An Effective Tool for Promoting Local Growth and Competitiveness? The Sustainability of the Subnational Business Environment Index

ii


Introduction
Since the turn of the century, as the global economy has become increasingly integrated and competitive, countrywide surveys of business climate have proven quite popular for comparing country progress and prospects for economic growth. In 2002, the World Bank used the Doing Business Index (DBI), to begin rating the world's economies based on the ease with which small and medium-sized enterprises conduct their business. Although the DBI is probably the most widely known and cited business index, today one need not search long to find a variety of other indices, each taking its own approach to calibrating country business environments given the interests of the designing organization. 1 Where one index focuses on the regulatory regime or on national competitiveness, another is primarily concerned with economic freedom or the general investment environment. The best country-level methodologies are sophisticated and well thought out, although every index faces some limitations. The target audience for these indices, as one might expect, comprises policy makers, journalists, development practitioners, and especially businesses. To increase their index scores national governments will scramble to demonstrate areas of improvement, and the annual announcements of country rankings are major events that draw international media attention.
By 2005, externally financed subnational business environment assessments began to emerge as tools for promoting local economic growth. Applied at the level of the state, province, municipality, and other subnational units, these assessments became increasingly popular. The subnational business index is a tool designed to quantify the degree to which a developing country locality supports business growth and investment. Following one methodology in particular-the so-called Economic Growth Index (EGI)-most of these indices emphasize the quality of economic governance. The EGI was conceptualized by Malesky and the Asia Foundation in Indonesia and Vietnam and then adapted as needed in a series of Asian countries (Malesky & Merchant-Vega, 2011).
Yet, a few indices do take an approach that attempts to measure the broader enabling environment for business. Although hard data (existing statistical records) are often incorporated, typically the index arrives at a score based on the results of surveys of a sample of firms within each jurisdiction, then aggregates each score according to an established methodology, and finally ranks the jurisdictions accordingly. The scores and rankings are used to enhance awareness of the quality of the business environment, particularly within subnational institutions of government, and to help identify recommended actions to improve it. The process is also meant to promote interaction around business climate issues among subnational government leaders, the business community, and other local stakeholders. Ranking the jurisdictions is thought to engender competition among them, spurring improvement-presumably, subnational units will seek to attract investment and gain political benefits by demonstrating that they have performed well in advancing their business environment. Much of the enthusiasm for the subnational business index has come from the Asia Foundation, donor agencies such as the US Agency for International Development (USAID), and private sector associations, which naturally have a direct interest in building probusiness climates. In fact, USAID and the Asia Foundation have been involved in most of the indices examined in this paper, and we believe that we captured all of the developing country indices that have been established with external support since the mid-2000s. 2 This work provides a comparative examination of the development, application, and sustainability of 13 subnational business environment indices (BEIs). All these indices were developed, and in few cases repeatedly implemented, since 2006 in countries in Asia, Eastern Europe, and Latin America. All 13 BEIs have been financed in whole or part by USAID and other international aid agencies.
We define sustainability as the continued use of the index after the aid program's support has come to an end. We argue that sustainability is fundamental to success because sustained use is not only the ideal intent of the subnational BEI, but is also needed for new policies and practices to be adopted and become routinized. Sustainability has long been closely associated with successful implementation, and the frame of reference for sustainable development is much narrower than it was when it first emerged in the 1970s (Kidd, 1992). Most commonly, sustainability is the transfer of knowledge and skills such that the activity is continued well after the termination of donor support (Taylor 2014). USAID, for example, has defined sustainable programs to be those that "continue or evolve under their own momentum or actions, without continued donor intervention" (USAID, 2015). To be sure, not all development interventions need to be sustained to be considered successful-an immunization campaign is a good example-but the subnational BEI is intended to work over time to gradually change policies, governance, and behaviors. Arguably, the longer the BEI is applied the more effective it is likely to be, and published reports indicate that repeated application can have positive effects whether or not sustainability is achieved. Nonetheless, sustainability is the critical, observable threshold that assures us that positive effects will endure and eventually be institutionalized. For subnational BEIs, in practice, this means that some domestic entity or collection of interested parties-a government body, private sector association, research institute, or nongovernmental organization of some kind-would have to find the BEI valuable enough to continue dedicating the necessary financial resources and personnel for it to be applied for the foreseeable future.
To help provide background and context, the next section provides a brief review of the rise of the major national BEIs and their features. We go on to compare and contrast the subnational BEIs and their application in practice. Then we examine, again in comparative perspective, the successes and failures of this group of BEIs; we provide a detailed presentation of El Salvador's Municipal Competitiveness Index (MCI). Finally, in the conclusion, we consider the implications of BEIs used as a tool for local economic growth.

Emergence of National-Level Indices and Economic Governance
International developments have given impetus to the emergence of BEIs. First, a liberal world order of open markets and free flow of trade and capital continues to predominate, even as rising states in the East and South challenge the economies of the Western, industrialized nations (Ikenberry, 2011). Today, ensuring economic stability, building domestic markets, attracting foreign direct investment, and promoting trade are the hallmarks of nations looking to grow and modernize. Brazil, Russia, India, South Africa, and China, along with several smaller nations in Asia and elsewhere, are joining the most dynamic and competitive economies in the world. Building or sustaining an environment in which international business is welcomed is considered a key strategy to drive continued progress.
