Asymmetric Yardstick Competition: Traditional Procurement Versus Public-Private Partnerships

We investigate yardstick competition between local jurisdictions in which pure rent-seeking incumbents undertake an identical infrastructure project choosing be- tween two contractual arrangements with different financing profiles, namely traditional procurement (TP) and public-private partnership (PPP). We show that a mixed regime, in which TP is used in one jurisdiction and PPP in the other, is likely to arise when projects are mildly lucrative, and/or jurisdictions have a moderate fiscal capacity. We find that, in the mixed equilibrium, incumbents provide different levels of public services, face different probabilities of re-election, and obtain different rents. The adoption of different forms of project governance permits incumbents to disguise themselves and undermine voters' ability to assess their performances. Therefore, yardstick competition is hindered, even if jurisdictions display identical revenue capacities.

The paper analyzes the design of simplified small business tax regimes in Eastern Europe and Central Asia and the impact of such regimes on small business tax compliance. Although many approaches for tax simplification exist, a general trend in the region is to offer small businesses the option to be taxed based on their turnover instead of net income. The study finds that many of the regimes in place are overly simplistic and neither take into account fairness considerations nor do they facilitate business growth and migration into the standard tax regime. Although revenue generation is not a main objective of such regimes, low revenue performance and the risk of system abuse by larger businesses should be issues of concern. More attention should therefore be devoted to improving the design of simplified regimes and monitoring their application. This will require in particular a more profound analysis of the economic situation and the tax compliance challenges in the small business segment and increased efforts to improve the quality of bookkeeping.

Introduction
Business taxation is at the core of the relationship between the state and its economic constituents. The transition towards market principles in the ECA region thus required a fundamental change in the principles underlying public revenue collection: A move away from the reliance on transfers, typically predetermined, by State Owned Enterprises (SOE) towards the assessment of actual taxable income for a growing number of private enterprises. As part of this major change in revenue administration and wider privatization and deregulation efforts, many transition countries established special programs to administer and support the growth and competitiveness of micro, small, and medium enterprises (MSMEs).
The development of small business activity during the transition resulted in major administrative challenges and a range of policy experiments to address these; not least regarding their tax treatment. Facing an ever-growing number of business "clients", the introduction of various simplified taxation schemes was partly an attempt to alleviate the administrative workload by tax policy makers. In practice, however, audit coverage remained relatively intense, given the common starting point of regular audits of all businesses before the transition (Engelschalk 2005). At the same time, starting with de Soto's influential work (1989), extensive informality of small firms and individual entrepreneurs had gained increased attention as a challenge for transition economies (Enste and Schneider 2002) with simplified tax policy and its administrative requirements commonly seen as one of the main policy levers.
In light of these ambitious objectives, the experience over the last 20 years casts some doubt on the effectiveness of simplified and preferential tax treatment in reducing compliance costs and burden to tax administrations, as well as in improving formal business creation and small enterprise growth. In some instances, widespread non-compliance and underreporting linked to simplified taxation may have turned into a major constraint for investment activities, as unfair competition from businesses that avail themselves of tax avoidance schemes increases economic pressure on formal businesses in the standard tax regime and reduces their competitiveness (WB 2012).
Research on a number of tax challenges for transition economies, in particular regarding the use of presumptive taxation regimes and the control of rampant corruption, has been scarce (Holmes 2002) aside from general guidance on MSME taxation (ITD 2006, IFC 2007 and select efforts to summarize country practices (Engelschalk 2005). Little evidence is on offer for policy makers interested in how to proceed in these areas of second-or rather third-best policy and administrative solutions, which are prevalent, given persisting capacity and resource constraints of both taxpayers and tax administrations.
This paper aims to contribute toward filling this gap based on documented country experience. We first provide an overview on general tax policy and administration trends (section 1), followed by a general discussion of country experiences in MSME tax policy in the region (section 2). The country practice informs a summary of lessons learned and policy recommendations, which are derived in section 3 of the paper.
The paper discusses exclusively the tax treatment of micro and small businesses. Most countries in the region have a general definition of what constitutes a micro and a small business either in a special SME law or in their commercial laws, and all countries define MSMEs for statistical purposes. These definitions generally refer to several parameters. In the Russian Federation, e.g., a business is considered to be small if annual turnover is not more than 11.2 million US$ and the number of employees does not exceed 100; 1 in Croatia the national accounting law sets small business thresholds of asset value below 4.8 million US$, annual revenues below 9.5 million US$, and average number of employees during the business year of not more than 50. Such definitions generally are not relevant for taxation purposes, however. Tax laws include special micro and small business thresholds, based on the revenue potential of the segment and its compliance capacity. These definitions generally are turnover-based and discussed more in detail later in the paper.

Changing Tax Revenue: Trends & Composition
Changes in the approach to business taxation across the ECA region are captured in a range of aggregate indicators. In 2012, central government tax revenue as a share of GDP ranged from 12.2 percent in the Slovak Republic to 24.1 percent in Georgia. While the ratio-a broad measure of the economy-wide tax burden-has dramatically increased in a range of economies, 2 it has remained relatively stable or declined in the majority of economies from 1995-2011.
More interesting than the aggregate numbers are the changes in the relative importance of direct and indirect tax instruments. As discussed below, the move to flat tax rates-one of the most prominent reforms in the region-had major implications for revenue collection. Its revenue impact varied depending on the degree and effectiveness of accompanying measures to increase the tax base, economic growth at the time of the reform, and complementing enhancements of tax administration (WB 2007). Generally, flat tax regimes, which tended to provide important alleviation of the tax burden in the upper income brackets, have reduced PIT revenues-Latvia, Lithuania, and Russia being an exception-and triggered heavier reliance on indirect instruments such as VAT and excises (Keen, Kim, and Varsano 2008).
More broadly, the spread of the VAT since the mid-1990s contributed to a more important role for indirect taxation in the revenue mix. Introducing a modern VAT has been a key component of tax reforms during the transition towards market economies in ECA, marking an important conceptual departure from previously levied product taxes. The general recommendations (not always followed, however) in designing VAT systems in transition economies included in particular (i) a simple (dual) 6 rate structure to facilitate administration and compliance, (ii) a comparatively high mandatory registration threshold to exempt the majority of small traders, while expecting an automatic but gradual increase in the tax base, as inflation would erode the real level of the exemption over the time needed to strengthen administrative capacity (Cnossen 1992, 232).
In many ECA economies, particularly in Central Asia, VAT revenues remain, however, largely based on VAT collected on imports, with domestic VAT contributing a small percentage to total revenues. The share of imports is thus an important driver of the productivity 3 of VAT. Source: IMF and WBG country reports.

