Policy Research Working Paper 10192 From Financial Development to Informality A Causal Link Salvatore Capasso Franziska L. Ohnsorge Shu Yu Prospects Group September 2022 Policy Research Working Paper 10192 Abstract Financial development reduces the cost of accessing external and especially banking sector development, on informal- financing and thus incentivizes investment in higher-pro- ity is causal. This causal link is established using a novel ductivity projects that allow firms to expand to the scale instrumental variable for domestic financial development: needed to operate in the formal economy. It also encour- financial development in other (neighboring) countries. ages participants of the informal sector to join the formal The causal link between informality and financial develop- sector to gain access to credit and financial services. This ment is stronger in countries with greater trade openness paper documents two findings. First, countries with less and capital account openness. The findings are robust to pervasive informality are associated with greater financial alternative specifications. development. Second, the impact of financial development, This paper is a product of the Prospects Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at syu2@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team From Financial Development to Informality: A Causal Link Salvatore Capasso, Franziska L. Ohnsorge, Shu Yu* JEL Classification: E26, G20. Keywords: Informal economy, financial development, cross-border financial integration, emerging market and developing economies. □ S. Capasso: Institute for Studies on the Mediterranean (ISMed-CNR); University of Naples Parthenope, Email: salvatore.capasso@ismed.cnr.it. F. Ohnsorge: World Bank; CEPR; CAMA, Email: fohnsorge@worldbank.org. S. Yu (corresponding author): World Bank, Email: syu2@worldbank.org. We gratefully acknowledge support from the World Bank Group Partnership Fund for the Sustainable Development Goals (SDG Fund). The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the World Bank and its affiliated organizations. 1 1. Introduction Informal economic activity is widespread around the world. On average, such activity accounts for about one-third of output, and informal employment captures almost one- third of total employment (Ohnsorge and Yu 2021).1 It undermines revenue collections, stunts productivity, hinders investment, and traps some of the most vulnerable workers in low-paying, unproductive employment. For policy makers in countries with widespread informality, it is a formidable challenge.2 However, without a clear understanding of the causes of informality, beyond its correlates, it is challenging to decide on the appropriate policy actions. This paper documents one cause of informality: underdeveloped financial systems.3 Financial development can influence the benefits and costs of informal economic activity undertaken by firms and households (Straub 2005). Firms in the informal sector are typically characterized by small scale, low capital-to-labor ratios, lack of investment, low productivity, a low propensity to implement new, and even high-return, technologies, and unskilled managers (Amin and Okou 2020; Capasso and Jappelli 2013; Dabla-Norris, Gradstein, and Inchauste 2008; Quintin 2008). By influencing firms’ behavior, financial development can encourage capital accumulation and productivity improvements and thus promote the transition of informal firms into the formal sector (Antunes and Cavalcanti 2007). The negative relationship between financial development and the informal economy is well-established. For different sets of countries, different time periods, different definitions of financial development, multiple different definitions of informality, and controlling for numerous alternative co-factors, many empirical studies have found a robust and significant result: greater financial development is associated with less informality (see Capasso, Ohnsorge, and Yu 2022 for a recent review). 1 In the literature, the term “informality” indicates all economic activities that are not, or only partially, covered by formal agreements. The term is very general and encompasses other more specific terms such as underground economy, shadow economy, or unofficial economy. In our analysis we refer to this general meaning. 2 For instance, Pappada and Zylberg (2017) showed that economic activity at the margin of informality can change the response of tax compliance to tax rates across countries. Prado (2011) showed that most countries, in a sample of OECD countries, would benefit from a reduction in informality. Restrepo-Echavaria (2014) demonstrated that mismeasurement of the size of the informal sector can create challenges for policy makers in understanding measured cyclical fluctuations. Ohnsorge and Yu (2021) offer a comprehensive review of policy challenges caused by large informal sectors. 3 Others have documented additional causes of informality at the firm or worker level. Kanbur (2017) and Loayza (2018) discuss the causes of informal activity at the firm and household levels. Ulyssea (2020) summarizes the effect of various policies (registration support, enforcement of employment laws, VAT credit system, trade agreements) on individual firms and of various social programs on individual workers. 2 Several theoretical studies have identified the various channels that may give rise to a negative relationship between financial development and informality, with causality running in either direction. These studies essentially compare benefits, such as access to finance, with costs, such as regulatory and tax compliance burdens, of operating in the formal sector. Broadly speaking, these studies differ in three dimensions: the modeling of financial market frictions (Straub 2005; Capasso and Jappelli 2013; Guo and Hung 2020; Bittencourt, Gupta, and Stander 2014), the incorporation of technological choices (Amaral and Quintin 2006; Elgin and Uras 2013), and the nature of informality choices (Antunes and Cavalcanti 2007, Quintin 2008; Blackburn, Bose and Capasso 2012). The main notion behind most of these studies is that in the presence of information asymmetries informal firms and individuals face a higher cost of access to credit since they are opaquer to external investors, with the result that it may be optimal to stay out of the credit market. As financial markets develop, financial contracts become more profitable, and the opportunity cost of accessing credit decreases as does the cost of formality. Moreover, since investments strongly depend on access to external financing, the choice between operating formally or informally often involves a technology choice. Conversely, aggregate constraints on the size of the financial system can also weaken the link between financial development and informality (Massenot and Straub 2016). Hence, an open question is the direction of causality. We show the causal impact of financial development on the extent of informal output for a large sample of countries using a novel instrumental variable in the two-stage least squares (2SLS) estimation. We do so in a panel setting and use a time-variant external instrumental variable, in contrast to previous studies.4 Such an approach brings the benefits of using long panel data where the fixed effects estimator becomes more consistent as the sample grows (Madsen, Islam, and Doucouliagos 2018; Davidson and MacKinnon 2006). Our choice of instrumental variable is inspired by a large literature that documents the link between domestic and foreign banking sector development. Specifically, we focus on one aspect of financial development that is likely to be most relevant for the vast majority of informal workers and firms: relationship banking, which is typically conducted through close interaction between the bank and the borrower, which often requires the establishment of bank branches. Drawing on the clear evidence of strong cross-country 4 Former studies also used the instrumental variable approach to identify the causal impact of financial development, but they tend to use time-invariant instruments or internal instruments. For instance, following La Porta et al. (1997), a large literature has explored the impact of financial development on economic development using legal origin as an instrument for financial development (see Levin 2021 and Amissah et al. 2021 for recent reviews). Other studies have instrumented financial development with variables that are also largely time-invariant, such as the legal system or latitude (Levine 1999; Levine, Loayza, and Beck 2000), religious composition (Beck and Levine 2002), and ethnic fractionalization (Naceur and Zhang 2016). Studies such as Beck, Levine, and Loayza (2000) and Rashid and Intartaglia (2017) used internal instruments via the GMM method. 