Fiscal Incentives and Shared Revenue Sources with Differentiated Sharing Rates

Abstract

A closer look at the estimating procedures that have been used to examine the empirical evidence of weak fiscal incentives at the local level in Russia in 1990s shows serious flaws in the treatment of proceeds from shared revenue sources with differentiated sharing rates widely employed in the intergovernmental fiscal relations in Russia at that time. This work analyzes two approaches suggested earlier to deal with proceeds from these sources when estimating fiscal incentives, and it develops an alternative method. Empirical results from this alternative approach suggest that the lack of fiscal incentives in Russia in mid-1990s was not as overwhelming as it appears from the estimates reported in the earlier works, which implies that the regional intergovernmental fiscal relations policies might have not been the likely cause of poor performance of the government and the economy in Russia at that time.

Keywords

fiscal federalism interbudgetary transfers incentives of local governments

The effects that fiscal incentives of subnational government can have on economic situation in a country have been studied in a number of articles mostly focusing on China (Qian and Roland 1998; Jin, Qian, and Weingast 2005) and Russia (Zhuravskaya 2000; Litwack 2002; Alexeev and Kurlyandskaya 2003). In particular, Zhuravskaya (2000) and Alexeev and Kurlyandskaya (2003) focus on fiscal incentives faced by Russia’s municipalities, and this is where the focus of this work is placed as well.

Zhuravskaya (2000) presents empirical evidence that shows that fiscal incentives faced by municipalities were almost nonexistent, and argues that the lack of such incentives in mid-1990s contributed to the poor performance of the government and the economy in Russia compared to other transitional economies, in particular China and Poland. Alexeev and Kurlyandskaya (2003) express doubt that regional governments would be interested in completely eliminating fiscal incentives as in such a situation local governments have no incentives to develop their revenue bases. They suggest that Zhuravskaya’s results might be wrong, and they proceed by reexamining her results while using their own data set from one of Russia’s regions. However, the estimates that they provide support Zhuravskaya’s findings by showing a complete lack of fiscal incentives faced by a sample of smaller, mostly rural, and less economically developed municipalities in the region. To justify these results, Alexeev and Kurlyandskaya suggest a number of plausible reasons that would explain why Russia’s regions chose to keep fiscal incentives faced by its municipalities at such a low level.

One might think that after this reexamination the case of fiscal incentives in Russian should be put to rest. However, a closer look at both works shows that there is an unresolved issue that neither work handles properly: dealing with the revenue sharing mechanism that included differentiated sharing rates widely used by Russia’s regions at that time. As I show in the following, the inability of each work to deal with this problem compromises the reliability of their empirical results as Zhuravskaya’s model suffers from misspecification, and Alexeev and Kurlyandskaya are forced to exclude from their sample the most economically developed municipalities in the region.

Previous Approaches and Their Problems

Both Zhuravskaya (2000) and Alexeev and Kurlyandskaya (2003) attempt to measure fiscal incentives by estimating the parameter k in the following equation:

Δ T = k \cdot Δ R + ϵ,

where $Δ T$ denotes changes in transfers, and $Δ R$ denotes changes in own revenues. However, there is a major difference in how Zhuravskaya and Alexeev and Kurlyandskaya specify equation (1) when separating transfers from own revenues or, more generally, the revenues that they put on the left-hand side and the right-hand side of equation (1). As I show in the following, while following the right intuition, Alexeev and Kurlyandskaya’s attempt to fix problems that are present in Zhuravskaya’s approach falls short of presenting a convincing argument and developing a well-justified alternative. My work completes this task by presenting an approach to the separation of the revenue components that should go on the left- and the right-hand sides of equation (1) that includes Alexeev and Kurlyandskaya’s method as a special case while showing how Zhuravskaya’s technique should be modified to specify equation (1) properly.

