Sovereign external borrowing and multilateral lending: Dynamics and crises

Avellán, Leopoldo; Galindo Andrade, Arturo José; Lotti, Giulia Working Paper Sovereign external borrowing and multilateral lending: Dynamics and crises IDB Working Paper Series, No. IDB-WP-936 Provided in Cooperation with: Inter-American Development Bank (IDB), Washington, DC Suggested Citation: Avellán, Leopoldo; Galindo Andrade, Arturo José; Lotti, Giulia (2018) : Sovereign external borrowing and multilateral lending: Dynamics and crises, IDB Working Paper Series, No. IDB-WP-936, Inter-American Development Bank (IDB), Washington, DC, https://doi.org/10.18235/0001443 This Version is available at: https://hdl.handle.net/10419/208141 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/3.0/igo/legalcode Sovereign External Borrowing and Multilateral Lending: Dynamics and Crises Leopoldo Avellán A rturo J. Galindo Giulia Lotti IDB WORKING PAPER SERIES Nº IDB-WP-0936 November, 2018 Office of Strategic Planning and Development Effectiveness Inter-American Development Bank November, 2018 Sovereign External Borrowing and Multilateral Lending: Dynamics and Crises Leopoldo Avellán A rturo J. Galindo Giulia Lotti Cataloging-in-Publication data provided by the Inter-American Development Bank Felipe Herrera Library Leopoldo Avellán. Sovereign external borrowing and multilateral lending: dynamics and crises / Leopoldo A vellán, Arturo Galindo, Giulia Lotti. p. cm. — (IDB Working Paper ; 936) Includes bibliographic references. 1. Development banks. 2. Capital movements-Developing countries. 3. Debts, External-Developing countries. 4. Debts, Public-Developing countries. 5. Fiscal policyDeveloping countries. 6. Financial crises-Developing countries. I. Galindo, Arturo. II. Giulia Lotti. III. Inter-American Development Bank. Office of Strategic Planning and Development Effectiveness. IV. Title. V. Series. IDB-WP-936 1300 New York Ave NW, Washington DC 20577 Copyright © Inter-American Development Bank. This work is licensed under a Creative Commons IGO 3.0 AttributionNonCommercial-NoDerivatives (CC-IGO BY-NC-ND 3.0 IGO) license (http://creativecommons.org/licenses/by-nc-nd/3.0/igo/ legalcode) and may be reproduced with attribution to the IDB and for any non-commercial purpose, as provided below. No derivative work is allowed. Any dispute related to the use of the works of the IDB that cannot be settled amicably shall be submitted to arbitration pursuant to the UNCITRAL rules. The use of the IDB's name for any purpose other than for attribution, and the use of IDB's logo shall be subject to a separate written license agreement between the IDB and the user and is not authorized as part of this CC-IGO license. Following a peer review process, and with previous written consent by the Inter-American Development Bank (IDB), a revised version of this work may also be reproduced in any academic journal, including those indexed by the American Economic Association's EconLit, provided that the IDB is credited and that the author(s) receive no income from the publication. Therefore, the restriction to receive income from such publication shall only extend to the publication's author(s). With regard to such restriction, in case of any inconsistency between the Creative Commons IGO 3.0 Attribution-NonCommercial-NoDerivatives license and these statements, the latter shall prevail. Note that link provided above includes additional terms and conditions of the license. The opinions expressed in this publication are those of the authors and do not necessarily reflect the views of the Inter-American Development Bank, its Board of Directors, or the countries they represent. http://www.iadb.org 2018 5 Sovereign External Borrowing and Multilateral Lending: Dynamics and Crises Leopoldo Avellán 1 Arturo J. Galindo 2 Giulia Lotti 3 Abstract Fiscal policy and net capital inflows in developing countries are procyclical. A large amount of literature has examined this phenomenon and explored its consequences for aggregate fluctuations. Multilateral development banks (MDBs) are an important source of external finance for governments and hence play a key role in financing the execution of fiscal policy. The literature has found evidence that government borrowing from MDBs is countercyclical, but how are MDB flows related to fiscal policy? Does this relationship depend on whether the government is running a deficit