3 linkages in banking activities over the past three decades, we instrument the extent of branch networks with the strength of branch networks in geographically close countries.5 This approach allows us to identify a causal link between financial development and the share of informal output in GDP. This paper documents two main findings. First, in line with former studies, less pervasive informality is associated with greater financial development. Second, the negative impact of financial development (especially banking sector development) on informality is causal. This causal link is stronger in countries with greater trade openness and capital account openness. The findings are robust to the use of alternative indicators of informality and financial development. As such, this paper makes two contributions to the literature: it is the first study to document a causal link between financial development and informal economic activity in a multi-country setting. The only previous studies that established such a causal link were specific to single countries and relied on microdata or sectoral data. Capasso and Jappelli (2013) use data on loan applications to show that access to finance increased formal activity in Italy; Catão, Pages, and Rosales (2009) and Moron, Salgado, and Seminario (2012) use household survey data to show that financial deepening increased formal employment in sectors that were most dependent on external finance in Brazil and Peru, respectively. Second, this study introduces a novel instrumental variable for domestic banking sector development: banking sector development in other (neighboring) countries. We establish that cross-country banking sector links can go beyond those established for lending (Kılınç, Seven, Yetkiner 2017; Bahadir and Valev 2015; Sever and Yucel 2021; Csonto et al. 2018), efficiency (Matousek et al. 2015; Nurboja and Košak 2017), interest rates (Rughoo and Sarantis 2014), or profitability (Olson and Zoubi 2017). Significant cross-country links exist even for banking system access, such as that captured by the extent of branch networks. The next section describes the data and correlations between financial development and informality. Section 3 introduces our novel instrumental variable for domestic financial sector development and the empirical approach. Section 4 presents the main results and results from a battery of robustness tests. The conclusion summarizes our findings and offers policy recommendations and avenues for future research. 2. Data choices and description The association between financial development and informality is well established 5 Indifferent contexts, financial development in other (neighboring) countries has been used as an instrumental variable for domestic financial development in studies such as De Haan et al. (2021) and Pleninger and Sturm (2020). Our study is the first to use this approach to establish a causal link between financial development and informality. 4 (Capasso, Ohnsorge, and Yu 2022). However, the direction of causality has yet to be identified, especially in a multi-country context. This section aims to demonstrate the presence of a causal link from financial development to informality by employing an instrumental variable approach. However, before we move on to the instrumental variable estimation, we document the association between output informality and financial development without any presumption of causality. 2.1 Data As summarized by Svirydzenka (2016), financial development materializes in different forms— as also reflected in the wide range of indicators of financial development used in the literature. Indicators range from domestic credit to the private sector to GDP ratios to equity market capitalization, bond market turnover, and the number of commercial bank branches per 100,000 adults, among others. Virtually all of these indicators, except those relating to equity market development, are significantly correlated with a range of informality measures (Capasso, Ohnsorge, and Yu 2022). Here, for our purpose, we focus on banking sector development and, specifically, access to banking services. Informal workers tend to be lower-income and lower-skilled, and informal firms tend to be smaller, less productive, and active in less capital-intensive sectors (Ohnsorge and Yu 2021). Hence, the main point of interaction of informal workers and firms with the financial system is likely to be banks, not equity or bond markets. Their transactions with banks, if any, are likely to be small in magnitude and founded on relationship banking since relationship banking, which can also be measured by a denser branch network, can decrease the probability of credit rationing despite higher screening and monitoring costs (Presbiter and Rebelotti 2014). This is the reason why, in the presence of a larger number of opaquer borrowers, banks can grant more credit by employing relationship lending technologies. One study for India, for example, showed that greater banking system access, including branch density, was associated with significant increases in enterprises in the informal sector (Raj, Sen, and Kathuria 2014). As such, we deliberately avoid using domestic credit to the private sector — a measure of banking sector depth rather than access— since aggregate credit tends to be dominated by large loans to large firms (Čihák et al. 2013). One study for the United States, for example, showed that the largest 8 percent of firms account for 68 percent of bank credit (Chorodow-Reich et al. forthcoming). Domestic credit to the private sector also tends to be highly cyclical for reasons unrelated to financial development (Claessens, Kose, and Terrones 2012). In contrast, our interest here is in financial development, stripped of its cyclical component to the extent possible and focused on the financial services relevant to informal workers and firms. That said, in Section 4.2, we test the robustness of our results by using domestic credit to the private sector as an alternative measure of financial development. Specifically, our proxy for financial development is the number of commercial bank 5 branches per 100,000 adults. The data are available in the World Development Indicators (WDI) for 183 countries from 2004-2018 (Table 1). Our proxy for financial development is significantly and positively correlated with other financial development measures commonly used in the literature, such as domestic credit to the private sector, the number of automated teller machines, account ownership, or the IMF’s financial development indices (Table 2). While there are many dimensions of informality, we focus on one for the purposes of this study: output informality (Elgin et al. 2021). For our analysis, output measures of informality are preferable to employment measures since finance is typically accessed to fund working capital or investment in physical or human capital, not employment. Specifically, output informality is proxied by Dynamic General Equilibrium (DGE) model- based estimates of informal output in percent of official GDP, as detailed in Elgin and Öztunalı (2012) and Elgin et al. (2021). DGE-based informal output as a share of GDP is available for up to 156 economies for 1990-2018. This measure of informality has been used in several previous studies of informality, including Elgin, Elveren, and Bourgeois (2021); Elgin and Iyidost (2021); Elgin, Williams, Oz-Yalaman and Yalaman (2022); and Granda-Carvajal and García-Callejas (2022). As a measure for the purposes of this study, it is preferable to the MIMIC measure of informality since the MIMIC measure includes, in its construction, indicators of financial development (specifically, cash demand). There are, however, limitations of the DGE-based estimates on informal output, such as reliance on a specific functional form or the choice of benchmarking (Elgin et al. 2021). In section 4.2, we run a series of robustness checks to address some of these limitations. 