Second, good governance and the establishment of well-functioning institutions are now widely recognized, if not always practiced, as critical to establishing an environment in which the private sector can thrive. Both academics and practitioners have long recognized a positive relationship between institutional strength and economic performance; where governance is good, growth tends to follow (Knack & Keefer, 1995;Kaufmann, Kraay, & Mastruzzi, 2007). As Kofi Annan (1998) famously reported, "Good governance is the single most important factor in eradicating poverty and promoting development. " Economically, this has meant improving transparency and accountability, public sector capacity, regulatory reform, and a host of other institutional advancements that are targeted in the various BEIs (Dixit, 2009).
A third development giving impetus to business climate indices, particularly subnational measures, is decentralization. In the past three decades or so, scores of countries have embarked on some degree of decentralizing reform-(re)establishing elected subnational leaders, devolving new functions to lower governmental levels, and transferring new resources (see Montero & Samuels, 2004;Bardhan & Mookherjee, 2006;Bird, Ebel, & Wallich, 1995;and Oxhorn, Tulchin, & Selee, 2004). Increased subnational authority carries important implications for subnational economic governance and the generation of a climate favorable to local economic growth. Potential changes in intergovernmental relationships are particularly consequential. Can ostensibly stronger subnational governments, notably cities, now more effectively generate the conditions required to attract investment?
A fourth development driving the rise of indices is the so-called aid effectiveness agenda that has swept through the international development community. In a series of major international fora over the past decade, donor agencies and aid recipients have been engaged in building more cooperative relationships around program development and implementation. 3 A major focus has been ensuring the achievement of program results. Development agencies and implementers-prodded by their congressional and parliamentary funders-are determined to demonstrate the effects of their aid programs. Consequently, tools such as indices that help evaluate program results, particularly quantitative analysis of outcomes and impact, have attracted considerable interest and intellectual energy. They are sought after because they assist aid agencies in designing and refining their programs. Efforts to create subnational business climate indices, like indices in other fields, address the demand to demonstrate results; indeed, as noted, all of the indices were or are currently financed by international aid agencies.
Governance as it relates to economic investment and growth, or "economic" governance, appears in national and subnational indices in various ways and to varying degrees. By economic governance, we mean "the structure and functioning of the legal and social institutions that support economic activity and economic transactions, " particularly the protection of property rights, enforcement of contracts, and collective action (Dixit, 2009, p. 5). Generally, governance is most apparent in the interaction between government-or, where government is not working well or at all, informal social institutions-and civil society (Dixit, 2009). The popular DBI is the prime example of a national index that is entirely focused on governance. The DBI evaluates a country's business climate by examining the regulatory burden, mostly regulations defined in law as opposed to their operation in practice, and always as they affect small and medium-sized enterprises. The DBI, now covering 185 countries and 10 topic areas, involves the critical governmentindividual interaction on economic matters that shape governance, such as starting a business or paying taxes. In fact, the report has been criticized for implying that it provides a comprehensive measure of the business environment when, in fact, it focuses only on government regulation (World Bank, 2008;World Bank Independent Review Panel, 2013). DBI has been assessing the regulatory environment at the subnational level since 2006. These reports examine the ease of doing business in the world's cities, often the largest urban areas, and have been completed in 30 countries since 2008. They usually involve a subset of the 10 topics covered in the national index. The cities are ranked accordingly, and their progress over time is evaluated in countries in which the subnational ratings have been completed more than once.
The five remaining national indices examined here focus on the overall business climate. Although economic governance is a significant aspect of each index, other institutions and policies not related to governance are included as well. Examination of the indicators shows that four of the remaining five indices dedicate a little less than half-between 41 and 46 percent-of their conceptual framework to economic governance. The most well-known of these is the Global Competitiveness Index (GCI). Established in 2005 by the World Economic Forum, and providing the thrust of the forum's annual Global Competitiveness Report, the GCI focuses on the factors that underpin national competitiveness. To the GCI's methodologists competitiveness means "the institutions, policies, and factors that determine the level of productivity of a country" (Sala-i-Martín et al., 2014, p. 4). Most of the GCI indicators organized under its institutions and infrastructure pillars can be considered governance, but the ratings framework includes pillars that address broader categories, such as technological readiness and market size. The GCI's methodologists have also given governance greater weight in countries in the early stages of development; innovation and efficiency receive most emphasis during the middle-income and advanced development stages, respectively (Sala-i-Martín et al., 2014, pp. 9-11). 4 In addition, the GCI's creators recently began applying a modified, truncated version of the GCI conceptual framework to 33 of the world's cities. City competitiveness is defined as a city's "ability to use available inputs efficiently to drive sustainable economic growth and prosperity" (Salai-Martín et al., 2014, p. 5). Economic governance is integral to this conceptual approach, particularly as it applies to institutions, the city's regulatory framework, and infrastructure services. This work allows comparisons among cities, but is not meant to serve as an index.