Flat Income Taxation and Simplified Small Business Regimes
The ECA region is not only a region with widespread use of presumptive tax regimes for small businesses; it is also a region in which many countries have introduced flat income tax regimes, an approach pioneered by Estonia in 1994. Following Estonia and its Baltic neighbors, the Russian flat tax reform in 2001 attracted global attention due to subsequent improvements in revenue collection. This triggered a wave of similar reform efforts throughout the region (see Table 3 below). The general objective was to promote economic growth through creation of a business-and investment-friendly environment for individuals and companies, as well as to achieve a high degree of tax fairness, 4 to simplify administration and compliance, and to introduce greater tax transparency.
The move towards flat income taxation typically affected a broad range of related taxation areas, 5 though the details of the reform programs differed as summarized by Keen, Kim, and Varsano (2008).
In the majority of countries, the introduction of a flat tax regime was not connected to the operation of presumptive small business tax regimes and had no impact on presumptive regime design and 7 operation. Remarkable exceptions are the Slovak Republic and Georgia. In both cases, the introduction of a flat income tax was combined with a broader simplification of the tax regime. In the Slovak Republic, the 2004 tax reform process aimed at eliminating a large number of exemptions and special regimes to introduce a consistent and comprehensive approach to direct taxation. As part of this process, the small business presumptive regime was replaced by a standard cost deduction ratio for the self-employed. This change of the small business taxation approach seems to have had a positive impact on voluntary tax compliance; consequently, the number of income tax returns submitted by self-employed increased by 14.6 percent in the first year of flat tax implementation. 6 Year over year change in percent 2.3 -2.2 3.7 14.6 Source: Saavedra (2007) A similar approach was taken in Georgia, where the move to a flat income tax was combined with a comprehensive and successful tax simplification approach. The 2005 tax reform reduced the number of taxes from 22 to 7; the number of required visits to the tax office dropped sharply and the estimated tax compliance rate increased from 35 percent to around 80 percent. Introduction of the flat tax was seen as the appropriate occasion to abolish the dysfunctional patent regime in place since 1998. Different from the Slovak Republic, no further simplification measures were foreseen for MSMEs, which were expected to comply with the standard income tax regime. Following the farreaching simplification of the general tax system, small businesses, however, still experienced an increase in compliance requirements. While taxpayer perceptions improved dramatically among large businesses, an increasing share of small businesses identified tax administration as a key barrier to doing business following the reform. This experience is one of the factors that explain the decision to reintroduce a presumptive tax regime in 2010. Another special case is Estonia, which introduced a simple flat tax regime early in the transition process, before a separate presumptive small business tax regime had been developed. The Estonian regime does not include any special rules or simplifications for small businesses. In an environment with a relatively highly educated and IT-literate small business community (more than 97 percent of corporate tax returns and 93 percent of PIT returns are filed electronically) and the non-existence of unofficial costs related to taxation (Dickinson 2012), a simple cash-based general taxation system proves sufficient to support small business tax compliance. The Georgian example demonstrates that even a successful flat tax introduction combined with a comprehensive tax system simplification does not automatically guarantee that special simplification rules for small businesses are no longer needed. In particular, the requirement to calculate and document business expenses and the risk of having disputes about the deductibility of such expenses can be considered an additional burden of flat tax regimes compared to presumptive tax regimes. Experience in some ECA countries, such as Bulgaria, has shown that the move to a flat tax did not substantially reduce the complexity of filing and documentation requirements, making compliance with the flat tax regime still burdensome for small business operators. This is even more of a challenge when the overall tax simplification measures combined with the flat tax introduction do not go far enough. The flat tax introduction in the Russian Federation, for example, was part of an exercise to introduce a new tax code, with the first part of the code becoming effective in 1999 and fundamentally reforming the system of tax administration, while the second part, dealing with specific taxes, was approved in 2000 and became effective from 2001. The centerpiece of the Russian reform-a single marginal personal income tax rate of 13 percent-was followed by an impressive increase in real personal income tax revenues of about 26 percent in the first year after its introduction. Using micro-level data, Ivanova, Keen, and Klemm (2005), however, provide cautionary insights, suggesting that attribution of the revenue performance to the PIT reform alone is questionable. 8 Moreover, the new Russian tax regime still consisted of around 40 different taxes, and small businesses remained confronted with an average of 9.56 types of taxes (Shetinin et al, 2005). Despite the flat tax introduction, the move to a more simplified tax regime for small businesses thus remained a valid concern, which was addressed with the introduction of the simplified tax system (STS) in 2003. A similar development occurred in Ukraine.
Pressure to introduce or maintain presumptive tax regimes with a low effective tax burden can also build up in the case of a flat tax regime which uses a high tax rate. Flat tax reforms in the ECA region did not necessarily lower the average and marginal tax rates for small businesses. Some people saw no change in their marginal tax rates, since many governments selected the marginal rate in one of the tax brackets that was previously used. This happened, for instance, in Lithuania, Latvia, and Georgia. 9 The rates selected in the early flat tax reforms in the Baltics were either corresponding to the highest marginal rate before the reform (Lithuania, Latvia), or in the middle of pre-reform rates (Estonia). With rates that remained high in comparison to those used in the preceding income tax 10 regime's brackets (Easterbrook, 2008), small business lobby groups continued to have grounds for requesting preferential tax treatment.

Administrative Approaches and Private Sector Perceptions
As Bird and Vazquez-Caro (2011) highlight, the effort of benchmarking tax administration performance is fraught with challenges, and the information provided in such exercises may not necessarily be of high relevance for policy makers aiming to improve their administration. A typical measure that is considered is cost of revenue collected, that is, the cost of administering the tax system compared to the total revenue collected by the tax administration. The information is not available for all countries in the region, but for those where information is provided, important differences can be observed. Estimates range from a cost share of less than 0.5 percent of total revenue in Estonia to more than 2 percent in Poland and the Slovak Republic. 10 Cross-country benchmarks are similarly challenging when comparing trends in the transparency, attractiveness, and integrity of the tax regime across countries. While a variety of benchmarking exercises have gained popularity among tax policy makers, their limitations are numerous. 11 Nevertheless, private sector perceptions, the legal and regulatory framework, and the structure of managerial system of the tax administration can help point towards general areas of interest.
The Paying Taxes report, prepared by the WBG and PwC as part of the Doing Business benchmarking exercise, captures information across three dimensions of the business taxation regime. 12 An analysis of the region-wide trends since 2004 suggests that Central Asian and European economies have been the most active in pursuing business taxation reform to lower administrative and financial burdens of businesses when compared to other regions (PwC 2014).
In line with these findings on the statutory regulatory environment, enterprise surveys, which were repeatedly conducted throughout the region to capture business opinions, showed an overall improvement in the perceived burden of the tax regimes from 2009-13 in the region.

II. Regional Overview: Development and Issues Regarding the Tax Treatment of MSMEs a) Income/Profit Taxes and Compliance Costs
In the ECA region, presumptive tax instruments were typically introduced in the late 1990s or early 2000s, with the objective of promoting private sector development and facilitating compliance management in an environment characterized by low tax administration capacity and a rapidly growing number of private small business operators. Prioritizing the effective compliance management of larger businesses to ensure sufficient revenue mobilization required tools to minimize administrative efforts for smaller entities. The goal was to encourage voluntary compliance of small businesses while allowing for a simple examination of low-revenue tax returns.
Moreover, a number of compliance burden and compliance cost studies highlighted the regressive features of tax compliance costs and stressed the need for developing simplified systems of taxation. Klun and Basic (2005) provide estimates for Slovenia and Croatia, and, similarly, survey-based analysis by the World Bank in Ukraine (2009), Uzbekistan (2008), Armenia (2010), and Georgia (2011) supports earlier findings of compliance cost assessment in the OECD, depicting a highly regressive burden. Given the high fixed cost component of tax compliance, the general trend identified in these surveys is hardly surprising: the smaller the businesses, the higher the tax compliance cost they face as a share of their turnover.
It is notable that even for businesses operating at more than $100,000 in turnover, measured compliance costs can still surpass 3% of their turnover level. The reasons for such high compliance burdens vary and include complicated reporting procedures and time spent on inspection visits and/or audits, a frequent challenge for MSMEs in the region (Engelschalk and Loeprick, 2011). Kireeva and Rudy highlight, for example, that in Belarus, to comply with the general tax regime, a small business with up to 50 employees on its books has to employ an average of two accountants. They estimate that monthly costs incurred by the SME segment for tax compliance may exceed $3.5 million. 13 13 National Report Belarus, in Lang et al (2008)

Figure 3: Regressive Tax Compliance Costs for Micro and Small Businesses in the Region
Source: IFC Tax Compliance Cost Surveys 2007-11.
Unlike other regions, presumptive regimes in the ECA region also frequently include incorporated businesses. In Africa or Latin America, the majority of countries restrict such regimes to nonincorporated businesses. A number of justifications can be made for the extension of presumptive regimes to the corporate sector. Compliance costs can be burdensome irrespective of legal status, and participation in the presumptive regime can be a tool to reduce these and thereby increase the competitiveness of small corporations. Moreover, small corporations do not necessarily have better in-house accounting capacity than non-incorporated businesses. On the other hand, however, different accounting obligations may already require more comprehensive records for corporations, and the risks of system abuse and downward migration increase markedly when corporations are eligible for simplified income tax treatment.
Attempts made in a number of countries to abolish presumptive tax regimes for small businesses have generally not been sustainable: -In Georgia, a patent system for small businesses was in place until 2005, when a new tax code was introduced. However, taxation of small businesses based on net income resulted in Tax Compliance Cost as a percentage of turnover

Ukraine Uzbekistan
Armenia Georgia high compliance costs, and in 2011, Georgia adopted a new simplified tax regime. Micro businesses with a turnover below GEL 30,000 ($18,100) are exempt from income taxation, while small businesses with a turnover below GEL 100,000 pay a presumptive tax based on turnover.
-Romania operated a Micro-Enterprise Tax (MET) regime with a 3 percent rate on turnover until January 2010, when the regime was abolished and small taxpayers were moved to the general tax regime. A year later, the system was reintroduced, and starting from January 2013, presumptive taxation even became mandatory for incorporated small businesses. The system was used by 92,000 taxpayers prior to its abolition, representing about 20 percent of eligible small businesses, and around 60,000 of these businesses immediately moved back to presumptive taxation after its reintroduction.
-Armenia had a turnover tax for businesses with turnover below AMD 30 million ($71,700). The turnover tax was abolished in 2008 as part of a major tax reform exercise, which also increased the VAT threshold to AMD 58.25 million ($180,000). Only the patent regime for micro businesses and a presumptive small business tax for a few selected activities, such as barber shops, remained in place. The repeal of the turnover tax regime was partly motivated by widespread abuse of the regime by larger businesses. However, it resulted in a significant additional compliance burden for many small businesses, and therefore presumptive taxation based on turnover was reintroduced in 2013.