2.2 Correlation between informality and banking sector development The scatter plot in Figure 1 indicates a negative relationship between the average level of output informality and the average number of commercial bank branches (per 100,000 adults) over the period 2004-2018. A large group of emerging market and developing economies are concentrated in the upper left corner where output informality is high, but the number of commercial bank branches is limited. This negative relationship is further confirmed by t-tests where the sample is split into country-year pairs with above-median and below-median shares of informality. We test for statistically significant differences in the simple average of the number of commercial bank branches for country-year pairs with above-median informality and those with below-median informality. On average, every 100,000 adults in countries with above- median informality have access to about 11 commercial bank branches— about half the number in countries with below-median informality. The t-test suggests that these differences are statistically significant. Of course, this association between informality and financial development could simply be driven by a set of confounding factors, such as per capita incomes or the quality of 6 institutions.6 To avoid results being biased by omitted variables, we run the following panel fixed-effect estimation: = + + γ + + + (1) where is the extent of informality in country i and year t, captures the level of banking sector development, and and control for country fixed effects and year fixed effects. The estimation includes a range of controls ( ) that can influence both informality and financial development. In particular, these are real GDP per capita (in constant 2010 U.S. dollars, in logarithms) and indicators of institutional quality (control of corruption, regulatory quality, and the rule of law). Data on real GDP per capita are drawn from the World Bank’s WDI. The indicators of institutional quality are obtained from the World Bank’s Worldwide Governance Indicators, with higher values corresponding to better institutional quality. As shown in Table 3, greater banking sector development is associated with less output informality in a statistically significant manner, even when controlling for other aspects of underdevelopment (such as poor institutions and low per capita incomes). This relationship remains statistically significant when using lagged financial development. While this establishes a correlation between financial development and informality, it does not establish causality. As discussed in the introduction, several studies argue for causality running in either direction. To establish causality, we next turn to an instrumental variable approach. 3. Causal link from financial development to informality We conduct the two-stage least squares estimation with a novel instrumental variable that has not yet been used in the literature on informality but that captures a well- established, quantitatively important driver of domestic financial development: financial development in other (neighboring) countries.7 This approach provides us with a time- variant external instrumental variable to take advantage of our panel setting (Madsen, Islam, and Doucouliagos 2018; Davidson and MacKinnon 2006).8 In the next subsection, we 6 See Ohnsorge, Okawa, and Yu (2021), Medina and Schneider (2018), and Ulyssea (2020) for recent reviews on determinants of informality. See Fergusson (2006) and Voghouei, Azali, and Jamali (2011) for reviews of factors contributing to financial development. Both informality and financial development are associated with institutional quality and economic development (typically captured by per capita incomes). 7 Distance-discounted financial development in foreign countries was used as an instrumental variable (IV) for domestic financial development in studies such as De Haan et al. (2021) and Pleninger and Sturm (2020) in different contexts, but our study is the first to use it to establish a causal link between financial development and informality 8 As suggested by Madsen, Islam, and Doucouliagos (2018) and Davidson and MacKinnon (2006), the benefits of using long panel data are that the fixed effects estimator becomes more consistent 7 demonstrate the validity of our instrumental variable and, in the subsequent subsection, detail our empirical strategy. 3.1 Instrumental variable: Foreign banking sector development International banking links are sizeable and have grown over the past two decades. They have grown particularly rapidly in emerging markets and developing economies (EMDEs; Aldasoro and Ehlers 2019; Feyen et al. 2020). In one-quarter of EMDEs, international claims— both cross-border (by foreign parent banks) and local (by local branches and subsidiaries of foreign parent banks)— on the nonfinancial private sector amounted to 11 percent of banking system assets and 12 percent of GDP in 2020. Foreign banks have contributed to domestic financial development, with studies showing that they introduce new banking technologies (Kumar and Ang 2014), spur competition and improve efficiency (Klein and Olivei 1999; Levine 2001), and expand access to banking services (IMF 2015). Foreign bank presence and other cross-border banking links have contributed to the convergence in cross-country banking sector development over time (Bahadir and Valev 2015). International banking links tend to thrive with geographical proximity. For instance, during 2003-08, foreign banks operating in Central, Eastern and Southeastern Europe were largely headquartered in Western Europe (IMF 2013). Gravity models applied to bilateral financial flows suggest that financial flows grow with geographic proximity (see, e.g., Head and Ries 2008; de Sousa and Lochard 2011; and Okawa and van Wincoop 2012). In one sample of 14 reporting countries and 186 partner countries, shorter distance was associated with greater international banking integration during 1995-2017 (Bouvatier and Delatte 2015). The importance of geographic proximity may reflect the trade and multinational corporate activities that tend to deepen with proximity as banks follow their clients to provide finance (see Floy and Manova 2015, for a review). It may also reflect similar operating environments, especially in economic or monetary unions. The significant commercial banking links between countries in the Eastern Caribbean Currency Union (ECCU) have in part been attributed to the joint operating environment with a single currency and a shared central bank (Csonto et al. 2018). The predominance of such regional banking links may account for the decline of convergence in banking sector developments as geographic distances grow (Bouvatier and Delatte 2015). Therefore, we expect the link between domestic financial development and financial development in other countries to be stronger for neighboring countries. At the same time, there is no reason to suspect that banking sector development in other (neighboring) countries would be directly associated with domestic informal economic activity. There is ample evidence that the informal sector is largely excluded from the domestic formal financial system (Farazi 2014; Perry et al. 2007). A domestic informal as the sample grows, whereas the instrumental variable parameter estimates can be severely biased in small samples. 8 sector that is largely excluded from the domestic formal banking system is even more likely also to be excluded from formal banking systems in other (neighboring) countries. Therefore, banking sector development in other (neighboring) countries is unlikely to directly affect, or be affected by, the extent of domestic informal economic activity. Of course, the possibility remains that financial development abroad correlates with something abroad that has cross-country spillovers to something at home that correlates with informality at home. Through this indirect channel, the exclusion restriction condition for our selected instrumental variable could be violated. However, two factors mitigate these concerns. First, we do not use the measures of financial development (such as credit to GDP) that are most strongly correlated with output. Instead, we use bank branches per person abroad as the baseline measure of financial development— a measure that is likely to be less strongly associated with output abroad than some other measures of financial development. Second, we use two distance-weighted measures for banking sector development abroad (detailed in the next section). As argued in Pleninger and Sturm (2020) and the sources cited therein, the exclusion restriction for such instrument is satisfied if two assumptions hold: (i) distance is exogenous (which it is) and (ii) any bias stemming from the endogeneity of financial development is independent of distance (which is also the case here).9 Indeed, as expected, there is a statistically significant, positive correlation between domestic banking sector development and regional banking sector development, at least in the dimension that matters for our purposes as discussed above: the number of commercial bank branches per 100,000 adults. Figures 2 A and B show, in scatter plots, the correlation between the home country’s number of bank branches and the distance - discounted average of bank branches in other countries in the region (A) or the distance- discounted average of bank branches in all other countries in the world (B). The strength of the correlation between domestic banking sector development and banking sector development in other (neighboring) countries, at least in the dimension that is relevant for our purposes (bank branches), is also apparent in the panel regression that constitutes the first stage of the instrumental variable estimation. 