To understand the challenge facing Zhuravskaya and Alexeev and Kurlyandskaya, consider the system of intergovernmental fiscal relations between regional and local budgets that was prevalent in Russia in the 1990s. This system can be described as a gap-filling equalization scheme: the amount of funds provided to municipalities by regions was determined by the difference between municipalities’ estimated expenditure needs and the amount of revenues that they were expected to receive from own sources and permanently assigned sources, revenue sources that were shared with local governments by federal and regional budgets on the derivation base permanently. The sharing rates for permanently assigned revenue sources were typically uniform for all municipalities in the region and usually remained unchanged over a number of years. After estimating the gap for the planned fiscal year, the regional governments typically proceeded by assigning an appropriate portion of revenues from what can be called discretionary assigned sources (or sources that also were shared with local budgets on the derivation base but not permanently) by setting specific sharing rates for each municipality depending on the size of the gap in a particular fiscal year. For municipalities with relatively small gaps, a certain share of revenues from discretionary assigned sources was sufficient to close the gap, and this is how they usually got the general purpose financial assistance provided to them by regional governments (hereafter, type A municipalities). Municipalities with relatively large gaps typically got all proceeds collected at their territories from the discretionary assigned sources as well as a certain amount of general purpose intergovernmental grants (hereafter, type B municipalities). These grants were aimed to reduce the gap between municipalities’ estimated expenditure needs and expected proceeds from the own, permanently assigned, and discretionary assigned revenue sources. As the size of municipalities’ tax bases and their estimated expenditure needs fluctuated from year to year, the regions adjusted the amounts of resources that they provided to municipalities by changing the sharing rates for discretionary assigned sources for type A municipalities, and by changing the amount of intergovernmental grants provided to type B municipalities.

The major difference between how Zhuravskaya (2000) and Alexeev and Kurlyandskaya (2003) specify equation (1) is related to the revenue from discretionary assigned sources. Zhuravskaya puts changes in those revenues on the left-hand side of equation (1) treating them as transfers; Alexeev and Kurlyandskaya put them on the right-hand side treating them as own revenues. At first glance, it might seem obvious that as the derivation-based revenues are a form of transfer distributed according to a specific formula, that is, in proportion of where the tax is collected (e.g., Martinez-Vazquez, Timofeev, and Boex 2006), changes in those revenues should be treated as changes in transfers, and Zhuravskaya (2000) does the right thing by putting them on the left-hand side of equation (1). However, Alexeev and Kurlyandskaya (2003) put changes in these revenues on the right-hand side of equation (1) claiming that in the particular case they are dealing with this is the right thing to do.

Interestingly enough, Alexeev and Kurlyandskaya do not see a contradiction between their and Zhuravskaya’s approaches assuming that the difference between these two methods is justified by the difference between two samples that these two works are dealing with; that is, Zhuravskaya deals with a set of type A municipalities whose sharing rates were frequently adjusted, and Alexeev and Kurlyandskaya deal with a set of type B municipalities whose sharing rates remained stable. Thus, Alexeev and Kurlyandskaya rely on stability of sharing rates as it is perceived by local authorities when separating a group of municipalities for which discretionary assigned revenues should be treated as own revenues, arguing that as type B municipalities “could be reasonably sure that, on the margin, the changes in collections of [the] taxes would not lead to changes in the sharing rates . . . the task of separating own revenues from transfers becomes relatively easy.”

To understand Alexeev and Kurlyandskaya’s logic, one has to notice that they do not follow the conventional definition of own revenues as those revenues whose tax base and tax rate are set by local authorities. Instead, they follow the definition that was used in the Russian law at that time, and they included in the category of own revenues not just local revenues and fees but also shares of federal and regional taxes assigned to municipalities on a permanent base. Accordingly, Alexeev and Kurlyandskaya argue that, because for type B municipalities the sharing rates for discretionary assigned revenues remain unchanged for an extended period, these revenues can be treated as permanently assigned revenues and therefore put on the right-hand side of equation (1).

One problem with this argument is highlighted by the word reasonably, which indicates uncertainty and blurs the line that Alexeev and Kurlyandskaya are trying to draw between the two groups of municipalities. Under the mechanism of revenue sharing described earlier, the probability of sharing rates remaining stable in the nearest future is determined by the size of the gap between municipalities’ expenditure needs and the sum of their own, permanently assigned, and discretionary assigned revenues: the bigger this gap, the smaller the probability that the sharing rates would be adjusted in the near future. Alexeev and Kurlyandskaya refer to a large number of poor municipalities in the region for which this gap was well over half of their expenditures. Presumably, these municipalities could be “reasonably” sure that their sharing rates would not be adjusted any time soon. In response to this argument, one might ask, How about those municipalities for which this gap was, say, 20 percent, or 10 percent, or maybe even 1 percent? Clearly, a 1 percent gap is too small to assure a municipality that its sharing rates would not get adjusted sometimes soon. A 10 percent gap would be more assuring, but probably not quite. A closer look at the problem shows that given the volatility of the local tax base, no municipality can be absolutely sure that its sharing rates will not be changed any time soon. For example, the poorest municipality might become rich if oil is discovered within its borders.