b) Basic System Design
Most of the regimes targeting micro-, small-, and medium-size businesses in the region have changed fundamentally since their first-time introduction, and many regimes get modified on a regular basis. While in the 1990s, simple fixed tax or patent regimes were also widely used for the small business segment, today, such regimes are largely limited to micro businesses, and a turnover-based approach has become the standard method for taxing small businesses. Experimentation with the tax treatment has been common in many countries. Kazakhstan, for example, first introduced simplified taxation for micro and small businesses in 1995. The system has been modified several times since, including a move from a fixed tax to a turnover-based tax and the replacement of a progressive presumptive tax by a single rate on turnover. In addition to the move towards a turnover-based calculation of the presumptive tax liability, it was increasingly recognized that the MSE segment of the taxpayer population actually consists of two different taxpayer groups: micro and small businesses. Further segmentation to differentiate between the small and the micro business categories constitutes the second major system reform trend. Frequently, segmentation was combined with an effort to improve local revenue mobilization, and the micro business tax revenues were allocated to local budgets.

c) Small Businesses and the VAT System
In theory, strong arguments can be made in favor of including small businesses in the VAT net. As a tax on consumption, the VAT chain would ideally stretch from the point of production (or import) to the point of sale to the final consumer, thereby including the retail sector or the provision of services to private consumers. Indeed, when VAT was introduced in the ECA region in the 1990s, a number of VAT laws included no or a very low VAT threshold. For example, the Russian VAT started with a very low threshold of Rub 100,000 (Russian Rubles) in 1992 (quickly increased to Rub 500,000), Hungary operated a threshold of $9,000, and Romania operated a threshold of $6,000 in the early 1990s (Jack 1996). The VAT systems in Belarus and Uzbekistan still operate without a threshold for incorporated businesses. From the business perspective, an advantage of being part of the VAT net could be for small businesses to facilitate interaction with VAT-registered businesses, assuming that a VATregistered larger business prefers ordering goods and services from clients who can issue VAT invoices. This might not always be the case, however; not all VAT-registered businesses are equally interested in formally deducting input VAT, and lowering the sales price of goods or services rendered could be an approach preferred to issuing a VAT invoice. In addition, as highlighted by Bird and Gendron (2007), in some countries, vibrant markets have been established by the trading of VAT invoices. Nevertheless, good practice suggests providing small businesses with an option to voluntarily register for VAT, even if their turnover is below the registration threshold, in case the business can prove that it is capable and willing to comply with VAT requirements (in particular keep the necessary books and records, issue VAT invoices, and operate cash registers).
In practice, there are several strong arguments against an approach of extending the VAT net to the small business segment. IMF experience has shown that setting too low of a threshold for VAT can significantly compromise the political and administrative feasibility of a VAT (Ebrill et al, 2001). Additionally, WBG/IFC tax compliance cost surveys show that joining the VAT regime substantially increases the tax compliance costs for small businesses. Feedback received from small businesses in European Union member countries demonstrates that small businesses consider VAT legislation as one of the 10 most burdensome EU laws. The compliance challenge is thus increased when VAT systems require extensive documentation, where taxpayers are subject to frequent VAT audits, or when filing procedures have not been streamlined and simplified. This is the case in a number of ECA countries, as comparative analysis indicates that the time required for VAT compliance in the region substantially exceeds compliance time in EU countries.

Figure 4: VAT Compliance Time across Regions
Source: PwC, The Impact of VAT Compliance on Business 2010.
A more detailed compliance cost analysis for VAT was conducted in Slovenia in 2001, two years after VAT was introduced with a relatively low threshold of SIT 5 million ($22,700), demonstrating the high compliance burden for small businesses. SIT 100 million-1 billion 0.73% Above SIT 1 billion 0.08% Source: Klun, 2003.
The risk to small business in complying with VAT requirements increases further in an environment of low administrative efficiency. Non-payment of VAT refunds or delays in the processing of refund requests can severely affect the liquidity of the business. Liquidity problems can also arise in case of

Hours Required to Comply with VAT
an accrual-based VAT system, when the VAT due has to be transferred to the Treasury before the business receives payment for goods or services from its customers. At the same time, the revenue benefits of including small businesses in the VAT net are minimal, as their contribution to total VAT collection is generally below 10 percent.
A reasonably high VAT registration threshold is the main tool for protecting small businesses from problems and costs related to VAT compliance. Country practice in the region varies considerably here, and a considerable number of countries apply a rather low threshold of less than $50,000 turnover, which also risks forcing many small businesses to join the VAT net.

Figure 5: Mandatory Registration Thresholds for VAT in ECA (in USD)
Source: Collecting Taxes 2012-13.
A number of FSU countries (Russia, Ukraine, and Belarus) have taken a different approach by integrating the VAT liability in the presumptive single tax regime. In these regimes, part of the single tax payment is considered as covering the VAT liability of the business. However, the benefits of including VAT in the single tax are questionable. From a tax administration point of view, the VAT net is not really broadened, and additional data to check VAT compliance of larger businesses is not generated. For businesses, the tax burden is increased with the application of a higher presumptive turnover tax rate, but the business cannot deduct input VAT or issue VAT invoices. Consequently, the competitive position of single taxpayers supplying VAT-registered businesses does not improve. Therefore, while a single tax approach may have many benefits in general, the inclusion of VAT in the scope of the tax does not generally seem sensible. In practice, a benefit of a VAT-inclusive single tax could be to protect small businesses from the administrative burden of asserting the exemption from VAT (see Box 1 below). The more appropriate course of action to take in this case would be to adjust the rules for small business VAT exemption.

Practical application of the VAT threshold in the Russian Federation
Russia operates a comparatively high VAT threshold. Businesses are not obliged to comply with the VAT regime if their aggregate turnover for three consecutive months is below Rub 2 million ($55,400). However, in order to benefit from this threshold, the business has to apply for an exemption from VAT and submit documentation required for turnover verification. The exemption is granted for a period of 12 months, after which it has to be renewed. Tax offices may refuse the exemption if they are not satisfied with the application. Should the turnover of the business exceed the threshold during the exemption period, the business has to comply retroactively with VAT obligations from the beginning of the month in which the threshold was exceeded.
While the introduction of a reasonably high threshold is a convenient solution for protecting small businesses from a high VAT compliance burden, the challenge to reduce obstacles for small business growth and to facilitate the transition into the VAT system remains, irrespective of the threshold level. Small businesses are likely reluctant to migrate into the VAT regime even in the case of a reasonably high VAT threshold. In Uzbekistan, for example, 58 percent of taxpayers who registered for VAT saw no advantage to their business in being a VAT taxpayer, and business operators cited many disadvantages related to joining the VAT net. Only 14 percent of businesses saw no disadvantages to VAT registration.