3.2 Empirical strategy Formally, we establish the causal effect of financial development on informality in the following two-stage least squares instrumental variable approach: ̂ + γ + δ + φ + (second stage) = + ℎ ∗ ℎ = + ℎ− + γ + δ + φ + (first stage) 9In future work, additional robustness tests could control for regional or distance-weighted GDP abroad to test the validity of our instruments, following the approach of Acemoglu et al. (2019). 9 In the first stage, we regress the level of banking sector development— proxied by the number of commercial bank branches per 100,000 adults— in country i and year t against three variants of banking sector development in other (neighboring) countries (ℎ− ∗ − ): (i) the unweighted average of the number of bank branches in all countries with common borders to country i, (ii) the geographic distance-discounted average of the number of bank branches in all countries in the same region excluding country i, and (iii) the geographic distance-discounted average of the number of bank branches in all countries excluding country i.10 The three variants of our instrumental variable both capture foreign financial development as well as the fact that the link between domestic and foreign financial sector development is stronger for neighboring countries. Both the regional and global distance-discounted averages are defined in this equation: ∗ 2 ℎ,− =∑ ( − ) ∗ ℎ ≠ ∑≠ The proxy for geographic proximity (i.e., geodesic distance) is obtained from the CEPII database. Common regions are defined as in World Bank (2021). In both the first and second-stage regressions, we control for the same set of control variables as in equation (1), including the level of per capita income (in 2010 U.S. dollars) and the three institutional quality measures taken from WGI (i.e., control of corruption, regulatory quality, and rule of law). All regressions include year fixed effects ( ) and country fixed effects ( ). Table 4 shows the results of the first-stage regression for an unconstrained sample for 2004-2018 (some of these observations will be dropped for the full estimation due to the lack of data on informal activity). Banking sector development in country i is significantly positively correlated with banking sector development in countries that are geographically close or in the same region. This supports our hypothesis that a well-developed banking system in one country “spills over” into a better developed banking system— of the type that would be accessed by informal firms and workers— in geographically close countries. Hence, we take this as an indication that banking sector development in other (neighboring) countries is a suitable instrumental variable for domestic banking sector development. 4. Empirical results As expected, the results of this exercise point to a causal impact of banking sector development on informal activity. The following subsection summarizes the main results. These are subjected to a battery of robustness tests in the subsequent subsection. 10Similar distance-discounted instruments have been used in studies such as De Haan et al. (2021) and Pleninger and Sturm (2020). 10 4.1 Baseline results Table 5 presents the estimated causal impact of banking sector development on the share ̂ , are of informal output. The estimated coefficients for our variable of interest, ℎ of the expected signs and statistically significant. In the first-stage regressions (this time with the matching sample to the second-stage regression), the number of bank branches in geographically close countries is statistically significantly associated with respect to a larger number of domestic bank branches, regardless of the precise definition of banking sector development in other (neighboring) countries. Ten more bank branches on average in the rest of the region are associated with 8-9 more domestic bank branches (all per 100,000 adults). The post-estimation test statistics confirm that the proposed instrumental variables are strong instruments. The second stage results document the statistically significant negative effect of banking sector development, as instrumented with the average number of bank branches in other (neighboring) countries, on the share of informal output in percent of GDP. Ten more bank branches per 100,000 adults reduce informal output by 1-2 percentage points of GDP. This is statistically significant, but also economically meaningful: it is equivalent to the median annual decline in the share of informal output between 2000 and 2018. 4.2 Robustness tests The above baseline results are robust to a series of alternative exercises, including the use of lagged independent variables, restricting the samples to subsamples that differ by relevant country characteristics, and employing domestic credit to GDP as a measure of financial development. 4.2.1 Lagged dependent variables The first-stage regression is open to concerns about endogeneity since the domestic banking system itself, especially if it is large, may affect foreign banking systems, especially if they are small. Hence, in the first-stage regression, the contemporaneous value of foreign bank branches is replaced with the lagged value of foreign bank branches. Table 6 shows that the main results are robust: greater banking sector development, as instrumented here, reduces the informal share of output. 4.2.2 Trade openness and capital account openness Our choice of instrumental variable— foreign banking sector development— rests crucially on two assumptions: first, that banking systems can transact easily across borders, including setting up foreign branches and subsidiaries, and second, that there are powerful incentives for banking systems to transact, including strong international trade or multinational corporate activity. In countries that are closed to foreign capital flows, such transactions should be severely impaired; in countries that are closed to international trade, there may be little incentive for foreign banks to follow their clients 11 abroad. Hence, in both cases, our choice of instrument should be less relevant. We would therefore expect the 2SLS regression to yield largely insignificant results in countries that allow limited trade or international capital flows. Tables 7 and 8 show that this is indeed the case. Our approach yields a significant effect of banking sector development on informality only in countries that are highly open to foreign trade (Table 7) and capital flows (Table 8). In these two tables, we split the sample into countries with above-median trade openness (defined as trade in percent of GDP) or capital account openness (defined as in Chinn-Ito 2006) and those with below- median trade openness or capital account openness. In the first-stage regression, the coefficient estimates are indeed somewhat smaller in countries that are less open to trade and capital flows than in those that are more open. In the second-stage regression, banking sector development, as instrumented here, has no statistically significant effect on informality in countries that are less open to trade and capital flows but does have a statistically significant effect in countries that are more open to trade and capital flows. 4.2.3 Alternative indicators of financial development We have deliberately chosen the number of commercial bank branches per 100,000 adults for the reasons laid out in Section 2. That said, an alternative indicator of financial development that has often been used in the literature is domestic credit to the private sector in percent of GDP. While this indicator is less suitable than the number of bank branches for our purposes, because it is “polluted” by cyclical movements and because it is mainly driven by formal-sector firms, Table 9 shows the results using domestic credit to the private sector (in percent of GDP) as a measure of financial development. The main results are robust: greater financial sector development leads to a lower share of informal output. 4.2.4 Alternative indicators of informality For our baseline results, we deliberately chose the share of informal output in GDP, for the reasons laid out in Section 2. That said, while conceptually less appropriate for the exercise conducted here, the DGE-based estimates on output informality of Elgin and Oztunali (2012) have some limitations. First, they rely on strong assumptions about the functional form of activity in the informal and formal sectors and about the relationship between formal and informal productivity (Orsi, Raggi, and Turino 2014; Schneider and Buehn, 2016). Second, like the MIMIC approach, they require base-year estimates of the informal economy from another independent study to calibrate the size of the informal economy (Ihrig and Moe 2004).11 11In the case of Elgin and Oztunali (2012), productivity in the informal economy was assumed to grow at the average of the growth rates of formal sector productivity and the capital stock (Solis- Garcia and Xie 2018). The model was calibrated by matching the productivity in the informal sector to the size of the informal sector in 2007 of the series reported in Schneider, Buehn, and Montenegro (2010). 