Two national indices are conceived around the concept of freedom. First, the Index of Economic Freedom (IEF), drawn from the ideas of Milton Friedman, is based on the notion that economic freedom-defined as "all liberties and rights of production, distribution, or consumption of goods and services"-leads to social and economic progress (Miller & Kim, 2014, p. 1). Revised in 2007, the IEF is simply an average of the scores on 10 so-called economic freedoms, which are also organized into four broad categories: rule of law, government size, regulatory efficiency, and market openness. The freedoms include property rights, freedom from corruption, freedom from government spending, labor freedom, and financial freedom, among others. Although the extent to which the connection between the IEF's 10 indictors and prosperity is questionable (Wares & Wetterberg, 2014, p. 5), governance is clearly taken into account. The second "freedom" index, which is published in the annual Economic Freedom of the World (EFW) report (Fraser Institute, 2016) is a cousin to the IEF. The EFW is organized according to broad categories of freedom: size of government, legal system and property rights, sound money, freedom to trade internationally, and regulation. The EFW's authors gather secondary data from international sources, and 41 percent of the indicators reflect aspects of governance.
The last of the major national indices is the Economist Intelligence Unit's Business Environment Ranking (BER). The BER "measures the quality or attractiveness of the business environment in the 82 countries (previously 60) covered by the [Economist's] Country Forecasts using a standard analytical framework" (Economist Intelligence Unit 2006, p. 1). Although the BER may appear conceptually similar to the DBI, it has much less to do with governance than it does the general environment for business. The BER covers 10 categories ranging from governance-related (e.g., political environment, taxes, and infrastructure) to the business climatefocused (e.g., foreign trade and exchange controls and the macroeconomic environment). About onequarter of the 91 indicators are directly related to governance. Clearly, the national-level indices are considerably different in conceptual and measurement approaches. They also differ in most respects from subnational indices discussed below. Although influenced by the conceptual developments at the national level (e.g., the Croatia Regional Competitiveness Index [RCI] is modeled on the GCI)-these subnational indices were developed independently of the national ones. Most of the 13 subnational indices examined in this study, including the El Salvador MCI, share the core EGI methodology that has been modified somewhat for country context in each case. Whereas the national-level indices cover much of the worldthe BER involves the fewest, at 82 countriessubnational rankings have been established in relatively few countries to date. Moreover, subnational rankings provide the basis for comparisons among subnational governments within a country; crosscountry comparisons would be of minimal value in any case because of contextual differences. Given the large number of subnational entities subject to consideration in any country, subnational indices generally do not cover all of the governments at the level of government subject to the assessment. A subset of the total number of states, cities, or municipalities is typically selected.

Comparing and Contrasting Subnational Indices in Operation
Our investigation of subnational indices identified 13 BEIs. All of them were developed as a means of measuring the degree to which subnational jurisdictions of developing countries support economic or business growth. 5 It is worth noting that other countries, especially industrialized nations, have developed subnational indices of their own for the purposes of reform. In the United States, for example, at least four state business climate indices have been used for decades, often with dubious results (Fisher, 2013). The European Union (EU) has created a competitiveness index-based on GCI methodology and applied for the second time in 2013-to better understand territorial competitiveness within the EU at the regional level (Annoni & Dijkstra, 2013). Table 1 compares five major aspects of the group of indices. The comparisons help provide insights into the value of the indices as tools for international development, particularly with respect to sustainability. First, we consider country and funding origins. Nine of the 13 indices are in Asia, 3 are from Eastern Europe, and 1 is in Latin America-none is in Africa. This distribution may well indicate the relative commitment of each region to and capacity for private-sector-led investment and growth. It also reflects the successful promotion of economic governance in Asia by international aid agencies and their country counterparts. The Asia Foundation, in particular, was active beginning in Indonesia in the early 2000s, providing its own institutional funding to support subnational indices or supporting them as a co-project-implementer with other donors and private-sector organizations (Asia Foundation, 2011). However, reflecting the diminished popularity of these indices within USAID and other donors, recently the Asia Foundation has been reluctant to dedicate funds for subnational indices unless another group is willing to help finance the effort (telephone interview, March 11, 2015 ). 6

Index Origins
All the indices in this set of 13 were developed with financial assistance from at least one international aid agency. External support has proven critical to their establishment and, for those that still exist, continuation. Several of the indices were created as particular activities within a larger development project-most of which were or are USAIDfinanced-aimed at generating economic growth and enhancing competitiveness. Other aid agencies that have supported subnational index development include the EU; Britain's Department for International Development (DFID); United Nations Development Programme (UNDP); the World Bank's International Finance Corporation (IFC); the German international development corporation (GIZ); and the former Australian Agency for International Development (AUSAID). The MCI in El Salvador, Kosovo's MCI, and Vietnam's Provincial Competitiveness Index (PCI), for example, were or are components of much larger USAID-financed economic growth projects.
Developing and implementing these indices is expensive, especially for extensive surveyingthe cost for one iteration reportedly ranged from $125,000 to $200,000 depending on the country-and without international assistance they might not be implemented at all.