or a surplus, or would this relationship change if the sovereign is going through a fiscal crisis? Do the differences in the scope and corporate structure between MDBs translate into different disbursement patterns? Beyond their impact to foster development in recipient countries, understanding the behavior of MDB flows is important to assess their contribution to macroeconomic stability. This paper answers these questions studying the co-movement of sovereign lending from MDBs with government expenditure and with private sovereign lending in different fiscal policy stances and during fiscal crises. The paper finds that multilateral sovereign lending is correlated with government expenditure, and that this correlation does not change if the government is running a surplus or a deficit. When considering total MDB lending, this comovement holds even in fiscal crises. Finally, the paper finds evidence of synchronization between multilateral development banks and the International Monetary Fund during fiscal crises. JEL codes: F21, F34, F41, F44, F53 Keywords: International government debt, capital flows, multilateral development banks 1 [email protected]; Inter-American Development Bank 2 [email protected]; Inter-American Development Bank 3 gl[email protected]; Inter-American Development Bank; and the Centre for Competitive Advantage in the Global Economy at the University of Warwick. We are grateful to Guillermo Vuletin, Erin Bautista and an anonymous referee for valuable comments. We also thank Andrés Gómez-Peña, Lina M. Botero and David Einhorn for editorial assistance. 5 1. Introduction Multilateral development banks (MDBs) were created after World War II to provide countries with financing directed toward development. Thanks to their high credit ratings, MDBs can borrow by issuing bonds on international capital markets at low costs. Given that their mandate is not to maximize profits, but rather sustain development activities, MDBs are subsequently able to lend to developing countries with only a narrow mark-up, even under grim domestic macroeconomic conditions or even when the government has no direct access to international financial markets. This unique financial model makes MDBs loans attractive to governments (Humphrey 2017). Indeed, their relevance as a reliable source of sovereign funding has increased over time, especially for lowand middle-income countries where their flows are larger than flows coming from private lenders. This rising role as development partners has made MDBs critical to the execution of fiscal policy and could even have positive macroeconomic externalities such as contributing to directly increase external liquidity, or indirectly, by catalyzing private capital inflows. This is particularly crucial nowadays given that globalization has increased markedly, with countries becoming more dependent on global financial conditions and hence more vulnerable to crises. The response to crises is then acquiring more relevance and MDBs can play a role in restoring the necessary confidence to attract global investors as fast as possible, minimizing output losses and preventing the spreading of crises to other countries. This paper studies the relationship between MDB flows and fiscal policy. It lies at the intersection of two strands of the literature. The first is the literature on the dynamics of international capital flows, spurred by the interest in financial crises over the last three decades. The large literature that looks at the cyclicality of international capital flows finds that, overall, net capital inflows are procyclical (Broner et al. 2013; Kaminsky et al. 2005). The procyclicality of capital flows can amplify business cycles, increasing consumption and spending in periods of capital flow bonanzas and imposing substantial adjustments when foreign capital no longer flows into the country (Levy Yeyati and Zuñiga 2015; De la Torre, Didier, and Pienknagura 2015). But when distinguishing by the lender of international government borrowing, there seems to be evidence of some heterogeneity in the behavior of capital flows. While private net lending to developing and emerging economies is procyclical (Galindo and Panizza 2018; Araujo et al. 2017; Levy Yeyati 2009; Dasgupta and Ratha 2000), there is scarce literature that looks at the cyclicality of multilateral institutions, emphasizing their countercyclical role(Galindo and Panizza 2018; Humphrey and Michaelowa 2011; Dasgupta and Ratha 2000), or that analyses the IMF lending responsiveness in crises (McDowell, 2017; Mody and Saravia, 2013). 