12 Therefore, we conduct a series of robustness checks using alternative measures of informality, including the survey-based share of self-employed in total employment, the MIMIC measure of informality, and a DGE-based measure of informality that is benchmarked to survey-based estimates of self-employment. The survey-based share of self-employment does not rely on strong assumptions but is more tenuously linked to credit than output-based measures. It is available for 160 economies over 2000-18. The MIMIC-based estimates cover a wider range of indicators associated with informality but include a measure of financial size in its construction. Estimates are taken from Medina and Schneider (2019) and are available for 155 economies over the period 2000-17. The model of Elgin and Oztunali (2012) benchmarked to survey-based estimates of self- employment is decoupled from MIMIC-based estimates of informality and does not require ad-hoc assumptions regarding the relationship between formal and informal productivity.12 It is available for up to 155 economies during 2000-18. Table 10 shows that the baseline results are largely robust to the use of alternative informality measures. All regressions that instrument the informal economy with commercial bank branches in the region or with distance-discounted commercial bank branches continue to indicate that greater financial development reduces the share of the informal economy. 5. Conclusion It is well-established in the literature that widespread informal economic activity is associated with underdeveloped financial systems through a wide range of mechanisms. What has not been well-established is the direction of causality, at least in a multi-country setting. In this paper, we use an instrumental variable approach to establish such a causal link. Using a novel instrument— financial sector development in other (neighboring) countries— we show that greater financial sector development, here proxied by access to banking services, reduces the share of informal output to GDP. The impact is stronger in countries with greater trade openness and capital account openness. These findings are robust to the use of alternative informality measures and various indicators for financial development. Confidence in a causal effect of financial development supplies policy makers with the information needed to direct policy action. It suggests that efforts to broaden access to the financial system can help reduce informal activity. This is likely to be particularly helpful when combined with other policies. 12See the Appendix for details regarding calibrating the model of Elgin and Oztunali (2012) to survey-based estimates of self-employment. 13 Future research could broaden this analysis to other measures of financial sector development. Here, we focus our exercise on banking sector development and specifically the extent of commercial bank branch networks. This choice reflected the theoretical and empirical support for a link between domestic and foreign banking sectors. However, other such links may exist in other parts of the financial systems and could be further explored. For example, in some regions (such as Sub-Saharan Africa), the use of mobile money and digital payments has grown rapidly in recent years (World Bank 2022; Yermack 2018). On average in EMDEs, the share of adults with a mobile money account has more than doubled from 6.5 percent of the people aged 15 years or older in 2014 to 14.8 percent only three years later, in 2017, and to 24 percent in 2021. That said, this is still well below the share of adults with an account at a financial institution (52.3 percent of the population aged 15 years or above in 2021). During our sample period of 1990-2018, data on mobile money use is only available for 2014 and 2017, too limited to yield meaningful econometric results. Future research, drawing on additional data, could explore the link between mobile money and informality further. 14 Tables and Figures Figure 1. Output informality and bank sector development (2004-18) Notes: Output informality is proxied by DGE-based estimates on informal output in percent of GDP. Financial development is measured as the number of commercial bank branches per 100,000 adults. Both measures are averaged over the period 2004-18 here. See section 2 for detailed data description. 15 Figure 2. Alternative measures of financial development spillover A. Distance-discounted commercial bank B. Distance-discounted commercial bank branches in the rest of the region branches in the rest of the world Notes: Financial development is measured as the number of commercial bank branches per 100,000 adults. All measures are averaged over the period 2004-18 here. See section 2 for detailed data descriptions. 16 Table 1. Data description Variable Description Source #ctry Perio Obs Mean SD. d Domestic credit to private sector refers to financial resources provided to the private World sector by financial corporations, such as Development Domestic credit to through loans, purchases of nonequity Indicators 183 2000 3,009 50.39 44.26 private sector securities, and trade credits and other (WDI) -18 accounts receivable, that establish a claim for repayment. Expressed in percent of GDP. Commercial bank The number of commercial bank branches per 183 2004- 2,591 17.44 16.41 WDI branches 100,000 adults 18 Real GDP per capita GDP per capita (constant 2010 U.S. dollars; 188 2000 3,525 8.47 1.45 WDI (in logs) in -18 logs). Perceptions of the extent to which public World power is exercised for private gain, including Governance Control of both petty and grand forms of corruption, as 189 2000 3,564 -0.07 0.99 Indicators Corruption well as "capture" of the state by elites and -18 (WGI) private interests. A higher value corresponds to better governance outcomes. Perceptions of the extent to which a country s citizens are able to participate in selecting their government, as well as freedom of 189 2000 3,580 -0.08 0.99 Rule of Law WGI expression, freedom of association, and a free -18 media. A higher value corresponds to better governance outcomes. It captures perceptions of the extent to which agents have confidence in and abide by the rules of society, and in particular the quality Regulatory quality of contract enforcement, property rights, the WGI 189 2000 3,551 -0.06 0.97 police, and the courts, as well as the -18 likelihood of crime and violence. A higher value corresponds to better governance outcomes. 17 Dynamic General Equilibrium (DGE) model- DGE-based output based estimates on informal output in percent Elgin et al. 156 2000 2,924 30.86 12.02 informality of official GDP, calibrated as in Elgin and (2021) -18 Oztunali (2012). Share of self- Survey-based self-employment in percent of Elgin et al. 163 2000 1,906 31.88 21.67 employment totalemployment. (2021) -18 Multiple Indicators Multiple Causes (MIMIC) Medina MIMIC-based model-based estimates of informal output in and 155 2000 2,790 29.67 12.38 output informality percent of official GDP. Schneider -17 (2019) DGE-based output Dynamic General Equilibrium model-based informality, estimates on informal output in percent of Authors 155 2000 2,283 43.55 31.11 calibrated to survey- official GDP, calibrated using survey-based -18 based self- self-employment (see Appendix for details). employment 18 Table 2. Correlations among various proxies for financial development [1] [2] [3] [4] [5] [6] [7] Commercial Domestic credit ATMs Account Financial Financial Financial bank branches to priv. sect. ownership development institutions markets Commercial bank branches 1 (per 100,000 adults) Domestic credit to private 0.476*** 1 sector (% of GDP) Automated teller machines 0.541*** 0.578*** 1 (ATMs; per 100,000 adults) Account ownership 0.529*** 0.677*** 0.682*** 1 (percent of age 15+) IMF’s “financial 0.539*** 0.809*** 0.754*** 0.799*** 1 development” index IMF’s “financial institutions” 0.689*** 0.802*** 0.792*** 0.827*** 0.928*** 1 development index IMF’s “financial markets” 0.366*** 0.733*** 0.648*** 0.696*** 0.956*** 0.777*** 1 development index Notes: Global Financial Development Database=GFDD. * p < 0.05, ** p < 0.01, *** p < 0.001. “Account ownership” is the percentage of survey respondents (aged 15 or above) who report having an account (by themselves or together with someone else) at a bank or another type of financial institution or report personally using a mobile money service in the past 12 months. The IMF’s “financial development” index has the following two components: “financial markets” development and “financial institutions” development”. The IMF’s “financial markets” development index captures access to, and depth and efficiency of, a country’s stock and debt markets. The IMF’s “financial institutions” development index measures how developed financial institutions are in terms of their dept h (size and liquidity), access (the ability of individuals and companies to access financial services), and efficiency (the ability of institutions to provide financial services at low cost and with sustainable revenues). Data on financial development over the period 2000-18 are available from the World Bank’s Global Financial Development Database, the World Development Indicators, the Global Findex Database, and the IMF’s Financial Development Index Database. 