The heavy presence of international funding does not mean that local organizations are not actively involved and, in some cases, taking the lead. Naturally, the business sector-the investment community-is often closely associated with supporting and marketing the index. Indeed, chambers of commerce, national competitiveness councils, and even a ratings agency have supported them. Councils in Croatia and the Philippines took the lead in promoting the development of the RCI and the Cities and Municipalities Competitiveness Index (CMCI), respectively, and remain the major proponents. The Vietnam Chamber of Commerce and Industry (VCCI) is a cosponsor of the PCI with USAID, and has provided institutional support from the start. The Economic Policy and Competitiveness Research Center (EPCRC) of Mongolia, which represents some of the largest companies in the country, initiated a PCI there. In some instances local organizations that have partnered with the USAID project implementers are being mentored to take over the implementation of the index, as is the case in El Salvador. As we will see, however, these efforts are not necessarily successful for long and, in some instances, neither the business sector nor local research organizations take enough interest in the index to ensure that it prospers. Both the heavy dependence on external funding and the support for competitiveness rankings by local business associations and research organizations have important implications for sustainability.

Methodological Features
The second major feature of any index is the methodology used to conceptualize, measure, score, and rank the business environment of each subnational entity. Index designers, like their national-index counterparts, devote considerable effort to the development and in many cases refinement over time of their methodologies. Various issues-in particular, the method of scoring-drive the methodological approach; however, two decisions of particular developmental interest are central to all the indices. First, index developers must decide how they want to conceptualize the notion of a "subnational business environment. " They must identify the defining parameters to be included in the conceptualization and, as indicated by the mix of national-level indices, there can be differing Again reflecting the popularity of the original approach of the Asia Foundation and its sponsors, Table 1 shows the concept of economic governance is the basis for 10 of the 13 indices examined here. In all 10 indices, economic governance refers to the nature and operation of the subnational government and its relations with the private sector and civil society more broadly with respect to issues of economic investment and growth. The economic governance approach emphasizes, in the interest of direct subnational reform, those parameters over which the subnational government has some authority. That level of authority-or degree of decentralization-is therefore a prominent consideration. Table 2 provides the defining parameters (or subindices) for each country index, each of which includes its set of relevant indicators. As easily seen, there is considerable similarity in the core aspects of indices in the countries that have adopted the alternative "business environment as economic governance" approach. Every one of the 10, for example, includes transparency and business participation in government, which measures the ease with which businesses can access relevant government information and the degree to which businesses are consulted or can advocate in decision making. The time required for regulatory compliance, meaning the costs associated with it, is also one of the parameters for all 10 indices. In addition, almost all (9 of 10) include measures for both entry costs-that is, the time required to receive a business license-and informal or illegal charges such as bribes or corruption. Another 7 indices include parameters for the level of satisfaction with or fairness in dispute resolution and the availability of business development services and facilities. Finally, 6 of the economic governance indices include measures for ease of access to land and the protection of the rights of property; the degree to which the subnational government is proactive in promoting investment and business growth; and business losses associated with crime and insecurity, an aspect over which the subnational government may have limited control.
However, there is also a considerable measure of difference in what each country considers important, even where they share the same conception of business environment (i.e., economic governance).
Four country indices include infrastructure as a distinct parameter, for example, while two more include infrastructure as part of business development services, perhaps because subnational governments in many countries have limited control over infrastructure development. Meanwhile, reflecting the distinctiveness of country contexts and further refinements, the Economic Environment Index (EEI) of Tamil Nadu, India, is the only index to include environmental sustainability, and Vietnam's PCI is the only one to include a measure of pro-state bias and workforce training (IFMR Research, 2009).
Again, turning to Tables 1 and 2, three indices-those in Croatia, Mongolia, and the Philippines-do not use the economic governance approach. Rather, although governance is an important element of their conceptualizations, these three are best described as an attempt to examine the full set of factors that affect the subnational business environment. Croatia's RCI, drawing from the methodologies of the World Economic Forum's GCI and the Institute for Management Development's GCI, examines the business sector and environment through a series of "competitiveness pillars" such as demographics; health, education, and culture; and investments and entrepreneurial trends, including governance aspects such as local government, rule of law, and infrastructure. Mongolia's PCI, an approach drawn from a Mongolia World Competitiveness study, relies on four areas of measurement: economic efficiency; government efficiency, including elements such as the provincial budget, business legislation, and institutional framework that are familiar to the economic governance approach; business efficiency; and infrastructure of various types. The Philippines CMCI is composed of three core components: economic dynamism, government efficiency, and infrastructure; these include several aspects of governance as well, including cost of doing business, transparency in local government performance, and tax rate levels.
A second major methodological issue for index developers is the source of their data. As Table 1 shows, virtually all of the indices in the sample of 13 rely heavily on the perceptions of businesspeople ascertained through extensive surveys of randomly selected firms. Perception surveys are, by nature, subjective; they provide the opinions of the representatives of private firms about the local business climate and thus raise questions about objectivity of the results. Perceptions, moreover, are subject to a variety of influences that may have little to do with governance or the larger business environment. Existing statistical data are often seen as a means of balancing the potential bias of surveys. Quality data may not be available subnationally, however; the national statistical agency may prove to be the most reliable source of data.

