appear to have stemmed in through wasteful expenditures and duplicity of functions, particularly with the central Government showing signs of bureaucratic expansion even after significant devolution of power to Provincial Councils, (d) the management of public funds in favour of capital formation has been negatively impacted during the post-devolution period, and (e) an abolition of the Provincial Council system could generate an additional GDP growth push of over around half-a-percent, nearly one-tenth of which would correspond to direct economic gains corresponding to saving of the recurrent expenditure on apex political and administrative bureaucracy of the Provincial Councils and investing such savings in development ventures not less productive than what is reflected by the prevailing ICOR of the economy.
Based on the above analytical inferences, it could be reasonably concluded that the devolution of power in Sri Lanka through the establishment of Provincial Councils enacted by the 13 th amendment of the Constitution, has not been based on any economic rationale; neither has it been able to demonstrate any economic justification through its existence over 25 years. On the contrary, the study reveals that the economic impact, if any, has been negative, in relation to public investment and to the economic growth potential therein.
The study also reveals that Sri Lankan case could possibly be another demonstration of the possible "economic ill-effects" of devolution when done in the absence of "prerequisite conditions" and where a nation would be forced to forfeit any transaction cost internalisation advantages it may have enjoyed prior to devolution (Bird and Vaillancourt, 1998). Further research requires the collection of more refined and disagregated data on provincial and national expenditure to investigate and identify such enabling and disabling determinants as may have been active with regard to Sri lankan devolution of power.