19 Table 3. Fixed-effect estimator: output informality and financial development [1] [2] [3] [4] Real GDP pc (in logs) -5.58*** -5.48*** -3.68*** -2.52*** (1.64) (1.35) (1.07) (0.95) Commercial bank branches (t) -0.04** -0.04** (0.02) (0.02) Commercial bank branches (t-3) -0.05*** (0.02) Commercial bank branches (t-7) -0.03* (0.02) Control of Corruption -0.33 -0.82 -0.59 (0.51) (0.59) (0.40) Regulatory Quality -0.06 -0.26 0.14 (0.64) (0.86) (0.62) Rule of Law 0.22 0.47 -0.56 (0.56) (0.70) (0.53) Observations 2,157 2,157 1,736 1,130 Number of countries 154 154 154 153 R-squared 0.56 0.56 0.52 0.46 Adj R-sq 0.55 0.55 0.51 0.45 Notes: Standard errors clustered at the country level in parentheses, *** p<0.01, ** p<0.05, * p<0.1. All fixed-effect regressions also control for year-fixed effects. See section 2 for detailed data descriptions. The sample covers up to 154 economies over the period 2004-2018, due to the data availability of commercial bank branches. Commercial bank branches are in per 100,000 adults. GDP per capita is in constant 2010 U.S. dollars. “Commercial bank branches (t -n)” capture “commercial bank branches” lagged by n years. 20 Table 4. The linkage between domestic and foreign banking development Dep var=Commercial bank branches [1] [2] [3] Real GDP per capita (in logs) 9.59*** 8.58*** 9.63*** (2.64) (2.62) (2.92) Control of Corruption 3.13** 2.74** 3.23** (1.24) (1.16) (1.29) Regulatory Quality 1.51 1.24 0.49 (0.93) (0.87) (0.92) Rule of Law -0.44 -0.25 -0.67 (1.20) (1.20) (1.21) Commercial bank branches in the region 0.83*** (0.15) Distance-discounted commercial bank branches (regional) 0.82*** (0.13) Distance-discounted commercial bank branches (global) 2.95*** (0.63) Observations 2,526 2,495 2,497 Number of countries 180 177 177 R-squared 0.26 0.30 0.20 Adj R-sq 0.26 0.30 0.19 Notes: Standard errors clustered at the country level in parentheses, *** p<0.01, ** p<0.05, * p<0.1. Fixed-effect models are estimated here. All regressions include year dummies. The data are for the period 2004-18. Commercial bank branches are in per 100,000 adults. GDP per capita is in constant 2010 U.S. dollars. See section 2 for detailed data descriptions. Data are between 2004 and 2018. 21 Table 5. Baseline results: Two-stage least squares estimation [1] [2] [3] 2nd stage: Output informality Real GDP per capita (in logs) -3.86*** -3.89*** -3.28* (1.49) (1.51) (1.68) Control of Corruption -0.09 -0.10 -0.04 (0.56) (0.57) (0.60) Regulatory Quality 0.09 0.09 0.14 (0.65) (0.65) (0.66) Rule of Law 0.17 0.17 0.19 (0.57) (0.57) (0.58) Commercial bank branches -0.16*** -0.15*** -0.20*** (0.05) (0.04) (0.07) 1st stage: Banking sector development Commercial bank branches in the region 0.88*** (0.17) Distance-discounted commercial bank branches (regional) 0.81*** (0.15) Distance-discounted commercial bank branches (global) 2.99*** (0.70) Observations 2,155 2,155 2,157 R-squared 0.99 0.99 0.99 Kleibergen-Paap rk LM statistic 22.59*** 23.70*** 17.66*** Cragg-Donald Wald F statistic 26.31 31.17 18.29 Notes: Standard errors clustered at the country level in parentheses, *** p<0.01, ** p<0.05, * p<0.1. Here common controls and constant terms are omitted for brevity. All regressions include both country fixed effects and year fixed effects. The data are for the period 2004-18. Commercial bank branches are in per 100,000 adults. GDP per capita is in constant 2010 U.S. dollars. The dependent variable at the second stage is DGE-based output informality (in percent of GDP). See sections 2 and 3 for detailed data descriptions and empirical methods. Data are for 2004-18. 22 Table 6. Robustness test: Lagged dependent variables in first-stage regression [1] [2] [3] 2nd stage: Output informality Real GDP per capita (in logs) -3.33** -3.31** -2.70* (1.43) (1.40) (1.53) Control of Corruption -0.16 -0.14 -0.05 (0.62) (0.63) (0.67) Regulatory Quality -0.05 -0.06 -0.03 (0.66) (0.65) (0.66) Rule of Law 0.35 0.37 0.41 (0.59) (0.58) (0.60) Commercial bank branches -0.16*** -0.16*** -0.21*** (0.04) (0.04) (0.06) 1st stage: Banking sector development Commercial bank branches in the region (t-1) 0.94*** (0.17) Distance-discounted commercial bank branches (regional; t-1) 0.86*** (0.15) Distance-discounted commercial bank branches (global; t-1) 3.36*** (0.72) Observations 2,000 2,021 2,023 R-squared 0.99 0.99 0.99 Kleibergen-Paap rk LM statistic 24.47*** 25.67*** 20.89*** Cragg-Donald Wald F statistic 29.29 35.16 22.02 Notes: Standard errors clustered at the country level in parentheses, *** p<0.01, ** p<0.05, * p<0.1. Here common controls and constant terms are omitted for brevity. All regressions include both country fixed effects and year fixed effects. The dependent variable at the second stage is DGE-based output informality (in percent of GDP). The data are for the period 2004-18. Commercial bank branches are per 100,000 adults. GDP per capita is in constant 2010 U.S. dollars. See sections 2 and 3 for detailed data descriptions and empirical methods. Data are between 2004 and 2018. 23 Table 7. Robustness test: Trade openness [1] [2] [3] [4] [5] [6] High trade openness Low trade openness 2nd stage: Output informality Real GDP per capita (in logs) -0.70 -0.58 -0.17 -8.86*** -8.83*** -9.16*** (1.51) (1.51) (1.85) (2.91) (2.89) (3.43) Control of Corruption 0.34 0.34 0.34 -0.77 -0.76 -0.83 (0.49) (0.50) (0.55) (0.86) (0.85) (0.92) Regulatory Quality 0.86* 0.87* 0.88* -1.12 -1.12 -1.14 (0.49) (0.50) (0.53) (0.96) (0.96) (0.94) Rule of Law -0.68 -0.67 -0.65 1.55* 1.55* 1.61* (0.65) (0.66) (0.69) (0.84) (0.84) (0.89) Commercial bank branches -0.18*** -0.19*** -0.23*** -0.04 -0.04 -0.02 (0.06) (0.05) (0.08) (0.10) (0.09) (0.15) 1st stage: Banking sector development Commercial bank branches in the region 0.93*** 0.64** (0.21) (0.29) Distance-discounted commercial bank branches (regional) 0.54** 0.51*** (0.22) (0.18) Distance-discounted commercial bank branches (global) 3.21*** 1.76 (0.86) (1.11) Observations 1,122 1,122 1,122 1,033 1,033 1,035 R-squared 0.99 0.99 0.99 0.99 0.99 0.99 Kleibergen-Paap rk LM statistic 16.63*** 18.66*** 13.86*** 4.61** 4.47*** 2.78* Cragg-Donald Wald F statistic 18.83 23.62 13.78 4.7 5.11 2.52 Notes: Standard errors clustered at the country level in parentheses, *** p<0.01, ** p<0.05, * p<0.1. Here common controls and constant terms are omitted for brevity. All regressions include both country fixed effects and year fixed effects. The dependent variable at the second stage is DGE-based output informality (in percent of GDP). “High (low) trade openness” are countries where trade flow (i.e., imports plus exports) as a share of GDP is above (below) median. Commercial bank branches are per 100,000 adults. GDP per capita is in constant 2010 U.S. dollars. See sections 2 and 3 for detailed data descriptions and empirical methods. Data are between 2004 and 2018. 24 Table 8. Robustness check: Capital account openness [1] [2] [3] [4] [5] [6] High capital account openness Low capital account openness 2nd stage: Output informality Real GDP per capita (in logs) -2.07 -1.84 -0.72 -5.75** -5.88*** -4.24 (1.57) (1.62) (2.20) (2.24) (2.26) (6.83) Control of Corruption 0.90 0.90 0.80 -0.73 -0.77 -0.28 (0.62) (0.64) (0.79) (0.81) (0.80) (2.31) Regulatory Quality 1.20** 1.22** 1.33** -0.81 -0.83 -0.61 (0.52) (0.52) (0.52) (1.09) (1.08) (1.70) Rule of Law -1.07 -1.04 -0.77 1.56* 1.52* 2.03 (0.65) (0.67) (0.74) (0.90) (0.87) (2.36) Commercial bank branches -0.14*** -0.16*** -0.22*** -0.33 -0.29 -0.81 (0.04) (0.04) (0.07) (0.26) (0.20) (2.17) 1st stage: Banking sector development Commercial bank branches in the region 1.01*** 0.35** (0.23) (0.16) Distance-discounted commercial bank branches (regional) 0.88*** 0.38** (0.18) (0.16) Distance-discounted commercial bank branches (global) 3.45*** 0.30 (0.96) (0.71) Observations 1,087 1,087 1,089 1,068 1,068 1,068 R-squared 0.99 0.99 0.99 0.98 0.98 0.95 Kleibergen-Paap rk LM statistic 17.35*** 18.59*** 14.08*** 4.37** 5.26** 0.20 Cragg-Donald Wald F statistic 19.85 22.94 12.91 4.80 5.39 0.18 Notes: Standard errors clustered at the country level in parentheses, *** p<0.01, ** p<0.05, * p<0.1. Here common controls and constant terms are omitted for brevity. All regressions include both country fixed effects and year fixed effects. The dependent variable at the second stage is DGE-based output informality (in percent of GDP). “High (low) capital account openness” are countries where Chinn-Ito index is above (below) median. Commercial bank branches are per 100,000 adults. GDP per capita is in constant 2010 U.S. dollars. See sections 2 and 3 for detailed data descriptions and empirical methods. Data are between 2004 and 2018. 