Figure 6: Business Perceptions on Disadvantages of being VAT Registered in Uzbekistan
Source: IFC, Tax compliance and reporting costs for businesses in Uzbekistan 2010. Facilitating compliance for smaller businesses in the VAT net should thus be an important part of a small business growth facilitation strategy. For many CEE countries, the EU VAT rules provide an orientation for such VAT simplification. Article 281 of the EU VAT Directive allows EU member countries to apply simplified procedures, such as flat rate schemes, for charging and collecting VAT from smaller VAT-registered businesses. The most widespread simplification measure in CEE countries is the use of VAT cash accounting schemes. The schemes allow the business to postpone the VAT payment to the date it receives payment for goods supplied and services rendered. Such schemes are in place for businesses with annual turnover below the following thresholds for these countries: Estonia at €208,646 ($236,000), Slovenia at €400,000 ($453,000), Latvia (for small businesses) at €100,000 ($114,000), Romania at RON 2,250,000 ($682,600), Serbia (since 2013) at SRD 50 million ($598,000), Bulgaria (since 2014) at €500,000 ($566,000).
Another measure with major impact on compliance and administrative costs is the reduction in the VAT filing frequency. Analysis in the EU suggests that for a micro business, the costs of filing a monthly VAT return amounts to more than €100 (around $140) per return. 15 In Hungary submitting three monthly returns instead of one quarterly return, increases filing costs by 35 percent. 16 A move from monthly to quarterly filing would therefore reduce annual filing costs of a micro business by $430. Quarterly filing for small businesses has now become a widespread compliance facilitation approach, although some countries, such as Bulgaria and Estonia, require all VAT-registered businesses to file a monthly VAT return. The EU definition of a micro business refers to businesses with fewer than ten employees and an annual turnover and/or annual balance sheet of not more than €2 million. While in many countries a reluctance of small business operators to join the VAT system can be observed, in some situations the opposite phenomenon may occur. In Romania, analysis conducted in 2010 showed that almost 380,000 small businesses with a turnover below the VAT registration threshold were voluntarily VAT registered. This meant that more than 60 percent of the VAT net comprised of small businesses, which according to system design, should have remained outside the VAT, complicating VAT administration for the tax offices and contributing only 1.3 percent to total VAT revenues. Closer analysis is required in such a situation in order to understand the dynamics that force small businesses into the VAT and increase their compliance costs. A more drastic approach is to exclude micro businesses from voluntary VAT registration. This approach was tried for some time in Serbia with the operation of a threshold for mandatory VAT registration at the level of SRD 4 million ($47,800) and a threshold for voluntary VAT registration of SRD 2 million ($23,900). The reform of the VAT law in 2012 abolished the threshold for voluntary registration, as it resulted in an obligation for registered businesses to deregister when the business turnover dropped below the SRD 2 million threshold. At the same time, the threshold for mandatory registration was increased to SRD 8 million ($95,600).

d) Small Businesses and Social Security Systems
Social taxes constitute an important component of the tax system in many ECA countries, and the need for compliance can create an obstacle for small business operators to formalize and legalize labor. For the self-employed, a number of ECA countries offer a preferential social tax treatment as an incentive for voluntary compliance and compensation for higher compliance costs. As Leibfritz (2011) 19 points out, such an approach creates a distortionary element in the tax regime and encourages employees to change their status from dependent employment to self-employment. Even without the explicit objective to reduce the social tax burden, tax policy makers face difficulties in applying the general social contribution regime to self-employed operating in a presumptive tax regime. While the level of social tax payment is generally a fraction of the net income of the selfemployed, a presumptive tax regime does not produce any information on the net business income, and the requirement to calculate net income just for social tax purposes would conflict with the simplification objective of the presumptive regime. An alternative calculation method in this situation is to apply the minimum wage as a tax base for the social tax, irrespective of the actual income of the self-employed. This was the case in Hungary under the EVA system and is applied for turnover tax payers in Poland. The result of this approach is that self-employed tend to have a much lower tax wedge than regular employees, and the incentives for salaried employees to become self-employed contractors (at least on paper) increases. 20 Such a trend can be observed in Ukraine, where it is assumed that the remarkable increase in the number of unincorporated small businesses is largely caused by salaried employees who register as small entrepreneurs and pretend to operate as independent contractors in order to secure the benefits of the simplified tax system (STS). Comparing small business development in Ukraine and in the Russian Federation, an OECD analysis finds that the 23 average unincorporated small business in Ukraine employs only approximately 1.6 persons. This figure has been falling, slowly but steadily, over the last six years; in Russia, by contrast, the average urban unincorporated small business in 2004 employed approximately 4.5 persons and the average number of employees rose. The fact that so many Ukrainians registered as individual entrepreneurs who appear to work entirely alone, reinforces the perception that many are not actually selfemployed entrepreneurs at all, but are seeking to exploit the benefits of the STS. A similar phenomenon may be observed in the small companies sector: the average number of employees in small companies fell from around 8 to 6.4 persons between 2000 and 2006. 21 A parallel but somewhat different issue is the social tax treatment of salaried employees in small businesses. Given the fact that social benefits like health insurance, unemployment insurance, and pensions are generally linked to the duration and level of contributions provided to insurance and pension schemes, the only transparent and reliable compliance method is for small business employers to calculate and transfer the precise employer contributions to these schemes and deduct the employee contribution from salaries paid. There is thus no difference in the approach between small and larger business employers. For obvious reasons, this approach is not favored by small business operators using presumptive tax regimes to comply with their business income tax obligations.
Turnover tax regimes are biased against taking on formal sector employees, because-unlike the standard tax regime-the costs of hiring labor cannot be deducted as expenses, and this effectively increases the small business tax burden. This bias is heightened by a social tax compliance burden which may equal, or even exceed, the presumptive tax compliance burden. While simple turnoverbased presumptive tax regimes are expected to promote business formalization and migration out of the shadow economy for the actual business entity, they simultaneously create incentives for labor to move into the shadow economy.
Some ECA countries have tried to find ways to mitigate this risk and reduce the social tax compliance burden. A typical approach is to integrate social tax into the presumptive tax regime. In Latvia, the micro business turnover tax of 9 percent exempts the business from withholding employee personal income tax (general PIT rate of 25 percent) and includes employer as well as employee mandatory contributions to the social security system (employer share of 24.09 percent; employee share of 11 percent of income). The system works in its basic design for businesses of up to five employees. In case of larger staffing numbers, an additional 2 percent on turnover is charged for each additional employee. Also, in case the monthly salary of an employee exceeds €700 ($792), the exceeding amount is taxed at a rate of 20 percent. In Ukraine, a business paying the unified tax of 6 percent on turnover does not have to comply with income tax, social security, property tax, and some local tax payment obligations. Unified tax revenues are allocated based on a fixed ratio: in case of legal entities, 42 percent of unified tax revenues go to the State Pension Fund, 15 percent to the State Social Security Fund, 23 percent to local governments, and 20 percent to the central government (in case of non-incorporated businesses, 43 percent of revenues are allocated to local budgets, while the central government does not benefit from unified tax collection).

24
The major difficulty with this approach is the lack of clear attribution of the social tax payment to the beneficiaries of social benefits. In particular, when the transfers to the social agencies do not depend on the number of employees hired, a relationship between benefits and payments cannot be established. It may also be difficult for individual employees to prove that they have acquired social benefits during the times worked for employers being presumptive taxpayers. There is thus no real alternative to imposing regular compliance with the social contributions system on small businesses in the presumptive tax regime. A certain incentive for remaining in the formal labor system can be provided, as in the case of the Russian simplified tax system, by allowing a deduction of payments made to social security agencies from taxable turnover.

Defining the Segment
Recognizing the fact that both the growth potential and the compliance capacity of micro businesses is substantially lower than in the case of small businesses, many ECA countries introduced specific tax instruments for the micro business segment. Thresholds defining the micro segment vary considerably across the region. The table shows that not all systems have managed to limit the application of micro business regimes to very small entities operating around subsistence levels. The analysis of the experience in the region shows that one risk of micro business segmentation is that the very simple (and often very preferential) micro regimes also become an attractive model for businesses above the size of a micro business, and pressure to extend the regime up to the VAT threshold level is building up. The Serbian experience provides a good illustration of this dynamic.

Access to the micro business patent regime in Serbia
The patent regime in Serbia was initially targeted at micro operations at a level below SRD 2 million turnover ($23,900). However, the threshold was soon increased to SRD 3 million, and in 2013, the system was extended further to businesses with a turnover of up to SRD 6 million ($71,700). The system in principle is targeted at "Any sole proprietor, who in view of circumstances, is unable to keep books" (Article 40 of the Income Tax Law); however, it is also accessible to the well-educated and self-employed. The 2013 system reform at least managed to deny patent regime access to accountants, auditors, tax advisors, and marketing agencies. The initiative taken by the MoF to also exclude doctors and lawyers was not well received by lawmakers, however, and the parliament voted against the initiative.
On the other hand, some countries, have demonstrated that despite such pressure, a better alignment of the system threshold with the concept of targeting micro businesses can be feasible. An example is Tajikistan, which has limited the application of its micro regime from a previous turnover of $41,600 to $20,800 in 2014.