4 Some have also analyzed the reaction of capital flows to crises in more detail. Broner et al. (2013) use a composite crisis indicator for banking, currency, and debt crises and find that in times of crises, capital flows decline. Dasgupta and Ratha (2000) test the response of net foreign direct 4 Specifically, Mody and Saravia (2013) find that the IMF responds more promptly to countries in severe crises, and McDowell (2017) when borrowers are more exposed to bond markets and short-term debt and the threat of capital flight is higher. 5 investment flows to balance of payments crises (1984–1989, 1995, and 1997), but do not find significant associations. Ratha (2005) examines cross-country data in 1980-2000 and finds that World Bank lending increased in the 1998-1999 Asian crisis. 5 Humphrey and Michaelowa (2011) examine the behavior of different institutions in years of global or regional economic crises (1982– 1983, 1995, 1998–1999, and 2009) 6 and find that in 1998–1999 and 2009 the World Bank and the Inter-American Development Bank increased their financial support. 7 Humphrey and Michaelowa (2013) study lending commitments by the same three MDBs but for a different set of countries and years: Bolivia, Colombia, Ecuador, Peru, and Venezuela in 1991-2010. 8 They find that the global financial crisis reduced World Bank lending, which they interpret as a supply restriction, while Inter-American Development Bank lending increased. Like the World Bank, Andean Development Corporation lending also decreased, most likely due to a spike in its own cost of funding. The second strand of the literature looks at the procyclicality of fiscal policy. A vast body of empirical literature has found that fiscal policy is procyclical for developing and middle-to-highincome countries, reinforcing the business cycle further, and harming the economy when conditions are already critical (Végh, Lederman, and Bennett 2017; Gerling et al. 2017; Bova, Carcenac, and Geurguil 2014; Frankel, Végh and Vuletin 2013; Reinhart and Reinhart 2008; Ilzetski and Végh 2008; Alesina, Campante, and Tavellini 2008; Talvi and Végh 2005; Kaminsky et al. 2005; Tornell and Lane 1999; Gavin and Perotti 1997; Cuddington 1989). Multiple factors can explain this behavior, including political economy distortions and the quality of governments (Avellán and Vuletin 2015; Ilzetski and Végh 2008; Talvi and Végh 2005; Tornell and Lane 1999). 9 Another cause can be that in bad times, when countries lack access to capital markets due to the procyclicality of capital flows, those countries are unable to adopt countercyclical fiscal policies. On the other hand, when international capital is plentiful, government spending increases excessively (Végh, Lederman, and Bennett 2017; Levy Yeyati and Zuñiga 2015; Frankel, Végh and Vuletin 2013; Kaminsky et al. 2005; Reinhart and Reinhart 2008). MDB lending has been analyzed mainly from the supply side (Humphrey 2014; Dreher et al. 2010; Dreher et al. 2009a, 2009b; Kilby 2006, 2011, among many others). 10 However, as Humphrey and Michaelowa (2013) point out, in recent decades countries have gained stronger financial and fiscal positions, some achieving high sovereign credit ratings and attracting investors more easily: ignoring the demand side in the analysis of MDB lending is no longer adequate. There is little research on the dynamics of MDB lending when a country’s demand for credit increases and we 5 The author finds that World Bank lending increases not only during crises, but more in general when debt service payment increases, and international reserves decline. 6 The authors also examine country crises defined based on the rankings for sovereign borrower risk in the annual Institutional Investor Index, on the overall fiscal balance of the central government as a share of GDP, and on international reserves divided by external short-term debt. 7 For the remaining years, the authors do not observe significant differences in multilateral lending. 8 Humphrey and Michaelowa (2011) focus instead on 10 countries in Latin America and the Caribbean in 1980-2009. 