25 Table 9. Robustness check: Domestic credit to private sector [1] [2] [3] 2nd stage: Output informality Real GDP per capita (in logs) -6.90*** -6.98*** -6.86*** (1.66) (1.62) (1.70) Control of Corruption -0.38 -0.36 -0.41 (0.47) (0.46) (0.47) Regulatory Quality -0.21 -0.21 -0.22 (0.57) (0.56) (0.57) Rule of Law 0.58 0.51 0.64 (0.61) (0.61) (0.64) Domestic credit to private sector -0.04** -0.04* -0.05* (0.02) (0.02) (0.03) 1st stage: Domestic credit Domestic credit to private sector in the region 0.53*** (0.15) Distance-discounted domestic credit to private sector 0.70*** (regional) (0.18) Distance-discounted domestic credit to private sector 2.21*** (global) (0.65) Observations 2,511 2,511 2,521 R-squared 0.99 0.99 0.99 Kleibergen-Paap rk LM statistic 11.90*** 12.55*** 11.47*** Cragg-Donald Wald F statistic 12.37 14.48 11.63 Notes: Standard errors clustered at the country level in parentheses, *** p<0.01, ** p<0.05, * p<0.1. Here common controls and constant terms are omitted for brevity. All regressions include both country fixed effects and year fixed effects. The dependent variable at the second stage is DGE-based output informality (in percent of GDP). Domestic credit to private sector is in percent of GDP (taken from WDI). See sections 3 and 4 for detailed data descriptions and empirical methods. GDP per capita is in constant 2010 U.S. dollars. Data are between 2000 and 2018. 26 Table 10. Robustness check: Alternative Share of MIMIC-based DGE-output informality, measures of informality2nd stage: self-employment output informality calibrated to self-employment [1] [2] [3] [4] [5] [6] [7] [8] [9] Real GDP pc (in logs) -2.52 -2.70 2.53 -8.86*** -8.81*** -7.91*** 2.58 3.06 7.85 (3.15) (2.99) (4.75) (1.10) (1.09) (1.54) (4.24) (4.08) (6.47) Control of Corruption -1.23 -1.42 -0.93 0.39 0.40 0.48 -1.35 -1.45 -1.10 (1.40) (1.37) (1.66) (0.51) (0.51) (0.58) (1.63) (1.62) (1.90) Regulatory Quality 1.25 1.07 1.61 0.03 0.04 0.11 -1.69 -1.95 -1.40 (1.56) (1.54) (1.69) (0.53) (0.53) (0.53) (2.55) (2.47) (2.52) Rule of Law -0.15 0.01 0.34 -2.82*** -2.82*** -2.84*** 0.91 1.09 1.30 (1.67) (1.68) (1.88) (0.57) (0.57) (0.61) (2.11) (2.13) (2.32) Commercial bank branches -0.22* -0.21** -0.45** -0.08** -0.09** -0.16** -0.37** -0.39*** -0.63** (0.13) (0.10) (0.21) (0.04) (0.04) (0.08) (0.17) (0.14) (0.26) 1st stage: Banking sector development Commercial bank branches in the region 0.82*** 0.89*** 0.80*** (0.20) (0.18) (0.19) Distance-discounted commercial bank 0.80*** 0.83*** 0.78*** branches (regional) (0.17) (0.15) (0.16) Distance-discounted commercial bank 2.84*** 2.84*** 2.77*** branches (global) (0.81) (0.70) (0.80) Observations 1,419 1,399 1,401 2,051 2,051 2,053 1,530 1,517 1,519 R-squared 0.97 0.97 0.97 0.99 0.99 0.98 0.97 0.97 0.96 Kleibergen-Paap rk LM stat. 16.98*** 20.65*** 12.27*** 21.87*** 23.15*** 15.87*** 16.83*** 20.24*** 11.73*** Cragg-Donald Wald F statistic 17.44 22.95 12.42 25.44 31.18 16.26 17.67 23.17 11.91 27 Notes: Standard errors clustered at the country level in parentheses, *** p<0.01, ** p<0.05, * p<0.1. Here common controls and constant terms are omitted for brevity. All regressions include both country fixed effects and year fixed effects. The dependent variable at the second stage are different measures of informality: share of self-employment (in percent of total employment; survey-based), MIMIC- based output informality (in percent of GDP), and DGE-based output informality, calibrated to the share of self-employment (percent of total employment). Commercial bank branches are per 100,000 adults. GDP per capita is in constant 2010 U.S. dollars. See sections 3 and 4 for detailed data descriptions and empirical methods. Data are between 2000 and 2018. 28 Appendix: Calibrating DGE estimates using survey-based self-employment data This appendix explains the benchmarking of the DGE model of Elgin and Oztunali (2012) to survey-based self-employment data. In Elgin and Oztunali (2012), employment in the informal sector, has the following form: 1 1− 1 1− −1+ = {(1− )(1−) [(1− ] } (1). ) After transforming equation (1), can be expressed as follows: 1− (1 − )(1 − ) = (2). 1 1− − 1 + [ ] (1 − ) Following Loayza and Rigolini (2011) and Fiess et al. (2010), data on self-employment provided by Elgin et al. (2021) can be used to define and calculate using equation (2). Following the earlier literature, is assumed to be equal to 0.36 while takes the country average of the depreciation rates reported in Penn World Tables 9.1 (PWT). Following Ihrig and Moe (2004), is assumed to be 0.425. Capital stock ( ) and formal employment ( ) are obtained from PWT 9.1 and updated using the World Development Indicators. Assuming a balanced budget for the government, is obtained as government spending in percent of GDP reported in PWT 9.1. Rewriting the production function of the informal sector ( ) using equation (2), we find to be a function of and : (1 − )(1 − ) 1 = ∗ ∗ ( )1− 1 1− − 1 + [ ] (1 − ) which gives >0. Since > 0, it is possible that the DGE estimates will move procyclically in the presence of large shocks in formal productivity. However, when other types of shocks occur at the same time, may not move procyclically. For instance, if 29 and experienced simultaneous shocks in different directions , might move countercyclically. References Acemoglu, D., S. Naidu, P. Restrepo, and J. A. Robinson. 2019. “Democracy Does Cause Growth.” Journal of Political Economy 127(1): 47-100. Aldasoro, I. and T. Ehlers. 2019. “Concentration in Cross-Border Banking.” BIS Quarterly Review June: 1-11. Amaral, P. S., and E. Quintin. 2006. “A Competitive Model of the Informal Sector.” Journal of Monetary Economics 53(7): 1541-1553. Amin, M. and C. Okou. 2020. “Casting a Shadow: Productivity of Formal Firms and Informality.” Review of Development Economics 24(4): 1610-1630. Amissah, E., S. Bougheas, F. Defever, and R. Falvey. 2021. “Financial System Architecture and the Patterns of International Trade.” European Economic Review 136: 103751. Ang, J., and S. Kumar. 2014. “Financial Development and Barriers to the Cross-border Diffusion of Financial Innovation.” Journal of Banking & Finance 39: 43-56. Antunes, A. R., and T. Cavalcanti. 2007. “Start Up Costs, Limited Enforcement, and the Hidden Economy.” European Economic Review 51(1): 203 –224. Bahadir, B., and N. Valev. 2015. “Financial Development Convergence.” Journal of Banking & Finance 56: 61 –71. Beck, T., R. Levine and N. Loayza. 2000. “Finance and the Sources of Growth.” Journal of Financial Economics 58: 261 –300. Beck, T., and R. Levine. 2002. “Industry Growth and Capital Allocation: Does Having a Market- or Bank-Based System Matter?” Journal of Financial Economics 64 (2): 147– 80. Berrak, B., and N. Valev. 2015. “Financial Development Convergence.” Journal of Banking & Finance 56 (C): 61-71. Bittencourt, M., R. Gupta, R., and L., Stander. 2014. “Tax Evasion, Financial Development and Inflation: Theory and Empirical Evidence.” Journal of Banking and Finance 41(1): 194 –208. 30 Blackburn, K., N. Bose, and S. Capasso. 2012. “Tax Evasion, the Underground Economy and Financial Development.” Journal of Economic Behavior and Organization 83(2): 243 –253. Bouvatier, V., and A. Delatte. 2015. "Waves of International Banking Integration: A Tale of Regional Differences," European Economic Review 80(C): 354-373. Capasso, S., and T. Jappelli. 2013. “Financial Development and the Underground Economy.” Journal of Development Economics 101 (March): 167-178. Capasso, S., F. Ohnsorge, and S. Yu. 2022. “Financial Development and Informality: A Literature Review.” The Manchester School. Catão, L., C. Pages, and M. R. Rosales. 2009. “Financial Dependence, Formal Credit, and Informal Jobs: New Evidence from Brazilian Household Data,” IDB Working Paper 118, Inter-American Development Bank, Washington, DC. Chodorow-Reich, G., O. Darmouni, S. Luck, and M. Plosser. Forthcoming. “Bank Liquidity Provision Across the Firm Size Distribution.” Journal of Financial Economics. Čihák, M., A. Demirgüč-Kunt, E. Feyen, and R. Levine. 2013. “Financial Development in 205 economies, 1960 to 2010." National Bureau of Economic Research, No. w18946. Claessens, S., M. A. Kose, and M. E. Terrones. 2012. “How Do Business and Financial Cycles Interact?” Journal of International Economics 87(1): 178-190. Csonto, B., A. Guerson, A. Myrvoda, E. Seworder, and S. Munoz. 2018. “Bank Network Analysis in the ECCU” IMF Working Papers 18/162, International Monetary Fund, Washington, DC. Davidson, R., and J. MacKinnon. 2006. “The Power of Bootstrap and Asymptotic Tests.” Journal of Econometrics 133(2): 421 –41. Dabla-Norris, E., M. Gradstein, and G. Inchauste. 