The Micro Business Tax Regime
The standard approach to micro business taxation in the region is the application of a patent regime. Frequently, these regimes are administered by local governments, and revenues go to local budgets.
Only in few cases, such as in Georgia, are micro businesses exempt from income taxation. Patent fees are generally determined by local governments, while national tax laws provide the basic structure of the regime and determine minimum and maximum patent rates.
The actual design of patent regimes varies considerably in practice. In a few countries, a very basic patent regime has been put in place with a fixed amount for all micro businesses, irrespective of business type and location. Much more frequent is the development of a detailed list of micro business activities and the determination of individual patent fees for each activity. A typical example of such an approach is the micro business regime in Bulgaria.
Bulgaria has experimented with its patent system for quite some time, both at the national and local level. While patents were initially administered by the national tax administration, the patent regime was transferred to local governments in 2008. The Law on Local Taxes and Fees lists 40 different categories of activities and specifies a minimum and maximum amount per indicator established. Local governments then determine the applicable amount for businesses on their territory, taking into consideration the precise location of the business, the economic importance, population density in the area where the business has been established, and the seasonal or permanent nature of the activity carried out. Hotels (  Distinguishing activities of similar nature: Many detailed patent lists could be simplified. In particular, the same patent rates are set for a variety of activities of similar nature, and thus unnecessarily complicate the list. The patent list for Bishkek, for example, includes eight different activities in the area of clothes manufacturing (activity list no. 77-84); however, the patent rate for all eight activities is exactly the same. Merging similar activities into a single category is an easy option to simplify convoluted systems.
 Mixed activities: The more detailed the activity list, the higher the probability that a micro business carries out more than just one listed activity. The Bishkek example given above, includes separate patents for the manufacturing of working clothes, other clothing and accessories, and underwear. In cases like this, micro businesses which are not highly specialized face the risk of multiplying their tax liability without necessarily achieving a higher profit than businesses that are active in only one area. There is therefore no clear justification for creating such a risk.
 The gap issue: Even a very extensive activity list can never be comprehensive. In fact, the more detail-oriented the list, the higher the probability that some micro business activities are missing. For affected business operators, it is then unclear how their patent rate will be determined or if they can access the regime at all.
 Determination and update of patent rates: It is a major exercise to set more than 100 different rates and ensure that they are properly updated and reflect changes in the business environment, profitability, and costs. The MoF and the tax administrations are typically not well equipped, and they do not have the required resources to perform this task properly. Sufficient data and information to allow a thorough calculation of appropriate rates for very specific activities is rarely available.
 Acceptance and fairness: As a result of the rate-setting problems, international experience shows that the acceptance level of detailed patent regimes in the micro business community is low. Business operators tend to consider the rate-setting process to be non-transparent. They also perceive the patent rate that they have to pay as arbitrary and unfair.
It is therefore desirable to streamline the detailed patent lists and categorize micro business activities more broadly. A reform in this direction has been implemented in Tajikistan.

Improving the Operation of a Patent Regime: Micro Business Tax Reform in Tajikistan
Tajikistan operates a patent system for individual entrepreneurs with a turnover of up to SM 100,000 (Tajik somoni) ($20,800) annual turnover and size of business premises up to 30 square meters. The cost of the patent is fixed according to type of activity, and varies from SM 240 ($50) per month for transportation of oil, liquid gas, and cement by specialized transport, to SM 30 ($6) for transportation by motor-scooter. The patent was initially only valid for one narrowly specified category of activities and it had to be renewed annually. A reform of the patent system in 2008 introduced major simplifications: -The new patent system offers an expanded scope of permitted economic activities. It reduced the number of patent categories from 49 patents (with 70 sub-activities) to 28 patents (covering 169 subactivities); -Tax inspection of patent holders is limited to verification that the individual entrepreneur has a patent, that the business activity is allowed per the patent, and that the patent-holder has a receipt of payment for the patent. Previously, the patent-holder was required to keep a cash register and maintain transaction logs, based upon which the tax inspector would determine tax obligations; -The patent does not have to be updated annually any more.
The patent system has become much more attractive after this reform, and 15 percent more patents were issued in 2008 compared to 2007.

Use of Additional Indicators
While a general patent regime establishes the same tax liability for all micro businesses active in a specific field, such an approach might be considered inappropriate, as the prospect of earning income depends critically on certain criteria, such as the business location or facilities. Some patent regimes have therefore introduced a small number of critical indicators to take account of the specific business situation. By far the most popular and, at the same time, most transparent indicators are the number of employees and the location of business premises. A system combining the use of these indicators is the Tax Card regime in Poland. The use of other indicators tends to be much more problematic. This is the case particularly where turnover elements have been introduced to create a progressive patent regime. In Croatia, for example, the lump sum tax amount depends on the level of the micro business turnover. The micro business in this case is still required to maintain a cash book and calculate its annual turnover. With regard to the calculation of the tax liability, a normal turnover tax could obviously be applied in this case and would eliminate the jump in the tax burden arising from moving from a higher turnover band.

Patent System Stability and Predictability
While patent systems have the potential to provide a transparent approach to micro business taxation and reduce micro business compliance costs, the lack of predictability of tax rates can be a major issue in practice, in particular when patent rates are set by local governments and are changed based on the revenue needs of the local government. In Lithuania, for example, a country where the patent regime extends to small businesses up to the VAT registration threshold, patent rates are determined annually by local councils, and the patent levels change substantially. While being an important revenue source for local governments, the unpredictability of the regime may add further disincentives to business formalization.

Figure 7: Volatility of patent rates in Lithuania
Source: Vilnius City Council data

Abuse Risks of Micro Regimes
For small businesses in the higher turnover range, patent-based taxation may be an easy and attractive way to reduce the overall tax liability. In principle, patent regimes may therefore create a risk of downward migration. In practice, such a risk can be largely mitigated by (i) an appropriate limitation of the application of the micro regime, and (ii) a design of the small business regime taking into account small business profitability and ability to pay.
In the case of Georgia, which exempted micro businesses from taxation, initial assessments of taxpayer behavior following the reform suggest that a noticeable downward migration of small businesses did not occur (Bruhn and Loeprick 2014). Downward migration is thus not a given, provided that the micro threshold is sufficiently low, while effective compliance management covers the segment, and the small business regime is operated reasonably well.
The system abuse risk increases with high patent system thresholds. In Kazakhstan, the upper threshold of the micro business (

Minimum Annual Price of a Business Certificate in Vilnius (in LTL)
31 extends to net incomes above the average annual wage of KZT 1.6 million 23 for businesses with profit margins above 40 percent. There is high popularity of the regime and ample anecdotal evidence that the high income, self-employed use the regime to lower their tax liability.

Figure 8: Number of Taxpayers Applying Simplified Tax Regimes in Kazakhstan
Source: Kazakh Revenue Service A patent regime may thus erode the small business tax regime and create major disincentives for small business growth when it reaches out to businesses up to the VAT registration threshold. This can be observed in Bulgaria, the Kyrgyz Republic, and in Lithuania. In the Kyrgyz Republic, small businesses have the choice to either apply the turnover-based small business tax regime (simplified regime) or the micro business patent regime, with the patent regime generally offering a comparatively favorable tax treatment, as demonstrated in Table 13. As a result, acceptance of the simplified regime is minimal and most small businesses request patents. With the patent regime not requiring any books and records, business growth and migration into the VAT, as well as the standard income tax regime becomes less likely, while the tax administration faces difficulties in effectively controlling for abuse of the regime.

Figure 9: Use of Tax Regimes in Kyrgyz Republic
Source: IFC: SME Survey 2009.

Defining a Small Business System Threshold
International good practice recommends aligning the upper threshold of the simplified small business tax regime with the VAT registration threshold. Businesses registered for VAT need to comply with advanced accounting standards, and should therefore also be in a position to easily calculate their net income for income tax purposes. Additionally, VAT compliance management benefits from the possibility of cross-checking information on income tax and VAT returns. Using the VAT threshold as a criteria to determine the ceiling of the simplified small business tax regime is less evident however in cases where the VAT threshold is either very high-thereby limiting the application of VAT to large businesses-or very low. In the first scenario the simplified regime would extend to the medium business segment and thus to a group of businesses which are able -and should be obliged -to calculate their net income, while in the second scenario a low VAT threshold also impedes compliance cost reductions on the direct taxation side.
One notable feature of the presumptive tax design in the ECA region is the frequent lack of proper system coordination between the presumptive tax regime and the VAT regime.

Approaches to Turnover-based Small Business Taxation
Turnover has become the most widely used base for small business tax systems, and its design varies considerably from country to country. The two key system design alternatives are:  Presumptive income tax regimes replacing only PIT or CIT, versus single tax regimes offering small businesses the option to pay only one tax instead of a variety of taxes. 24  Single tax regimes with a single tax rate for all types of small businesses (such as in Azerbaijan with regional tax rate variations), versus multiple rate regimes using different tax rates for different business categories to account for different (assumed) average profit margins, particularly in the trade versus service segment.