9 Political economy distortions include political pressures or rent-seeking activities that call for expansionary fiscal policy in good times (Avellán and Vuletin 2015; Ilzetski and Végh 2008; Talvi and Vegh 2005; Tornell and Lane 1999). The quality of governments is captured by regulatory quality, government effectiveness, control of corruption and rule of law in Avellán and Vuletin (2015), by the legal-political institutional infrastructure and fractionalization of power in Tornell and Lane (1999). 10 See Humphrey and Michaelowa (2013) for a more complete of papers. 5 aim to address this gap by empirically exploring the relationship between MDB lending to the government and public expenditure in line with different fiscal policy stances. Learning how MDB flows regularly co-move with the execution of fiscal policy is also important because it furthers understanding of how these institutions exacerbate or dampen capital flow cycles. Furthermore, assessing how the relationship of MDB flows and government spending evolves during fiscal crises can shed light on the role of MDBs as external liquidity providers. This paper begins with a discussion of the evolution of net flows for the public sector from MDBs and private creditors since the 1980s. For most countries, net flows from MDBs are larger and less volatile than net flows from the private sector, but there is some heterogeneity depending on the country’s income level. To measure how MDB lending 11 is systematically related to the demand of borrowing countries, we look at government expenditure. The analysis shows that when countries’ government expenditure increases, net borrowing from foreign creditors increases as well, and this is true for all creditors, confirming the correlation between fiscal policy and capital inflows found in the literature. No asymmetries are found when assessing whether this behavior differs when the government is in surplus or deficit. Given the co-movement of every creditor with the fiscal policy within a country, the analysis checks whether there are cases where external private capital markets and MDBs move in opposite directions. The analysis finds that while in the 1980s there was a positive relationship between net flows from MDBs and the private sector, this relationship became negative in the 1990s and non-significant in recent decades. The shift in the association between net flows might be related to the high frequency of fiscal crises in the 1990s. 12 To explore the issue above further, the analysis turns to the behavior of net flows around different types of fiscal crises. In particular, a distinction is made between fiscal crises due to credit events, exceptionally large official financing, implicit domestic public debt default, and loss of market confidence. The analysis finds that private creditors and MDBs behave differently in times of fiscal crises. While private net flows are negatively associated with credit crises, net flows from regional developments banks (RDBs) are positively correlated. In crises where there is exceptionally large official financing – that is, when countries ask for support from the International Monetary Fund (IMF) – the analysis finds that all MDBs increase their financial support, which illustrates strong synchronization among these institutions. While all the MDBs’ net flows are negatively correlated with implicit public defaults, net flows from private creditors do not change. In the case of crises driven by a loss of market confidence, private net flows decrease, but MDBs do not change their net flows. Finally, the analysis looks at whether the relationship between government borrowing and expenditure changes at times of country-specific fiscal crises. It is noted that overall the 11 Throughout the paper we define lending as net flows and not only approvals. 12 The analysis was unable to disentangle whether there is crowding out of private net flows due to net flows from multilateral development banks, as only simple correlations were examined. 5 association with total multilateral lending remains positive, with few exceptions when we disaggregate by the type of multilateral organization. The remainder of the paper is organized as follows. We provide a description of the data in Section 2. We discuss the empirical strategy in Section 3 and present the results in Section 4. We perform some robustness checks in Section 5 and give some final conclusions in Section 6. 