Another problem with Alexeev and Kurlyandskaya’s attempt to separate a group of municipalities for which discretionally assigned revenues can be treated as permanently assigned revenues is what can be called the triggering effect. Consider a type B municipality whose tax base had been steadily growing until at some point the region decided to cut its sharing rates. For the sake of simplicity, assume that the local authorities remain “reasonably sure” that their sharing rates are not going to change until it actually happens. Presumably, when this municipality ceases to be a type B and turns into a type A municipality, a dramatic change in the treatment of its discretionary assigned revenues should take place as these revenues are swinging from the right-hand side of equation (1) in year t − 1 (the last year when the sharing rates remained stable) to the left-hand side of equation (1) in year t (the first year when the sharing rates were changed). No matter how small an adjustment in the sharing rates is, the whole bulk of the discretionary assigned revenues should be moved from the right to the left-hand side of equation (1), which creates a worrisome lack of continuity in this approach. A similar problem will take place if we consider a type A municipality whose deteriorating tax base eventually turns it into a type B municipality: presumably at some point the local authorities will get used to stable sharing rates and will be “reasonably sure” these sharing rates are not going to change at which point we will have to swing the whole amount of discretionary assigned revenues from the left-hand side to the right-hand side of equation (1). Perhaps, this lack of continuity prevented Alexeev and Kurlyandskaya from combining their approach when dealing with type B municipalities with Zhuravskaya’s approach when dealing with type A municipalities. Instead, Alexeev and Kurlyandskaya report their results for type B municipalities only and leave the question of what kind of fiscal incentives type A municipalities were facing unanswered.

The preceding criticism shows that Alexeev and Kurlyandskaya’s attempt to come up with a more reliable separation of those revenues that belong to the right-hand side and to the left-hand side of equation (1) does not reach its goal; that is, their argument to justify the decision to treat discretionary assigned revenues as permanently assigned ones for type B municipalities is not well grounded, and they neither explain how to deal with fiscal incentives faced by type A municipalities nor suggest how their approach can be combined with Zhuravskaya’s method if it is at all possible. Thus, there is still the need to address the problem. I turn to this solution next.

Filtering Out the Effect of Changes in Sharing Rates

To figure out how to deal with changes in revenue from discretionary assigned sources when measuring fiscal incentives, one must notice that the amount of revenue received by the local budget from a discretionary assigned source, DAR, is the product of the sharing rate, S, and the amount of collections from this source, C, and, therefore, its total change, $Δ D A R$ , can be calculated as the differential change:

Δ D A R = S Δ C + C Δ S .

The first component on the right-hand side of equation (2), $S Δ C$ , takes into account the change in collections $Δ C$ given that sharing rate S does not change, while the second component, $C Δ S$ , takes into account the change in the sharing rate $Δ S$ given that collections C do not change. There is a distinct difference between these two components in terms of the ability of the regional and local governments to control them. The latter component is under exclusive control of the regional government as it reflects its decision to adjust the sharing rates: this component clearly belongs to the left-hand side of equation (1). The former component—one resulting from changes in collections—can be controlled by local authorities as long as they are able to affect at least to a certain extent the following two factors in their jurisdictions: the size of the tax bases and fiscal discipline. By putting $C Δ S$ on the left-hand side while keeping $S Δ C$ on the right-hand side of equation (1), one can measure how much revenue the region claws back through the adjustments in sharing rates for discretionary assigned sources in response to increases in collections from these sources.

This separation of changes in discretionary assigned revenues can be applied to the two types of fiscal incentives effectively measured by Zhuravskaya and Alexeev and Kurlyandskaya. If, like Zhuravskaya, one would like to measure how much resources is withdrawn by the region in response to a given increase in own revenues, then of course $C Δ S$ should appear on the left-hand side of equation (1). $S Δ C$ should not appear on either side of the equation as this is neither the parameter that is being offset (which is the own revenues) nor the instrument of the offset (which is $C Δ S$ ). By putting $S Δ C$ on the left-hand side of equation (1), Zhuravskaya adds a component that is not only irrelevant to the offsets performed by the region but also correlates with changes in own revenues on the right-hand side of the equation as both parameters are linked to the size of tax bases and the level of fiscal discipline in the municipality, which is likely to create a positive bias in the estimate of k if such specification is chosen. As Zhuravskaya’s sample included relatively prosperous municipalities whose shares of discretionary assigned revenues were likely to be small, the distortive effect of putting $S Δ C$ on the left-hand side of equation (1) on the estimating procedure was probably modest. However, for municipalities with larger sharing rates, especially for type B municipalities whose sharing rates were set at their maximums, the distortion of the estimating procedure could be quite large.