III. Use of Presumptive Regimes: Lessons Learned a) Low Take-up and the Design of Appropriate Rate Structures
Two conclusions can be made regarding the use of presumptive tax regimes. First, experience shows that newly introduced presumptive regimes are not automatically widely applied by small businesses.
Regime take-up rates tend to be low initially, and only increase as a result of taxpayer information campaigns and trust being built in the small business community, with regard to the practical application and potential benefits of the regime.
Second, the use and popularity of presumptive regimes is not directly correlated with the turnover rate applied. Turnover tax regimes with a low tax rate do not automatically attract more businesses than regimes with a higher tax rate. Azerbaijan is an example of a country applying different turnover rates depending on the location of the business; while a 4 percent rate on turnover is applied for businesses located in Baku, businesses in the regions only have to pay a 2 percent turnover rate. Survey results suggest that this rate differentiation has no measurable impact on the level of system use.

Figure 10: Use of Tax Regimes in Azerbaijan
Source: IFC, Study of Small and Medium Enterprises in Azerbaijan, Baku 2009.
The level of the tax burden as a percentage of turnover may impact the composition of the presumptive taxpayer group. A system with higher turnover tax rates may become less attractive for businesses with low profit margins. This is an unwelcome result, as businesses with lower profit margins, which frequently operate in retail trade, generally have less capacity to comply with net income taxation than service businesses and professionals. The system thus risks being biased towards segments of the small business population which have less need for compliance simplification. An example of such a phenomenon is the micro-enterprise tax regime in Latvia, which has a comparatively high rate of 9 percent on turnover. In practice, this translated into a situation where the regime generated very little interest among small traders. Its primary users were selfemployed professionals and, in particular, small limited-liability companies.

SMEs and Individual Entrepreneurs: The Choice of Tax Regimes in Azerbaijan
Simplified tax regime Standard tax regime Exempt from all taxes Interestingly, the same phenomenon can be observed even in countries with a much lower turnover tax rate, as shown in the example of Tajikistan, which operates only a 4 percent turnover tax (5 percent in the case of turnover above SM 200,000). While the turnover tax regime is applied by 80 percent of small tourism operators, 75 percent of small consultancy service providers, 61 percent of medical service providers, and 67 percent of consumer service businesses, the take-up rate is only 39 percent for retail businesses and 35 percent for construction companies. A gradual change in the composition of businesses using the simplified tax regime has also been experienced in Hungary with the operation of the Simplified Entrepreneurial Tax (EVA) and the gradual increase of the EVA tax rate.
EVA was introduced in Hungary in 2003 as a flat tax, paid on sales revenues. Eligibility was limited to private entrepreneurs and business entities that had been in business for at least two years, had an annual income (including VAT) not exceeding Ft 25 million (approximately $115,000), and with individuals as owners. EVA payers were not subjected to income tax, dividend tax, company car tax, or value-added tax. The EVA rate was initially 1 percent, levied on turnover with some minor adjustments. EVA quickly became very popular, and the number of businesses applying for taxation under the simplified tax increased from 59,000 in 2003 to 83,000 in 2004, and close to 100,000 in 2006, when the tax rate was increased to 25 percent. Surprisingly, the participation level remained stable and still covered around 96,000 businesses in 2010. As the system was attractive primarily for businesses with margins exceeding 60 percent, it seemed to be used extensively by engineers, lawyers, and accountants, who operate with low cost/income ratios, implying that EVA reduced their tax burden. Some entrepreneurs used EVA even if it entailed a higher tax burden, because it reduced their administrative costs. 25 In 2010, the EVA rate was further increased to 30 percent, and to 37 percent in 2012 before being abolished in 2013.
The risk of factual discrimination of low-profit trading businesses can be avoided by introducing a rate differentiation in the turnover tax regime. Small business in the trade segment can be offered a lower turnover tax rate, while service businesses and in particular, self-employed, can be taxed at higher rates. While such an approach is still a rarity in the region, it has been tried successfully by some ECA countries. Poland, which has developed the most sophisticated rate differentiation, demonstrates that the benefits of the approach are twofold: it eliminates system entry barriers for businesses in retail and wholesale trade, and improves presumptive tax revenue generation, as higher turnover rates on high profit segments considerably increase the overall presumptive tax collection. Source: Grabowski, Polish presentation for the ITD regional conference on SME taxation, Tbilisi 2011.
The major drawback of applying the approach to practice is the lack of reliable data on average small business profit margins, which could justify the scope and level of rate differentiations. Data from statistical offices in ECA countries frequently proved insufficient as a basis for rate differentiation, and attempts made to collect information on average profit margins in various small business segments through targeted taxpayer surveys generally did not provide sufficiently reliable data. Consequently, such survey results are difficult to interpret and provide only a limited basis for policy decisions. An example of such a situation is the IFC SME survey conducted in Tajikistan in 2009: The survey shows major differences in profit margins within the service segment, while most activities listed generally only show a relatively small overall profit margin range between 10 percent and 15 percent. This does not explain the considerable differences in simplified system use discussed above, and probably indicates that real profit margins are substantially different from the margins provided for the survey.

Figure 13: Small Business Profit Margin Analysis in Tajikistan
Source: IFC SME survey 2009.
A reliable profit margin differentiation for tax purposes requires input from tax administrations, particularly through the collection and analysis of SME tax audit data, and ideally as a combination of targeted risk-based audits, as well as a number of random audits conducted to verify margin estimates. Such audit data, however, are not generally available, or at least have not yet been systematically analyzed.
With an incomplete and unreliable basis for a turnover rate differentiation, an approach similar to Poland, distinguishing five small business segments becomes questionable. More appropriate in this situation would be a simpler two-rate system, as practiced in Kosovo or the Kyrgyz Republic. A further alternative to consider could be the approach introduced in Georgia in 2011: The small business tax regime for businesses with a turnover of up to GEL 100,000 ($57,500) has a standard rate of 5 percent for small businesses in all segments; however, a business can benefit from a reduced rate of 3 percent if it can document, by presenting corresponding invoices, that business expenses amounted to at least 60 percent of turnover. The rate differentiation is not directly linked to the business activity, but is supposed to reflect the actual business profit margin. A certain disadvantage of the approach is the need for the tax administration to verify invoices presented by small businesses and the risk of disputes about the tax recognition of these invoices, but it provides an innovative method for increasing the fairness of a turnover-based presumptive tax regime. A similar approach is currently being discussed for the trade segment in Armenia.

b) Presumptive Regimes, the Business Environment, and Compliance Management
Surprisingly, despite the extensive use of presumptive tax regimes in the region, analysis of the system impact is still sporadic. Of particular importance is impact analysis on tax compliance costs and business formalization.

Compliance Cost Reduction
Business surveys provide some indication of compliance costs associated with different tax regimes. A simple comparison of compliance costs in Ukraine, for instance, indicates a cost reduction of almost 50 percent for small businesses operating in the presumptive regime. Relative costs, % Similar survey-based reviews consistently identify accounting and tax filing obligations as the major element of compliance costs, thus supporting general claims of cost reduction following the introduction of simplified accounting and filing requirements for presumptive taxpayers.
Fully exploiting the cost-benefit potential of presumptive taxation therefore requires appropriate simplification of bookkeeping and reporting requirements. A full alignment of bookkeeping standards with the requirements of a turnover-based tax system in principle would only require the business to keep records of its turnover. This approach, however, would not provide sufficient data for a presumptive taxpayer risk analysis and identification of businesses abusing the regime. Therefore, tax administrations generally require more extensive documentation of business transactions. Nevertheless, compared to accounting requirements in the general tax regime, a major simplification of the accounting burden can be achieved. Small businesses also consistently highlight problems with the number of taxes with which they are required to comply. Armenian businesses, for example, cited the number of taxes as their most widespread concern with regard to the operation of the tax system. 26 In Belarus, businesses complained that a normal SME in the general tax regime has to pay on average 12 taxes and duties. 27 Single small business taxes, which combine various national (and in some countries, local) taxes into 42 one tax payment, therefore create additional compliance cost reductions for small businesses. This can be a major benefit in practice. The single tax in Ukraine, for example, substitutes 12 other taxes and duties in addition to income/profit tax, VAT, and land tax.