Intergovernmental Linkages
The level of intergovernmental coordination or cooperation among levels of government in the promotion of pro-business policies is arguably as important to improving the subnational business environment as the specific functions and autonomy of the subnational governments themselves. Rarely is any governmental function carried out in isolation; there is typically some measure of necessary collaboration among all parties in the relationship. In index design and implementation, however, the quality of intergovernmental collaboration per se is given little attention. Precisely because they are broadly conceived, the three subnational indices-Croatia RCI, Mongolia PCI, and Philippines CMCI-do capture, to some extent, the nature of the intergovernmental relationships as they relate to the business environment. Mongolia's PCI, for example, incorporates measures of fiscal decentralization, the institutional framework, and physical and social infrastructure (i.e., health, education) encompassing all levels. Croatia's RCI also captures education policy at all levels, infrastructure, and the rule of law. Interestingly, the CMCI in the Philippines reports developing a "framework [that] integrates these [three] pillars of local economic development and competitiveness up to the regional, national, and global levels" (National Competitiveness Council of the Philippines, 2014). Some of the CMCI measures are concerned with integration, such as compliance with national directives to local government units, as an indicator of government efficiency.
Among the economic governance approaches, occasionally an indicator will reflect the evaluation of intergovernmental linkages. The proactive government pillar in the Malaysia BEI, for example, is described as "A measure of the effectiveness of federal, state, and local government programmes and of businesses' awareness of major pro-economic development programmes initiated by the federal government. " The relevant indicators include "percentage of firms aware of the [national] Economic Transformation Programme" and "percentage of firms that disagree or strongly disagree that ministries with influence over their industry don't understand the industry" (Terpstra Tong, Merchant-Vega, & Terpstra, 2012, p. 73). In El Salvador, for MCI 2013, index designers broadened the indictor for public security to include the national as well as local security efforts, a reflection of the need for intergovernmental coordination to address the problem. Mostly, however, the nature and relative importance of national-subnational linkages in economic governance are left either unassessed or ignored.

Longevity and Sustainability
Ideally, the subnational index is meant to be applied repeatedly over a period of years so as to provide a standardized measure of the quality of the business climate in a municipality, province, or other subnational unit. Over time, it can serve as means of comparative benchmarking or a gauge of progress or regress in the prospects for subnational investment and growth. However, longevity is often not the case. Index longevity, as represented in this sample, is fair at best. As shown in Table 1, more than half (7 of 13) were discontinued after being implemented a mere one or two times; 5 were implemented only once. Varied explanations can be cited for this lack of success, but most cases involved at least one of the following: The index became politicized, there was insufficient interest in it, or funding for continued use was unavailable. In Bulgaria, for example, the index ranking of the municipalities became controversial because those that did not score as well as others complained that developmental differences were not taken into account; rather than conduct a second iteration, the Institute for Market Economics decided to develop a regional approach for which data are also more easily obtainable (Skype interview, February 6, 2015). In Cambodia, the private sector was too small to engage effectively and the government was not sufficiently interested; in Malaysia, lack of government and private sector interest and the need for a source of funding were major problems (telephone interviews, December 30, 2014; January 21, 2015; February 2, 2015). In Bangladesh, the District Economic Governance Index (DEGI) was part of the government's larger donor-supported program for private-sector development that was discontinued when a new government with its own agenda took power. In addition, funding for another iteration was unavailable (telephone interviews, January 21, 2015, andMarch 11, 2015).
Conversely, we see that some indices do exhibit considerable longevity. As seen in Table 1, five of the remaining six indices have been applied at least three times; one of the six has been implemented just twice. All six appear to be continuing to function well and, as of this writing, are probably going to be implemented at least once more. The oldest index-Vietnam's PCI, which began in 2005-has been implemented for 10 consecutive years and is now highly recognized within the country and beyond. With USAID support, the Kosovo MCI has been used for 5 consecutive years and is expected to continue as long as funding is provided. Croatia's RCI has been instituted every 3 years since 2007, although the source of funding for possible future iterations is not clear.
The question of longevity naturally raises the issue of sustainability. Sustainability is more than longevity.
A BEI is sustainable when at least one in-country entity-governmental, civic, private-sector, or other-finds the tool valuable enough after external support has ended to continue financing it with its own resources for the foreseeable future. As shown in Table 1, in several cases a local organization immediately joined as a partner or later took on the index's implementation, but in these cases a larger project or another source of external support continued to finance it. These cases show that the localization of implementation may be a key first step toward sustainability. In fact, local organizations have taken on the implementation for each of the six indices that continue to be used and, in five indices, local organizations have been involved from the start: Croatia, El Salvador, Mongolia, the Philippines, and Vietnam. The sustainability status of each of the six is shown in the final column of Table 1.
Significantly, two of the six indices-the El Salvador MCI and Philippines CMCI-are advanced enough to be considered sustainable. The next iteration of the MCI, planned for 2017, is to be financed by the local Salvadoran research organization, the Advanced School of Economy and Business (Escuela Superior de Economía y Neogocios, ESEN), which also covered half the costs of the last application in 2013. The Philippines CMCI does not use surveys, but is based entirely on data collection; the data are intended to be integrated into the national data collection system. CMCI likely is less expensive than other indices and, although it receives some marketing support from USAID, is led by and has received seed funding from the country's National Competitiveness Council. The Regional Competitiveness Councils are expected to obtain resources for data collection from their local government units in the region (e-mail communication, January 28, 2015).