If, like Alexeev and Kurlyandskaya, one wants to estimate fiscal incentives to raise own and assigned revenues, then $S Δ C$ should appear on the right-hand side of equation (1) along with changes in own revenues and changes in permanently assigned revenues. The left-hand side of equation (1) should include $C Δ S$ as in the previous case. When dealing with type B municipalities whose sharing rates remain stable, this approach reproduces Alexeev and Kurlyandskaya’s method where the left-hand side includes intergovernmental grants, and the right-hand side includes changes in own and assigned revenues.

Other specifications of equation (1) based on the separation of the effect of changes in the sharing rates described earlier can be arranged. In particular, it can be used to test whether higher-level officials respond the same to changes in own revenues as they do in responding to changes in the assigned revenues. These tests, however, are beyond the scope of this article.

To illustrate the applicability of this method and to take advantage of the available data, in the next section, I provide empirical estimates that complete Alexeev and Kurlyandskaya’s analysis by extending it to type A municipalities in Rostov Oblast.

Empirical Results and Their Interpretation

Based on the above separation of changes in discretionary assigned revenues, we can specify the left- and the right-hand sides of equation (1) in the following way:

Δ T = Δ M + C Δ S,

Δ R = Δ O w n + Δ P A R + S Δ C,

where $Δ M$ denotes changes in intergovernmental grants, $Δ O w n$ denotes changes in local taxes and fees, and $Δ P A R$ denotes changes in permanently assigned revenues.

Except for the separation of changes in discretionary assigned revenues into two components as suggested above, the following two models closely follow Alexeev and Kurlyandskaya’s models based on the comparison of actual and planned numbers:

\begin{aligned} T_{A, t} - T_{P, t} & = a + k_{1} \cdot (R_{A, t} - R_{P, t}) + k_{2} \cdot r i c h \cdot (R_{A, t} - R_{P, t}) + b_{3} \\ \cdot r i c h + b_{4} \cdot P o p + b_{5} \cdot Y 96 + b_{6} \cdot Y 97 + ϵ, \end{aligned}

\begin{aligned} T_{P, t + 1} - T_{P, t} & = a + k_{1} \cdot (R_{A, t} - R_{P, t}) + k_{2} \cdot r i c h \cdot (R_{A, t} - R_{P, t}) + b_{3} \\ \cdot r i c h + b_{4} \cdot P o p + b_{5} \cdot Y 96 + ϵ, \end{aligned}

where subscripts A and P denote the amounts of either own revenues or transfers actually received by the municipality or planned by the region for the municipality, Pop—municipal population, rich—dummy variable for type A municipalities, and Y96 and Y97—the respective year dummies.

Model (5) implies that the region adjusts intergovernmental grants and sharing rates during the same year in which actual own and assigned revenues deviate from the planned numbers, and model (6) implies that this deviation induces the region to make adjustments in the following year. The regressions are run using ordinary least squares (OLS), fixed effects, and random effect techniques. The estimates of the coefficients in these two models are presented in table 1.

Table 1.

Regression Results for Equations (5) and (6).