Business Formalization and Compliance
Little thorough analysis is available with regard to the impact of simplified small business taxation on business formalization and the compliance behavior in the ECA region. 28 At first sight, when considering registration trends, the experience in many ECA countries looks impressive. In the short term, presumptive systems may have a relatively low initial take-up rate, likely due to a lack of information on the part of small business operators about the regime and its benefits, or distrust in its stability and implementation. Generally, however, participation tends to grow remarkably over time. However, such growth figures do not provide any information on the actual effects of presumptive tax regimes on the formalization decision of businesses. What is generally lacking in ECA countries (and worldwide) is an analysis of the previous tax situation of the small businesses joining the regime. The presumptive taxpayer population include (i) tax-registered small businesses changing their tax status and moving from the general into the presumptive regime (see also the example of Uzbekistan below); (ii) newly established businesses registering with the tax authority and selecting the presumptive taxation approach; (iii) some-hopefully not too many-larger businesses migrating into the presumptive regime. It is therefore completely unclear whether and to what extent growth in registration can be attributed to informal businesses deciding to formalize.
Several analytical efforts on the dynamics around presumptive tax regimes provide an indication that they have contributed to improved voluntary tax compliance. For Ukraine, Thießen (2002) estimates, looking at the years 1999 and 2000, that the presumptive tax for small and medium-size businesses has reduced the shadow economy by 11-14 percent. More recently, the Russian SME Resource Center conducted a survey and analysis of businesses using the simplified taxation system in the Russian Federation. 29 The analysis concluded that the adoption of the simplified tax system resulted in the legalization of about 30 percent of SME income. Similar revenue increases have been reported in Ukraine, following the introduction of the fixed tax regime, and the tax administration found that the amount of taxes collected from businesses that decided to move into the fixed-tax regime increased by four times compared to the amount collected from the same group of businesses before the fixed tax was introduced. 30

Presumptive Regime Threshold and Business Development Distortions
While the existence of a simplified presumptive tax regime may have considerable benefits for the SME segment, there are also multiple risks associated with the impact of the presumptive regime on revenue generation and the integrity of the tax regime, as well as potential detriments for business growth and development. These detriments result, in particular, from a major tax burden and compliance cost increase for businesses migrating from the presumptive into the standard tax regime, and an undesirable incentive for larger businesses to migrate downwards into the presumptive regime.
As a result, country analysis tends to reveal the erosion of the general tax regime due to the attractiveness and popularity of presumptive tax regimes. This erosion needs to be balanced against the benefits of the regime and may not always be as serious as could be assumed when glancing at anecdotal evidence. Sometimes, a skewed taxpayer distribution may simply reflect the predominance of the micro and small business segment in an economy.
Presumptive tax regimes for many small businesses not only offer the possibility of facilitating tax compliance and reducing compliance costs, but also reducing the actual business tax burden. There is no sound justification for such tax liability reduction, which violates the ability-to-pay principle; 31 in practice, however, many presumptive tax regimes result in a comparatively low tax burden compared to similar businesses operating in the standard tax regime.
Upward Migration 30 Semikolenova, Taxation of Small and Medium Enterprises, Kiev, 1999. Indeed, the percentage of small businesses taxed on a presumptive basis, which eventually voluntarily migrate into the standard tax regime, is low. This observation as such is not sufficient, however, to assume that the presumptive regime threshold establishes a business growth barrier. Such a conclusion would require the additional diagnosis that presumptive taxpayers tend to grow to a level close to this threshold and then stop showing any further increase in their turnover. Many small businesses tend to remain small and operate at the lower, and not at the upper, end of the presumptive regime turnover scale. In Ukraine, for example, looking at the distribution of unified taxpayers according to turnover levels, Alm and Saavedra (2006) find a remarkable upward migration within the system. In the five years, 2000-04, the total number of individual simplified taxpayers more than tripled. However, while the increase in the lowest band has been modest (an increase by 69 percent), taxpayers with payment obligations in the highest four bands (more than Hrv 200,000) increased by more than 12 times. This is only to a small extent due to inflation, as annual inflation rates were below 10 percent in the years analyzed. Still, more than two-thirds of presumptive taxpayers are grouped into the three lowest bands of the regime and operate not even close to its upper threshold, suggesting that the large majority of small businesses in the regime are not concerned with a potential system threshold growth obstacle.
The same pattern can be observed in Tajikistan. Analysis of the composition of taxpayers in the turnover tax regime shows that close to 100 percent of individual business operators and, more This review of taxpayer distributions does not imply that a system threshold barrier for small business growth does not exist. Instead, it indicates that the number of presumptive taxpayers affected by the system threshold is comparatively small. In addition, it is important to take into account that the presumptive regime threshold is not necessarily the only obstacle for business growth. Even without the need to abandon the preferential presumptive regime treatment, reaching the VAT threshold can become a major barrier for small business growth. A strategy to facilitate the transition from the presumptive into the standard tax regime therefore needs to include, as a core element, the facilitation of VAT compliance procedures for medium-size businesses, in particular through offering VAT cash accounting schemes. A cash accounting option could also be considered for income taxation of medium-sized businesses. While a number of Central and Eastern European countries have reduced the barrier for migration out of the presumptive regime by better aligning their small business tax regime with the general tax regime (see below), Russia, Ukraine, and Belarus have considerably increased their presumptive regimes' eligibility thresholds. In the Russian Federation, the threshold for the application of the simplified tax system (STS) for incorporated businesses was raised in 2010 from RUB 26.8 million to Rub 60 million turnover (with the staffing threshold of a maximum of 100 employees remaining unchanged); similarly in Ukraine, the presumptive regime threshold increased in 2012 for incorporated businesses from UAH 1 million to UAH 5 million, with the maximum staffing of 50 employees also remaining unchanged.
Increasing the threshold, to a large extent, eliminates the growth obstacle for businesses with a turnover below the old threshold, but does not provide an overall solution to the problem. As Alexeev and Conrad (2013) point out, a business at the RUB 60 million threshold may still face a RUB 400,000 tax increase as a result of a RUB 1 increase in turnover. It may also increase problems of system operation, and the risk of system abuse becomes more acute. An effective strategy for facilitating business migration should, instead of raising the system threshold, at least include the following components: (i) aligning the tax burden in the presumptive regime with the tax burden in the standard tax regime; (ii) imposing basic bookkeeping standards on presumptive taxpayers; (iii) introducing some compliance facilitation measures, in particular for VAT compliance, for mediumsize businesses; (iv) offering targeted taxpayer services for business migration. In addition, a riskbased audit approach aiming at identifying businesses that should be moved into the standard tax 32 Based on data from 2009, when Georgia did not offer presumptive income taxation to MSMEs. 100000 150000 Gross Revenue from Income Tax return regime, should be applied to avoid unfair competition for medium-size businesses in the standard regime.

Downward Migration
In a number of cases, migration of businesses from the standard tax regime into the presumptive tax regime is perfectly legitimate. It can either be a result of a shrinking business turnover-making the business eligible for the presumptive regime-or a business that always qualified for presumptive taxation but opted for being taxed in the standard regime and subsequently changed the system selection.
Tax administrations frequently report that they found incidences of widespread presumptive regime abuse by larger businesses. An extreme case is Ukraine, where the tax administration found cases of larger firms splitting into 20 or more small businesses, thereby qualifying them for the presumptive tax regime. As such, government officials are concerned that as much as 50 percent of all presumptive taxpayers in the system may be fraudulent. 33 Apart from such anecdotal evidence, OECD analysis shows that while the number of small companies increased between 2000 and 2015, the total industrial output of small businesses decreased from 8.1 percent to 5.5 percent during that time. These, along with the decrease in the average number of small business employees from 8 percent to 6.4 percent in 2006 suggest that companies close to the simplified system thresholds of 10 employees or 50 employees either fragment the business or under-report employment in order to remain eligible for presumptive taxation. Despite the concentration of small businesses in booming, consumption-oriented sectors like retail trade, the officially reported consolidated financial results for small businesses in all sectors except health, education, and social services was negative in 2004/05, suggesting large-scale concealment of profits. 34 Unusual migration trends can be an important indicator of system abuse, and the non-existence of a substantial medium-size business category can indicate the downward migration of medium-size businesses. Both elements can be observed in Kazakhstan, where the number of individual entrepreneurs with a registered turnover below the presumptive regime threshold of KZT 40 million suddenly increased significantly in 2009, coinciding with a reduction of the presumptive tax rate from 5 percent to 3 percent. The number of individual entrepreneurs above the eligibility threshold declined, and the number of incorporated businesses in the medium taxpayer (9,339 in 2009) segment was comparatively low. These might be seen as an indication of downward migration dynamics. In a situation where the number of businesses migrating into the presumptive regime continuously and substantially exceeds the number of businesses migrating out of the presumptive regime, the base of the standard tax regime erodes, tax revenue collection decreases, and competition increases for the few businesses remaining in the standard regime. As an example, migration trends in Uzbekistan from 2010-12 provide reasons for concerns about the long-term effect of its presumptive tax regime. In Uzbekistan, the number of net income taxpayers decreased by 25 percent over a period of two years and is now less than 10 percent of the taxpayer population. The artificial splitting up of businesses in order to abuse the presumptive tax regime results in a reduction in the overall cost efficiency of business operations. Assuming rational business decision making such splitting-up is therefore only attractive if savings in both tax payments and compliance costs exceed these efficiency losses. Eliminating the differences in the tax burden between the standard and presumptive regimes substantially reduces the incentives for these business divisions. Also, efficiency losses are higher, the smaller the newly generated business entities need to be to qualify for the presumptive regime. The split-up option is therefore more attractive in a country like Ukraine, which has a presumptive regime threshold of 50 employees and UAH 5 million ($520,000) turnover, than Latvia, which has a turnover threshold of €100,000 ($113,000) and a staffing threshold of five employees. An essential step towards reducing system abuse risks is to define a presumptive regime threshold which limits regime application to small businesses that face capacity constraints and compliance difficulties with the standard tax.
Transparent rules for the application of the regime are a second essential element. This relates in particular to the level of bookkeeping required and imposed. While a simplified cash-based bookkeeping standard is appropriate for the operation of a turnover-based presumptive tax regime, this standard is not always sufficiently enforced to verify the size of business operations in practice. approach in this respect has been taken by the Russian Federation, where the increase of the presumptive regime threshold was combined with the introduction of comprehensive accounting requirements for companies. While this is not necessarily an appropriate approach for all countries, analysis in the Kyrgyz Republic has demonstrated that the enforcement of a higher level of accounting can be well in line with the actual practice of small corporations.