2. Data To study the dynamics of international government lending, this paper focuses on net flows received by the government during the year, that is, disbursements minus principal repayments. The World Bank’s World Development Indicators are used as a source for net flows in current U.S. dollars from MDBs, 13 RDBs, 14 the World Bank, 15 other multilateral institutions, 16 and private creditors. 17 RDBs, the World Bank and other multilateral institutions are all part of the MDBs, but they are also analyzed separately to explore potential differences in sovereign lending. The sample includes 108 countries and totals 3,411 observations with non-missing net flows in the 1980–2015 period. High-income countries and countries that have fewer than 20 observations for GDP are excluded from the analysis. 18 The analysis also uses nominal GDP (in local currency units [LCUs] or in U.S. dollars) from the World Bank’s World Development Indicators. The countries in the sample are depicted in Figure 1. To investigate the relationship between net flows and fiscal policy, the analysis uses general government total expenditure and primary fiscal balance from the IMF’s World Economic Outlook. The former is defined as total expense and the net acquisition of nonfinancial assets (in LCUs), and the latter is defined as net lending/borrowing plus net interest payable/paid (interest expense minus interest revenue, also in LCUs). To analyze the behavior of capital flows around fiscal crises, the analysis uses the database of fiscal crises prepared by Gerling et al. (2017). Those authors define fiscal crises as periods of extreme funding difficulties that result in a disruption in the normal debt dynamics and in the 13 Public and publicly guaranteed multilateral loans include loans and credits from the World Bank, RDBs, and other multilateral and intergovernmental agencies. Excluded are loans from funds administered by an international organization on behalf of a single donor government. These are classified as loans from governments. 14 Net flows from RDBs include concessional and non-concessional financial flows. Concessional flows cover disbursements made through concessional lending facilities, and non-concessional financial flows cover the remaining flows. RDBs include the African Development Bank, Asian Development Bank, European Bank for Reconstruction and Development, and Inter-American Development Bank. 15 Net flows from the World Bank are the sum of net flows from the International Bank for Reconstruction and Development, the founding and largest member of the World Bank Group, and the International Development Association, the concessional loan window of the World Bank Group. 16 “Other multilateral organizations” is a residual category that includes the Caribbean Development Fund, Council of Europe, European Development Fund, Islamic Development Bank, Nordic Development Fund, and similar entities. 17 Public and publicly guaranteed debt from private creditors includes bonds that are either publicly issued or privately placed; commercial bank loans from private banks and other private financial institutions; and other private credits from manufacturers, exporters, and other suppliers of goods, as well as bank credits covered by a guarantee of an export credit agency. 18 They are excluded because the analysis will later calculate the trend of GDP, and it is important not to base the calculations on too few observations. Countries with fewer than 20 observations are Aruba, Afghanistan, Faeroe Islands, Iraq, Myanmar, Montenegro, Somalia, Serbia, São Tomé and Principe, and South Africa. 12 !",$ =&"+(")+*2-+",$ +*2-.+",$ ∗0-. +*21.+",$ ∗01. +*3-456+*3-.456∗0-. +*31.456∗ 01. +C*D-C456∗+",$ +*D-.C456∗+",$ ∗0-. +*D1.C456∗+",$ ∗01. +,",$ . (2b) Hence, β1L will capture the relationship between government expenditure and net flows in lowincome countries in fiscal surplus, the sum of βL + βLM will capture that relationship in lower-middleincome countries, and the sum of βL + βUM will capture that relationship in upper-middle-income countries. β1L+ β3L will measure the association between government expenditure and net flows in fiscal deficit in low-income countries, the sum of β1L+ β1LM + β3L+ β3LM will measure that association in lower-middle-income countries, and the sum of β1L+ β1UM +β3L+ β3UM will measure that association in upper-middle-income countries. To assess the presence of co-movements between private and net flows from MDBs, the