Independent variable	Dependent variables, per capita
	Equation (5)			Equation (6)
	Changes in the same year grants and assigned revenues $T_{A, t} - T_{P, t}$			Changes in the following year planned grants and assigned revenues $T_{P, t + 1} - T_{P, t}$
	OLS***	Fixed effects	Random effects	OLS***	Fixed effects	Random effects
Constant	−67.2 (28.3)	−36.4 (22.9)	−67.2 (29.2)	−248.6 (19.2)	−217.5 (18.9)	−248.6 (19.2)
$R_{A, t} - R_{P, t}$	0.11** (0.24)	0.09** (0.24)	0.10** (0.22)	−0.64 (0.19)	−0.53 (0.27)	−0.64 (0.19)
$R i c h * (R_{A, t} - R_{P, t})$	0.23 (0.32)	0.27 (0.31)	0.25 (0.29)	0.91** (0.27)	0.87* (0.39)	0.91** (0.27)
Rich	127.4 (43.8)	—	127.3 (55.8)	110.7 (37.9)	—	110.7 (37.9)
Pop	0.13 (0.12)	—	0.13 (0.15)	0.16 (0.10)	—	0.16 (0.10)
Year 96	51.5 (39.7)	52.5 (34.7)	52.0 (33.8)	194.4 (25.9)	189.2 (29.0)	191.4 (25.9)
Year 97	20.8 (36.6)	21.2 (31.1)	21.0 (30.8)	—	—	—
Adjusted R ²	0.09	0.04	0.12	0.39	0.30	0.42

Note: OLS = ordinary least squares. The Hausman test does not reject random effects specifications. Standard errors are in parentheses.

*The coefficient is greater than 0 at 5 percent significance level.

**The coefficient is greater than (−1) for k ₁ and greater than 0 for k ₂ at 1 percent significance level.

***Cluster robust standard errors provided in parentheses.

These results suggest several conclusions. The estimates of parameters in model (5) show that neither type A nor type B municipalities were facing clawbacks resulting from the adjustments in the same year adjustments in intergovernmental grants or discretionary assigned revenues, which is consistent with Alexeev and Kurlyandskaya’s finding for type B municipalities. The estimates of parameters in model (6) are also consistent with Alexeev and Kurlyandskaya’s findings for type B municipalities as these show that type B municipalities were facing a complete clawback of changes in their own and assigned revenues through adjustments in the next year planned intergovernmental grants and discretionary assigned revenues. This consistency with Alexeev and Kurlyandskaya’s results for type B municipalities is to be expected as the approach suggested in this work includes Alexeev and Kurlyandskaya’s method as a special case when sharing rates remain stable, which is the case for type B municipalities. The evidence of the clawback effect produced by changes in the following year planned transfers is even stronger than in Alexeev and Kurlyandskaya’s work as not just two, but all three specifications for model (6) produce the estimates of the coefficient k ₁ that are not significantly different from minus one.

However, when it comes to type A municipalities—that is, those municipalities whose revenues were adjusted through changes in sharing rates—the results are strikingly different from those received for type B municipalities. According to the estimates, type A municipalities were able to avoid a complete clawback of changes in their own and assigned revenues not only within a current fiscal year but also when the time came to reevaluate their fiscal aid in the following year. The difference between the regional policy toward type A and type B municipalities regarding fiscal incentives is supported by the estimates of the coefficients of the dummy variables for type A municipalities in interaction with changes in own and assigned revenues in model (6). The hypothesis that the dummy coefficients are smaller than zero can be rejected at 1 percent significance level for OLS and random effect specifications and at 5 percent significance level for the fixed effects specification.

The estimates received in this work show that as Alexeev and Kurlyandskaya demonstrated previously those municipalities in Rostov Oblast in 1996–1998 whose changes in own and assigned revenues were adjusted by changes in intergovernmental grants (type B municipalities) were indeed facing a complete lack of fiscal incentives. However, these incentives were quite strong for those municipalities whose changes in own and assigned revenues were adjusted by changes in sharing rates (type A municipalities).

There are two questions that these findings raise. One question is why discretionary assigned revenues are so much more “stickier” than intergovernmental grants when it comes to adjusting changes in municipalities’ revenues by the region. There are two possible explanations of this observation. One is that the richer municipalities might be considered by the region as locomotives of economic growth whose incentives to foster local tax base and improve fiscal discipline should not be compromised even for the sake of equalization. At the same time, the economic potential of the poorer municipalities might be considered too low and much beyond repair to worry about providing them with fiscal incentives at the cost of equalization. Such a situation will be consistent with an equalization mechanism where the region maintains a bottom line below which no municipality can drop regardless of how weak its local tax base is. As long as a municipality stays in the category of poor municipalities, any changes in its economic situation would not affect the overall amount of revenue available to it neither if the economic situation improves nor if it deteriorates, which is good news for equalization but bad news for incentives. Under such an equalization mechanism, only relatively rich municipalities will enjoy overall increase in their revenues if their economic situation improves, the downside of which is that if their economic situation deteriorates then the negative effect on the overall amount of revenue available to them will be also quite noticeable, which is a good news for incentives but a bad news for equalization.