Figure 21: Actual Bookkeeping Practice of Small Business in the Kyrgyz Republic
Source: IFC SME Survey.
Due to the limited transparency and control, possibilities in system abuse risks are particularly high when patent regimes with no bookkeeping requirements are extended to the small business segment. Additional safeguards need to be put in place in this case to counteract the access of larger businesses to the patent regime. In Tajikistan, for example, while the patent regime turnover threshold was relatively high at SM 200,000 ($42,000), the additional eligibility criterion of operating the business without any (non-family) employees hampers system access for larger businesses. As a result, the system abuse level seems rather moderate, with about 3.5 percent of patent holders reporting a turnover above the patent regime threshold in an anonymous survey. 35 Nevertheless, in 2013, the Tajik government used a general revision of the tax code to lower the patent threshold to SM 100,000 in order to better monitor the regime.
Additional system eligibility criteria and specific anti-abuse provisions create further barriers for system abuse. Such criteria can be, as in the case of Russia, the requirement that the majority of 35

SMCs
Farmers shares (75 percent in Russia) are owned by private individuals, that the business has no branches or representative offices, or that it does not operate in any high risk areas such as financial services, manufacturing of excisable goods, or trading in minerals. In Latvia, in addition to restrictions regarding the number of employees, only limited liability companies having solely individuals as shareholders, can apply for the micro enterprise regime.

Monitoring MSMEs
While the approach practiced in the early phase of transition to target a 100 percent audit of all registered businesses has been abolished in all countries in the region, there still seems to be an inappropriately high share of administrative capacity consumed by visiting small businesses that have low potential tax yield. From a business perspective, compliance costs are increased due to time and resources required to prepare for audits, be available during the audits, and respond to queries following the audits. The main focus of any audit program for the small business segment needs to be the identification of taxpayers who abuse the presumptive tax regime for tax minimization purposes by either substantially under-declaring their turnover or by artificially splitting up business operations. It is therefore necessary to define criteria that indicate when an artificial separation of business activities is evident. Audit activities should focus on small businesses with a turnover close to the upper threshold of the presumptive regime to verify if, according to their actual turnover, these businesses should have migrated into the medium-size business category. Moving to a more risk-based approach to tax audit, targeting in particular businesses which should be transferred to the general tax regime and major cases of turnover under-reporting, therefore needs to be a priority for tax audit reform in the region. Many Central and Eastern European countries have successfully implemented reforms in this direction, and some FSU countries, such as Kazakhstan and the Kyrgyz Republic, are following this path.

ECA Average
Even a well-designed and well-administered MSME regime generates only a very small and often even negligible percentage of total tax revenues. 36 This renders the administration of presumptive tax regimes relatively unattractive for national tax administrations, and a number of countries in the region have therefore transferred the administration of presumptive tax regimes to local governments. In these cases, the successful coordination of different levels of administration is critical for effective compliance management.

Small Business Taxation in Albania
Since the introduction of the first simplified scheme in the early 1990s, Albania has tested different special regimes and moved towards a decentralized administrative approach aiming to account for regional differences. Since 2005, income tax for small businesses with a turnover of up to ALL 8 million is administered at the local level. For micro businesses with less than ALL 2 million in turnover a simple patent applies, for businesses with a turnover between ALL 2 million and 8 million a turnover tax is used (with 7 different turnover and 3 district categories).

Overview on Presumptive Taxation Policies for MSMEs in Albania
Year System/Revisions 1992 Special tax regime for individuals (trading activities, handicrafts, and other services)

1993
"Law for small business tax": fixed tax and a tax based on gross revenues (rates of 3%, 5%, and 8% on gross income)

1998
Eligibility extended to legal entities; turnover threshold of ALL 5 million introduced. Turnover tax of 4% applied to all small business with turnover between ALL 2-5 million.
Fixed patent for businesses with annual turnover < ALL 2 million (differentiated by sector and location).

2002
Alignment of presumptive and VAT threshold at ALL 8 million; Fixed tax at the local level.

2005
Reduction of the rate of the simplified profits tax from 4% to 3% from 2005. Assignment of administration (and revenue) from small business tax to the local level.

2006
Turnover rate schedule introduced for 7 turnover levels and differentiated by region and sectors

2008
Introduction of balance sheet requirement for all small business with a turnover above ALL 2 million

2010
VAT threshold lowered to ALL 5 million

2014
Simplified net profit taxation with reduced rate of 7.5% for businesses with turnover between ALL 2 million and ALL 8 million. Administered by central tax administration and no longer by local governments. 36 Tajikistan, where small businesses account for 4.2 percent of total tax collection (patent tax 1.4 percent, simplified tax 2.8 percent), is an example with a comparatively high level of micro and small business contribution to total tax collection. In 2009, the Polish tax card and turnover tax regimes amounted to a total around $500 million in tax collections, thereby accounting for just 8.5 percent of individual business tax revenues in the country. In Armenia, following the reintroduction of the turnover tax for small businesses the tax accounted for just 1.1 percent of total tax revenue (EBRD 2013) The assignment of the small business tax administration to local government was rather unusual, resulting in important coordination challenges between municipal authorities and regional GDT offices. Most notably, the tax base of the municipalities was at risk from small business growth.
Following the assignment of small business taxation to the local level, revenue dropped sharply. In Tirana, for example, between 2005 and 2010, total revenue collected from the sector dropped by almost a third, from ALL 1.4 billion to 1 billion, despite an increase in the number of registered small taxpayers. 37 There are several explanations for this decline in revenue. Political considerations resulting from conflicting agendas at the local, regional, and central level are reported to interfere with efficient tax administration. USAID (2009) highlights that the reduction in the SBT rate is the main driver of declining revenue. Finally, the transitional challenges and the need to build municipal authority capacity in managing small taxpayers were frequently highlighted as the main driver of the sharp decline in revenue, particularly in light of coordination challenges. In some municipalities, a different coding system was developed to monitor their local tax base (instead of relying on the tax ID issued by the national registration office), undermining efficient information sharing. This lack of coordination facilitated the abuse of the patent or small business regimes, as businesses that were broken up in multiple small entities and registered with different municipalities cannot be properly tracked.

d) Possibilities to Further Improve the Presumptive Regime Design
One characteristic of simplified taxation in the ECA region is the continuous reform and modification of presumptive tax regimes. The directions of these reforms are not similar throughout the region, and there is thus no clear development of a new general architecture of presumptive taxation regimes. As a general observation, the difference in the approach to presumptive tax design between Central and Eastern European countries and FSU countries is increasing. While in the countries of the Former Soviet Union, a single rate turnover tax is still the predominant approach to simplified small business taxation, CEE countries have begun phasing out pure turnover taxes and are aiming at a better alignment of presumptive and standard tax regimes.
One method in achieving such alignment is to replace turnover as small business tax base with simplified net income calculation. With this approach, small businesses are integrated into the standard income tax regime; however, the requirement to calculate business expenses and determine net business profit is replaced by a lump sum cost deduction. Similar to turnover tax regimes, the business thus only has to calculate its gross income. Lump sum cost deduction ratios may be established for different categories of small business activities to reflect average income/expense ratios. The standard income/profit tax rate is then applied on the presumed net business income.