following models are estimated for each subperiod (1980s, 1990s, 2000s, 2010s), for each country grouping, and for the whole sample, as in Broner et al. (2013): EFG",$ =&"+(")+*H;<IJ)K",$ +,",$ (3) H;<IJ)K",$ =&"+(")+*EFG",$ +,",$ , (4) where MDBi,t (Privatei,t) are net flows to the public sector from MDBs (private creditors) in country i in year t, scaled by trend GDP, de-meaned, and standardized by the country-level standard deviation of the net flows. The variable αi represents country fixed effects, γit represents countryspecific trends, and εi,t is the error term, clustered at the country level. The analysis then turns to assessing the dynamics of net flows from different agents in fiscal crises by estimating the following equation: !",$ =&"+(")+*><LM_M;N,",$ +,",$ , (5a) where yi,t are the different types of net flows scaled by trend GDP and standardized at the country level; fisc_crh,i,t are dummies for the years of fiscal crises of type h (credit events, exceptionally large official financing, implicit domestic public debt default, and loss of market confidence) in country i and year t, αi are country fixed effects, and γit are country trends; and εi,t is the error term, clustered at the country level. When heterogeneities in the relationship between net flows and fiscal crises by income group are allowed, equation (5a) becomes: !",$ =&"+(")+*-><LM_M;N,",$ +*-.><LM_M;N,",$ ∗0-. +*1.><LMOPN,",$ ∗01. +,",$. (5b) To check what happens to the relationship between net flows and government expenditures at times of crises and to the relationship between fiscal crises and net flows once government is controlled for, the following is estimated for fiscal crises: !",$ =&"+(")+*2+",$ +*3><LM_M;N,",$ +C*D><LM_M;N,",$ ∗+",$ +,",$ , (6a) where β1 will capture the association between net flows and government expenditure outside of fiscal crises, while the sum of β1 + β3 will measure the association between net flows and 13 government expenditure when the country experiences a fiscal crisis. β2 will capture the relationship between net flows and fiscal crises once government expenditure is controlled for. Finally, to allow for heterogeneities in the relationship between net flows and government expenditure in fiscal crises by income group, the following is estimated: !",$ =&"+(")+*2-+",$ + R *2S+",$ ∗0S 1. ST-. +*3-><LMOPU,V,W + R *3S><LMOPU,V,W ∗0S 1. ST-. +C*D-><LMOPU,V,W ∗+",$ + R *DS><LMOPU,V,W ∗+",$ ∗0S 1. ST-. +,",$.CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC(6b) C The next section presents the main empirical results, and the Appendix further describes the results by income group. 4. Results 4.1. Net Flows and Government Expenditure The analysis begins by showing the relationship between net flows from different agents and fiscal policy, specifically government expenditure, and by estimating equation (1a). As can be seen from Table 2, net flows from private creditors, MDBs, RDBs, the World Bank and other multilateral organizations all co-move with government expenditure: the more a country spends, the larger net flows it receives from external borrowers. Table 2. Net Flows to the Government and Government Expenditure Source: Authors’ calculations. Note: This table reports the correlations between net flows to the government from different agents and government expenditure (G). Both net flows and government expenditure are scaled by trend GDP, de-meaned, and standardized by country standard deviations. The sample period is from 1980 to 2015. Standard errors in parentheses are clustered at the country level; *** p<0.01, ** p<0.05, * p<0.1. FE: fixed effects; MDB: multilateral development banks; RDB: regional development banks; WB: World Bank; Others: other multilateral organizations. MDB RDB WB Others Private β 0.0641*** 0.0231** 0.0292*** 0.0243*** 0.0380*** se (0.020) (0.010) (0.009) (0.006) (0.011) Country FE Yes Yes Yes Yes Yes Country-trends Yes Yes Yes Yes Yes Average NFL 0.203 0.0218 0.0527 0.00833 0.0481 Average G 7.869 7.655 7.774 7.871 7.536 No. of countries 108 98 108 107 106 No. of observations 2,371 2,054 2,324 2,258 1,989 R-squared 0.307 0.226 0.326 0.150 0.130 14 The analysis then assesses if the relationship between net flows and government expenditure changes depending on whether or not government is running a primary fiscal deficit. The results of estimating equation (2a) are reported in Table 3. β1 is found to be