A second explanation is the difference between how regional officials perceive assigned revenues and intergovernmental grants: assigned revenues might be taken as if local authorities had some sort of claim in these revenues while intergovernmental grants might be taken as resources in which local governments have no claim. This attitude is reflected in the Russian law that not only includes in the category of “own” revenues permanently assigned revenues but also requires that a certain share of regional revenues was shared with local budgets on average. This requirement does not make those revenues municipalities truly own according to their conventional definition that defines own revenues as those for which a government should set tax rates or tax base (e.g., imposing a tax burden on its own accord and at its own responsibility). However, it might create an impression that the assigned revenues somehow “belong” to local budgets and discourage regional officials from taking these revenues away from local budgets as freely as they take away intergovernmental grants.

A second question that these findings raise is how can one explain the discrepancy in the results for type A municipalities received in this work and Zhuravskaya’s estimates, whose analysis covers type A municipalities as well but produces very different results? This discrepancy cannot be explained by the difference in the specification of equation (1) because, as explained earlier, Zhuravskaya’s misspecification should produce a positive bias in the estimates, not a negative one. Nevertheless, her results show a complete offset of changes in municipalities’ revenues (which implies that coefficient k in equation [1] is equal to minus one), while the results received in this work show no such offset (which implies that coefficient k is equal to zero).

One might try to explain this discrepancy by arguing that both results are correct but that they reflect the situation in different time periods as Zhuravskaya’s data cover the period from 1992 to 1997, while the data used in this work cover the period from 1996 to 1998. One might interpret this discrepancy as the evidence that fiscal incentives in Russia improved from the early to the late 1990s. Also, one might argue that Rostov Oblast was not representative of the rest of the country in 1990s in providing strong fiscal incentives to its larger type A municipalities.

However, a more probable explanation of this discrepancy is that the self-selection bias that Zhuravskaya’s estimates are likely to exhibit. The self-selection might have occurred not in one but in two stages as Zhuravskaya collected her data. First, the questionnaires used by Zhuravskaya were sent to a relatively small self-select group of eighty cities members composing the League of Russian Cities. It is likely that those municipalities that were facing a larger clawback were more likely to join the league to protect their interests more effectively. Second, of the eighty members of the league to whom the questionnaires were sent, only thirty-five municipalities provided usable data, and it is likely that only those municipalities whose authorities felt that they were particularly mistreated by the regions as a result of the clawback had the incentive to put the effort into collecting these data.

Conclusion

One of the reasons that have been put forward to explain why Russia lagged behind other countries in economic growth in the 1990s is the inefficiency of its intergovernmental fiscal relations, in particular the lack of fiscal incentives faced by local governments as the regions presumably deprived them of fiscal independence by offsetting any increase in own revenues with changes in transfers. The main empirical evidence supporting this view is presented by Zhuravskaya (2000), who finds a complete lack of such incentives in a sample of thirty-five large Russian cities in twenty-nine regions for the 1992–1997 period. Finding these results implausible, Alexeev and Kurlyandskaya (2003) revise them using their own data set, but their estimates provide similar results at least for a subset of smaller, mostly rural, and less economically developed municipalities. A closer look at Alexeev and Kurlyandskaya’s work shows that their focus on a particular subset of municipalities is caused by their inability to handle shared revenue sources with differentiated sharing rates when measuring fiscal incentives faced by larger, mostly urban, and more economically developed municipalities. As a result, they focus their attention on municipalities with stable sharing rates and report their empirical results for this specific group of municipalities rather than for all municipalities in the region. The larger, more urbanized, and more economically developed municipalities are left outside of the scope of their work, which significantly limits the applicability of their results, especially for the purpose of revising Zhuravskaya’s findings as her sample includes those municipalities whose characteristics are similar to those municipalities that Alexeev and Kurlyandskaya effectively excluded from their analysis.

Thus, the goal of this article is to answer the following question: why Alexeev and Kurlyandskaya are not able to deal with shared revenues with differentiated sharing rates and how these revenues should be handled when measuring fiscal incentives? I come up with an answer to the first part of this question by arguing that what prevents Alexeev and Kurlyandskaya from resolving this problem is that they are relying on the Russia’s law definition of own revenues, which takes stability and predictability as the core characteristics of own revenues, rather than following the conventional definition of own revenues as those revenues that come from the sources for which local governments set tax bases and tax rates. Accordingly, Alexeev and Kurlyandskaya treat shared revenues with fixed sharing rates as municipalities’ own. On the contrary, I argue that shared revenues should be treated as transfers and not as own revenues regardless of whether they are permanently assigned to local budgets or used as a tool to adjust the amount of resources available to them.