The Vietnam PCI is now managed by the VCCI and has recently received funding from USAID for a few additional years. The plan for this PCI is to eventually find a domestic source of support or devise a business model, as previously attempted, that can sustain it (telephone interview, January 19, 2015). Along with the Croatia RCI and Mongolia PCI, the Kosovo MCI is conducted by a local survey firm, but it remains to be seen if it will be considered valuable enough for a local partner to assume the costs of implementation. A major challenge in securing local financial support is being able to secure funding from an impartial domestic source that will not compromise the integrity of the index.

Explaining Success and Failure
In international development the degree of success of an intervention is substantially influenced by the local context. A fundamental measure of success of the subnational BEI, its sustained use, likewise depends on key contextual factors. In this small sample at least, the size and economic-or, better yet, political-strength of the private sector appears to be an important aspect. Where one finds a dynamic sector coupled with market-oriented national policy, the soil for an index appears generally more fertile. National governments facing an increasingly competitive and global marketplace are concerned about maintaining national competitiveness and attracting foreign direct investment. Firms seeking to enhance their competitive positions are more likely to recognize that supporting a pro-business index is in their self-interest. Businesses have greater incentive to engage the index, promote its use, lobby key stakeholders for its continuation, and even take over its application.
Vietnam, which is ranked by the GCI among the top half in the world in competitiveness and is a relatively large market (ranked 32nd), is the best example of this scenario. 7 The VCCI has been the local partner for the index-with USAID funding and external advisory support-since it began in 2005. As the nominal representative and advocate for the private sector interests in the country, and as a quasi-governmental organization within a single party state, VCCI is an influential proponent. The PCI has become a widely recognized reform tool. In the Philippines, which is a relatively large and fairly competitive economy (ranked in the top half of countries by the GCI), CMCI proponents have been embedding the necessary collection of data into the national statistical agency's regular data collection process. Conversely, the indices developed in Indonesia and Tamil Nadu, India, two countries with relatively large, fairly competitive markets, were not used more than twice. It may be that an index in countries of such size simply cannot resonate widely enough to capture sufficient attention and support.
Countries with a desire to improve their economic competitiveness also appear interested in using subnational indices for the long term. El Salvador, which is ranked by the GCI at 105th, among the bottom quarter in the world, faces economic pressures as it seeks to integrate into the world economy. The competitiveness of the Salvadoran economy is part of a heated policy debate, and the government is engaging with the international community on reforms to bolster investment, growth, and employment through public-private partnerships and other initiatives. The engagement of the Salvadoran Advanced School of Economy and Business (ESEN) in implementing the MCI from the start and ESEN's commitment of its own resources reflects the prominence of these issues, its belief in the efficacy of the tool for encouraging subnational policy reform and investment, and its mission as a business school. Although the Salvadoran government has not expressed an interest in financing MCI, it is using some of the sub-indices for performance measurement.
In Mongolia, to cite another example, the PCI attempts to identify the strengths and weaknesses of provincial competitiveness and generate reliable data that will allow policy makers to develop stable and effective long-term development policies. The PCI is sponsored by EPCRC, an organization that was founded by the leading companies in the country under the auspices of the president. Thus, the index has strong industry and government connections in a country that is not very competitive internationally (its GCI ranking is 102 of 138). The case of Malaysia, however, demonstrates that highly open, highgrowth economies are by no means a guarantee of interest in a subnational index. Aimed at inspiring "local authorities to improve their service delivery and develop actionable policy agendas to benefit local business, " the BEI gained no traction from the government despite initial interest by one agency (Terpstra Tong et al., 2012, p. 3). Neither the political leadership nor the private sector proved interested; the index was applied just once. In Malaysia, in some quarters, the BEI approach was viewed as too political (telephone interview, February 2, 2015).
A subnational index appears less likely to prosper where the private sector is small and the economy relatively closed or just beginning to open. In Cambodia, where the Provincial Business Environment Scorecard (PBES) was implemented twice, the business sector was seen as too small and state dependent to advocate for the index. According to the GCI, Cambodia ranks among the bottom 25 percent of the world's countries in market size and business sophistication. Moreover, the index proved to be of little interest to an authoritarian government characterized by centralized decision making; in these environments, no business association is sufficiently interested to take it up. In Kosovo, repeated implementation through a USAID project has almost certainly encouraged pro-business policies and practices. Finding a professional organization with the capacity to administer the MCI over the long term has proven difficult, although a local firm has been found to conduct the surveys (telephone interview, January 29, 2015). Private companies in Kosovo are relatively small. The major Kosovo business associations reportedly do not engage member businesses and do not represent their interests; moreover, the chamber of commerce reportedly did not see the value in the MCI unless it could control it (USAID/Kosovo, 2012, p. 16) (telephone interview, January 29, 2015). 8 The ability of an index to wade through the politics that accompany its establishment and still maintain stakeholder support appears to be a significant factor in its success. Evaluating and ranking subnational governments tends to rankle those governments that do poorly or do not improve as expected. The process may be criticized as unfair or biased; it can become politicized or subject to manipulation as subnational governments compete for the highest score. Index proponents may thus find it difficult to transfer responsibility for its application to a local counterpart-be it a government ministry or research institute-which needs to be seen as impartial. Controversy regarding the perceived fairness of the Bulgaria Local Business Environment Index (LBEI) rankings was a major reason for its subsequent replacement with a regional, less governmental approach (telephone interview, February 6, 2015).