2008. “What Causes Firms to Hide Output?” Journal of Development Economics 85 (1-2): 1-27. De Haan, R. Pleninger, J. Sturm. 2021. “Does Financial Development Reduce the Poverty Gap?” Social Indicators Research 161: 1-27. Elgin, C., and O. Öztunalı. 2012. “Shadow Economies around the World: Model-Based Estimates.” Working Papers 2012/05, Bogazici University. Elgin, C., and B. Uras. 2013. “Is Informality A Barrier to Financial Development?” 31 SERIEs 4(3): 309 –331. Elgin, C., M. A. Kose, F. Ohnsorge, and S. Yu. 2021. "Understanding Informality." Discussion Paper 16479, Center for Economic Policy Research, London, UK. Elgin, C., A. Elveren, and J. Bourgeois. 2021. “Informality, Inequality and Profit Rate.” Applied Economics Letters 28 (12): 1017 –20. Elgin, C. and E. Iyidost. 2021. “Velocity of Money and Shadow Economy.” Applied Economics Letters, DOI: 10.1080/13504851.2021.1961117 Elgin, C., C. C. Williams, G. Oz-Yalaman, and A. Yalaman. 2022. “Fiscal Stimulus Packages to COVID-19: The Role of Informality.” Journal of International Development, https://doi.org/10.1002/jid.3628. Farazi, S. 2014. "Informal Firms and Financial Inclusion: Status and Determinants", Policy Research Working Paper 6778, World Bank, Washington, DC. Fergusson. 2006. “Institutions for Financial Development: What Are They and Where Do They Come From?” Journal of Economic Surveys 20 (1): 27-70. Feyen, E. N. Fiess, A. C. Bertay, and I. Zuccardi Huertas. 2020. “Cross-Border Banking in Emdes: Trends, Scale, and Policy Implications.” Policy Research Working Paper 9393, World Bank, Washington, DC. Foley, F., and K. Manova, 2015. "International Trade, Multinational Activity, and Corporate Finance," Annual Review of Economics, Annual Reviews, 7(1): 119-146. Granda-Carvajal, C., and D. García-Callejas. 2022. “Informality, Tax Policy and the Business Cycle: Exploring the Links.” International Tax and Public Finance, https://doi.org/10.1007/s10797-021-09717-7. Head, K., and J. Ries. 2008. "FDI as An Outcome of the Market for Corporate Control: Theory and Evidence." Journal of International Economics 74(1): 2-20. Ihrig, J., and K. S. Moe. 2004. “Lurking in The Shadows: The Informal Sector and Government Policy.” Journal of Development Economics 73 (2): 541-557. International Monetary Fund. 2013. Central, Eastern and Southeastern Europe: Regional Economic Issues. International Monetary Fund, Washington, DC. International Monetary Fund. 2015. “Chapter 2 International Banking After the Crisis: Increasingly Local and Safer?” Global Financial Stability Report 2015/April. International Monetary Fund, Washington, DC. 32 Kanbur, R. 2017. “Informality: Causes, Consequences and Policy Responses.” Review of Development Economics 21 (4): 939-961. Kılınç, D., Ü. Seven, and H. Yetkiner. 2017. “Financial Development Convergence: New Evidence for the EU.” Central Bank Review 17(2): 47-54. Klein, M., and G. Olivei, 1999. "Capital Account Liberalization, Financial Depth, and Economic Growth," Working Papers 99-6, Federal Reserve Bank of Boston. La Porta, R., F. Lopez-de-Silanes, A. Shleifer and R. W. Vishny. 1997. “Legal Determinants of External Finance.” Journal of Finance 52: 1131 –1150. La Porta, R. , F. Lopez-de-Silanes, A. Shleifer and R. W. Vishny. 1998. "Law and Finance”, Journal of Political Economy 106: 1113 –1155. Levine, R. 1999. “Law, Finance, and Economic Growth.” Journal of Financial Intermediation 8: 36 –67. Levine, R., N. Loayza and T. Beck. 2000. “Financial Intermediation and Growth: Causality and Causes.” Journal of Monetary Economics 46: 31 –77. Levine, R. 2001. “International Financial Liberalization and Economic Growth.” Review of International Economics 9(4): 688-702. Levine, R. 2021. “Finance, Growth, and Inequality.” IMF Working Papers 2021/163, International Monetary Fund, Washington, DC. Madsen, J., M. Islam, and H. Doucouliagos. 2018. “Inequality, Financial Development and Economic Growth in the OECD, 1870 –2011. ” European Economic Review 101: 605 –24. Massenot, B., and S. Straub. 2016. “Informal Sector and Economic Development: The Credit Supply Channel.” Economic Inquiry 54(2): 1046-1067. Matousek, R., A. Rughoo, N. Sarantis, and A. G. Assaf. 2015. “Bank Performance and Convergence During the Financial Crisis: Evidence from the ‘Old’ European Union and Eurozone.” Journal of Banking and Finance 52: 208-216. Medina, L., and F. Schneider. 2018. “Shadow Economies Around the World: What Did We Learn Over the Last 20 Years?” IMF Working Papers 18/17, International Monetary Fund, Washington, DC. Medina, L., and F. Schneider. 2019. “Shedding Light on the Shadow Economy: A Global Database and the Interaction with the Official One.” CESifo Working Paper No. 7981. 33 Morón, E., E. Salgado, and C. Seminario. 2012. “Financial Dependence, Formal Credit and Firm Informality: Evidence from Peruvian Household Data.” Working Paper 288, Inter-American Development Bank, Washington, DC. Nurboja, B., and M. Košak. 2017. "Banking efficiency in South East Europe: Evidence for Financial Crises and the Gap between New EU members and Candidate Countries," Economic Systems 41(1): 122-138. Ohnsorge, F., and S. Yu. 2021. The Long Shadow of Informality: Challenges and Policies. Washington DC: World Bank. Ohnsorge, F., Y. Okawa, and S. Yu. 2021. “Chapter 4 Lagging Behind: Informality and Development” In The Long Shadow of Informality: Challenges and Policies , edited by F. Ohnsorge and S. Yu. Washington, DC: World Bank. Okawa, Y., and E. van Wincoop. 2012. “Gravity in International Finance.” Journal of International Economics 87(2): 205-2015. Olson, D. and T. Zoubi. 2017. “Convergence in Bank Performance for Commercial and Islamic Banks During and After the Global Financial Crisis.” The Quarterly Review of Economics and Finance 65: 71-87. Orsi, R., D. Raggi, and F. Turino. 2014. “Size, Trend, and Policy Implications of the Underground Economy.” Review of Economic Dynamics 17 (3): 417-436. Pappadá, F., and Y. Zylberberg. 2017. “Austerity and Tax Compliance.” European Economic Review 100: 506 –24. Perry, G., W. Maloney, O. Arias, P. Fajnzylber, A. Mason, and J. Saavedra-Chanduvi. 2007. Informality: Exit and Exclusion. Latin American and Caribbean Studies. Washington, DC: World Bank. Pleninger, R., J. Sturm. 2020. “The Effects of Economic Globalisation and Ethnic Fractionalisation on Redistribution.” World Development 130: 104945. Prado, M. 2011. “Government Policy in the Formal and Informal Sectors.” European Economic Review 55(8): 1120 –36. Presbitero, A.F. and R. Rebelotti. 2014. “Geographical Distance and Moral Hazard in Microcredit: Evidence from Colombia.” Journal of International Development 26(1): 91 –108. Quintin, E. 2008. “Contract Enforcement and the Size of the Informal Economy.” 34 Economic Theory 37 (3): 395-416. Raj, S. N., K. Sen, and V. Kathuria. 2014. “Does Banking Development Matter for New Firm Creation in the Informal Sector? Evidence from India.” Review of Development Finance 4(1): 38-49. Rashid, A. and M. Intartaglia. 2017. “Financial Development— Does It Lessen Poverty?” Journal of Economic Studies 44(1): 69 –86. Rewilak, J. 2017. “The Role of Financial Development in Poverty Reduction.” Review of Development Finance 7(2): 169 –176. Restrepo-Echavarria, P. 2014. “Macroeconomic Volatility: The Role of the Informal Economy.” European Economic Review 70: 454 –69. Rughoo, A. and N. Sarantis. 2014. “The Global Financial Crisis and Integration in European Retail Banking.” Journal of Banking and Finance 40: 28-41. Schneider, F., A. Buehn, and C. Montenegro. 2010. “New Estimates for the Shadow Economies All Over the World.” International Economic Journal 24 (4): 443-461 Schneider, F. and A. Buehn. 2016. “Estimating the Size of the Shadow Economy: Methods, Problems and Open Questions.” IZA Discussion Paper 9820, IZA-Institute of Labor Economics, Bonn, Germany. Sever, C., and E. Yucel, 2021. "Does Informality Hinder Financial Development Convergence?:" Working Papers 2021/02, Bogazici University, Department of Economics. Straub, S. 2005. “Informal Sector: The Credit Market Channel.” Journal of Development Economics 78 (2): 299-321. Svirydzenka, K. 2016. “Introducing a New Broad-based Index of Financial Development.” IMF Working Paper 16/5, International Monetary Fund, Washington, DC. Ulyssea, G. 2020. “Informality: Causes and Consequences for Development.” Annual Review of Economics 12: 527-546. Voghouei, H., M. Azali, and M. Jamali. 2011. “A Survey of the Determinants of Financial Development.” Asian-Pacific Economic Literature 25 (2): 1-20. World Bank. 2021. Global Economic Prospects. June. Washington, DC: World Bank. World Bank. 2022. The Global Findex Database 2021. Washington, DC: World Bank. Yermack, D. 2018. “FinTech in Sub-Saharan Africa: What Has Worked Well, and What 35 Hasn’t.” Working Paper 25007. Working Paper Series. National Bureau of Economic Research. 36