This claim on its own would not be sufficient, as essentially that is the logic that Zhuravskaya follows when separating own revenues from transfers in her work. The cornerstone of this work is an observation that any given change in the amount of revenue from a shared source can be separated into two components: one, attributed to changes in sharing rates, and another—attributed to changes in collections, and only the former component should be treated as an instrument that the regions use to adjust the amount of resources available to local budgets. This approach, with some qualifications, informs Alexeev and Kurlyandskaya’s technique when applied to municipalities with stable sharing rates, but it conflicts with Zhuravskaya’s treatment of changes in shared revenues.

When this approach is applied to the data set from Rostov Oblast, Alexeev and Kurlyandskaya’s findings are confirmed as the results show that the subset of smaller, mostly rural, and less economically developed municipalities in Rostov Oblast were facing a complete lack of fiscal incentives. However, the rest of the municipalities in the region were facing quite strong fiscal incentives. Thus, as Alexeev and Kurlyandskaya suspect, Zhuravskaya’s results might be wrong; specifically, the lack of fiscal incentives in the 1990s was not as prevalent as she argues, at least when it comes to larger, more urbanized, and economically developed municipalities. On a larger scale, this observation contradicts Zhuravskaya’s claim that Russia’s regions’ inefficient intergovernmental policies was an important factor contributing to Russia’s economic decline in 1990s.

In the end, what does this analysis suggest? Perhaps revisiting two closely connected questions—what was happening with fiscal incentives at the local level in Russia in 1990s and how should shared revenues with differentiated sharing rates be treated when measuring fiscal incentives—would not be justifiable for their own sake. However, as my work demonstrates, the problem is not that not enough has been done to address these questions but that what has been done has not been done right: Zhuravskaya overlooks the role that shared revenues with differentiated sharing rates play in determining fiscal incentives, which results in the misspecification of her empirical model, and Alexeev and Kurlyandskaya’s attempt to figure this problem out falls short of reaching the goal, which forces them to leave an important subset of municipalities in a region out of their analysis.

This work corrects Zhuravskaya’s and Alexeev and Kurlyandskaya’s mistakes and improves our understanding of how to analyze the fiscal incentives of shared revenues. The immediate effect of this improvement is the empirical evidence that demonstrates that not all municipalities in Russia in 1990s were lacking fiscal incentives as claimed in the previous studies, and, therefore takes the blame off presumably inefficient regional intergovernmental policies as one of the contributors to the country’s economic decline at that time. The longer-term effect of this improvement is our ability to deal with shared revenues with differentiated sharing rates should we run into this mechanism in one of the former centrally planned economies again or should the idea of using differentiated sharing rates cross the minds of inventive policy makers somewhere else.

Footnotes

Acknowledgements

I am grateful to Jorge Martinez-Vazquez, Andrey Timofeev, and the anonymous referees for their comments and discussion. I am also grateful to Michael Alexeev and Galina Kurlyandskaya for providing the data used in this work.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Alexeev

Michael

Kurlyandskaya

Galina

. 2003. “Fiscal Federalism and Incentives in a Russian Region.” Journal of Comparative Economics 31:20–33.

Jin

Hehui

Qian

Yingyi

Weingast

Barry R.

. 2005. “Regional Decentralization and Fiscal Incentives: Federalism, Chinese Style.” Journal of Public Economics 89:1719–42.

Litwack

John M.

2002. “Central Control of Regional Budgets: Theory with Applications to Russia.” Journal of Comparative Economics 30:51–75.

Martinez-Vazquez

Jorge

Timofeev

Andrey

Boex

Jameson

. 2006. Reforming Regional-local Finance in Russia. Washington, DC: The World Bank.

Qian

Yingyi

Roland

Gerard

. 1998. “Federalism and the Soft Budget Constraint.” American Economic Review 88:1143–62.

Zhuravskaya

Ekaterina V.

2000. “Incentives to Provide Local Public Goods: Fiscal Federalism, Russian Style.” Journal of Public Economics 76:337–68.