Again, the perception of the index as political explains in part why the BEI in Malaysia-today among the world's more competitive economies (ranked 25th by the GCI) and larger markets, but also a highly authoritarian one-party system-was not taken up. When successful, after a few iterations of the index understanding of the process improves, it is increasingly accepted, and it can become a valued source of data and advice for policy makers seeking reform. Trust in the fairness of the process becomes essential, and index researchers are careful to guard methodological objectivity. In Croatia, the RCI has faced repeated questions from national and county officials unhappy with the results or changes in their rankings. In response to the concerns of subnational officials, researchers have been careful to detail the reasons for lower scores or, in other cases, emphasize the importance of evaluating the level of improvement from year to year as opposed to declaring absolute rankings (telephone interview, July 17, 2015 Designed to measure the local business climate, the MCI assesses municipal capacity to create, reform, and enforce policies that stimulate private-sector development. The MCI is used to help create a favorable environment for investment and job growth in municipalities through the improvement of municipal governance and the conditions for local economic development. The municipal scores and rankings serve to encourage healthy competition among municipalities and promote sharing of lessons and best practices among them. Currently, 108 of the most populous Salvadoran municipalities are evaluated and ranked on a weighted 10-point scale. Surveys and indicators are developed with input from a series of stakeholder meetings involving business owners, organizations, and municipal officials. Municipal data are drawn from a variety of sources, such as the official Daily Record, municipality websites, and published reports of government agencies. The MCI was revised substantially in 2013 to expand the public security sub-index, increase confidence in MCI's measurements, and provide a final baseline for future applications. The 2013 application is also notable because ESEN contributed half of the cost of its implementation and agreed to take over the entire process, cost included, for 2017 and beyond (RTI International, 2009;RTI International & ESEN, 2013).
With the aid of the MCI website, index results are widely disseminated and promoted when released. Workshops, expert panels, and media events to discuss, publicize, and promote use of the tool were organized through the larger USAID project. In 2009, for example, for the first iteration of the MCI, the project convened an event in San Salvador to detail and publicize the report and then held three municipal workshops on consecutive days outside the capital that brought together public-and private-sector representatives to discuss actions for improving business climates. MCI results are of considerable interest to elected officials-their subnational governments, after all, are being evaluated and ranked against each other-and to the business sector, the media, and other stakeholders. In El Salvador, MCI scores and rankings and the publicity surrounding themalong with support from the USAID-Municipal Competitiveness Project, through which MCI was created-appears to have helped spur municipalities to institute policy changes and in some cases dramatically increase their scores. The municipality of Zacatecoluca, for example, advanced from being one the worst ranked in 2009 to place in the top 25 percent by 2013.
In the 2013 results, the average score for the 108 municipalities was 5.96. Three of the sub-indices-Illegal Payments, Entry Costs, and Time to Comply with Regulations-ranked above average, with scores of 8.18, 7.00, and 6.23, respectively. The remaining five-Municipal Services, Public Safety, Rates and Taxes, Transparency, and Proactivity-ranked below average, with scores of 5.91, 5.63, 5.24, 5.22, and 4.97, respectively, suggesting more room for improvement in these areas. Six municipalities supported a ranking of excellent; 16 were highly ranked, 64 were average, 16 were low, and 6 were very low (RTI International & ESEN, 2013).
how they identify manipulation, and the precautions taken to maintain the integrity of the rankings" (Malesky et al., 2015, p. ii).
Developing a subnational index is sometimes accompanied by follow-on activities aimed at promoting reform. Such activities may be financed by the same project that financed the index. This was the case with MCI in El Salvador. In Kosovo, the USAID-financed Business Enabling Environment Program (BEEP), which funded the MCI there, subsequently used MCI scores to conduct intensive municipal diagnostics and identify the reforms likely to have the strongest impact. BEEP staff then made presentations across the country to recommend the specific improvements (USAID/Kosovo, 2013, p. 16). Largely, however, this group of indices has not been more deeply integrated into larger development projects; some have tended to operate independently or to be pushed aside. The failure to use index results as baselines for municipal progress, for programing guidance, and as indicators of progress is seen by some researchers as a major missed opportunity. It may explain why some indices have not seen sustained application (telephone interview, December 30, 2014).

Conclusion
This examination of subnational BEIs prompts the ultimate question, does the subnational index work as a tool for international development? It is important first to be clear about what these indices do and do not accomplish in support of pro-business reform. On one hand, the subnational index raises the public profile of subnational economic competitiveness and the prospect of reform of the business environment; marketing the tool is a big part of it. The indices provide a valuable evaluation of the environment, albeit largely from the business perspective, especially in those cases in which the index relies mostly or entirely on opinion surveys of local business representatives (to varying degrees, the use of hard data alleviates this bias). Moreover, the index identifies national and subnational policy-making constraints and opportunities. Drawing comparisons and a measure of competition among subnational governments, these indices can motivate subnational efforts to examine their strengths and weakness and make improvements.
On the other hand, indices do not provide comprehensive diagnostics of the subnational business climate. To varying degrees, they measure the perceptions of local businesses. Perceptions are widely variable, however, and respondents may have disparate understandings of the quality of the environment that may have little to do with actual progress or regress. In Kosovo, for example, researchers are unable to explain abrupt "inexplicable" shifts in the scores for some municipalities (USAID/Kosovo, 2012, p. 17). The economic governance-based indices, moreover, are more narrowly defined than broad-based approaches, although it should be noted that the EGIs are explicitly developed to address those areas over which subnational governments exercise influence. The indices in Croatia, Mongolia, and the Philippines, which as a group are less reliant on perceptions, are broader and more comprehensive. Yet, these three indices suffer from data quality issues and offer less guidance with respect to subnational government reform because subnational officials are not responsible for, or able to, address all of the policy areas covered by the indices.
Sustainability, we have argued, is a fundamental test of an index's developmental success. We should emphasize, however, that whether or not sustainability has yet to be achieved, repeated application of an index-especially when project support is provided-is likely to help generate positive instances of reform and change in businessrelated practice at the subnational level. The indices for Croatia, El Salvador, and Mongolia have been implemented three times as of this writing, and all remain active. Mongolia's PCI proponents report, for example, increased motivation to improve competitiveness within the 20 provinces of the index (personal e-mail from an index implementer, July 6, 2015). Kosovo's MCI implementers have attributed a shift toward market-oriented governancespecifically the voluntary elimination or suspension of the business license fee in 22 municipalities-to the index (USAID/Kosovo 2013, 18). Vietnam's PCI is the most exemplary of them all, having been conducted every year for more than a decade. In 2016, it reported improvements in three business areasentry costs, transparency, and time costs (Malesky et al., 2016). The challenging question is whether reported gains fostered by these BEIs are substantial enough or become sufficiently infused over a fairly short period of time to have a permanent influence on local business environments and economic progress. The longevity of the PCI provides the strongest case in the affirmative.
With regard to sustainability per se, there is some reason for optimism. El Salvador's MCI and the Philippines' CMCI (albeit now relying entirely on hard data) can be considered sustainable for the foreseeable future. The Vietnam PCI is annually updated and most likely to be sustained, although repeated attempts at sustainability demonstrate the general difficulty of the task. As noted in Table 1, three more indices are potentially sustainable, depending on whether local sources of funding can be found. One clear lesson appears to be that early involvement of a committed, credible local partner with the resources or ability to generate local support is a major part of the recipe for success. Assuming a positive outcome in Vietnam, almost a quarter of the 13 cases examined here are currently sustainable and, taking the most optimistic view, nearly half will continue to be applied by in-country organizations with local financing for years to come. The 7 cases that were discontinued early on-only one making it to a second iteration-should give pause as to the obstacles to sustainability. The future is always uncertain, however; an index that is by all appearances sustainable and widely consulted today could be dissolved at any point for any number of reasons, while another could be quickly rejuvenated with a sudden shift in the local setting.
The index is a promising tool for the promotion of policy reform in countries in which the conditions appear ripe, the index is well developed, and its value is taken seriously. Where improving competitiveness is a lead item on the policy agenda-in localities or countries looking to better integrate into the world economy and to incentivize domestic and foreign investment-securing the views of the private sector through an index of this type is reasonable and may prove highly beneficial in the long run. Such tools can raise the "collective voice" of the private sector and can serve as a means of advocating for probusiness policies. The indices clearly have influence: In Vietnam, the provinces feel the pressure to reform, while in El Salvador, index results are making their way into electoral campaigns. Mongolia sees its index as an important source of inputs for long-term development policy. Given that the business sector is a prime stakeholder in economic affairs, its voice should be among the many to be heard.
The limitations that are likely to face any index initiative cannot be dismissed. An index can be undermined by manipulation, by the politics surrounding it, or by a desire to avoid the political implications altogether. They are widely viewed as expensive to carry out. Finding an impartial entity or, as we have seen, a business-support organization that will respect the integrity of the process can be exceedingly difficult. The important contextual factors-size of the private sector, national government interest, liberalization of the economymay ultimately not prove that helpful, given other constraints, in institutionalizing the index. Even when we have a sense of the conditions that favor sustainability, identifying those conditions in advance to the point of ensuring BEI success in practice is hardly a definitive endeavor. The experience of this admittedly small sample indicates that, most of the time, the index will not be repeatedly used. The story of this group of BEIs serves as a word of caution against the widespread adoption of expensive tools or interventions perceived to be effective despite limited experience as to their actual level of success.
The positives and negatives need to be weighed against a variety of contextual and other factors that can help or hinder the process. As is often the case in development, there is no easy, clear-cut answer. Ultimately, it remains up to the practitioners, policy makers, government agencies, and stakeholders to determine-hopefully after carefully taking past experience into account-whether the subnational BEI continues to be worth the investment.