significantly positive for all the net flows, indicating that when a government is in fiscal surplus and increases its government expenditure, net flows from every creditor increase (Table 3). β3 is never significantly different from zero, which implies that the behavior of foreign creditors does not change when the country is in deficit. Indeed, as can be seen from the sum of β1 + β3, the relationship of net flows and government expenditure stays positive for every creditor. Overall, Table 3 suggests that when a country increases its total expenditure, net flows from international creditors increase, irrespective of whether the country is in fiscal primary deficit or surplus. Table 3. Net Flows to the Government and Government Expenditure, Asymmetries Source: Authors’ calculations. Note: The table reports the correlations between net flows to the government from different agents and government expenditure (G), exploring different behaviors to positive/negative primary fiscal balances. Both net flows and government expenditure are scaled by trend GDP, de-meaned and standardized by country standard deviations. The sample period is from 1980 to 2015. Standard errors in parentheses are clustered at the country-level; *** p<0.01, ** p<0.05, * p<0.1. FE: fixed effects; MDB: multilateral development bank; RDB: regional development banks; WB: World Bank; Others: other multilateral organizations. 4.2 Net Flows from Multilateral Development Banks and Private Creditors Given that the relationship of private and multilateral lending with government expenditure is the same, it is important to explore further differences between the two types of lending. MDB RDB WB Others Private G (β1) 0.0577*** 0.0189* 0.0349*** 0.0245*** 0.0365*** se 1 (0.015) (0.010) (0.007) (0.007) (0.012) Fiscal Deficit (β2) 0.2398*** 0.1489 0.2060*** 0.1162* 0.2185*** se 2 (0.075) (0.090) (0.063) (0.064) (0.070) Fiscal Deficit # G (β3) -0.0019 -0.0041 -0.0051 -0.0028 -0.0071 se 3 (0.005) (0.004) (0.003) (0.003) (0.005) Country FE Yes Yes Yes Yes Yes Country-trends Yes Yes Yes Yes Yes Average NFL 0.164 0.00394 0.0318 -0.00783 0.0345 Average G 7.377 7.170 7.333 7.378 7.335 β1+ β3 0.056*** 0.0148* 0.0298*** 0.0216*** 0.0294** No. of countries 106 96 106 105 104 Observations 2,250 1,960 2,210 2,145 1,903 R-squared 0.314 0.232 0.340 0.145 0.144 15 Table 4 presents the correlations between net flows from MDBs and private creditors (equations 3-4). Despite some differences in magnitude by income group, overall, the different net flows were positively associated in the 1980s, but their relationship turned negative in the 1990s. In the most recent decades, however, no sign of co-movement between the two is found. Table 4. Correlations of Net Flows to the Government from Multilateral Development Banks and Private Creditors Source: Authors’ calculations. Note: The table shows the correlations between net flows to the government from MDBs and private creditors for upper-middleincome, lower-middle-income, and lower-income countries. Net flows are scaled by trend GDP, de-meaned, and standardized by country standard deviations. The sample period is from 1980 to 2015. Standard errors in parentheses are clustered at the countrylevel; *** p<0.01, ** p<0.05, * p<0.1. MDB: multilateral development bank; PRIV: private creditors. The question arises as to whether the negative correlations observed in the 1990s might be related to the higher frequency of fiscal crises during that decade. Indeed, when the number of years affected by fiscal crises is considered in the sample, a prevalence of fiscal crises during the 1990s is identified (Figure 2). Hence, the analysis now turns to the dynamics of net flows during fiscal crises. 4.3. Net Flows and Fiscal Crises When fiscal crises are considered regardless of their type (credit events, exceptionally large official financing, implicit domestic public debt default, and loss of market confidence), it can be seen that at times of crises there is a retrenchment in private creditors’ net flows, while MDBs, RDBs and the World Bank tend to increase net flows to the public sector (Table 5). Net flows from other multilateral organizations do not change during fiscal crises. Table 5. Net Capital Flows Dynamics in Fiscal Crises Source: Authors’ calculations. 1980s 1990s 2000s 2010s 1980s 1990s 2000s 2010s 1980s 1990s 2000s 2010s 1980s 1990s 2000s 2010s MDB = β PRIV (3) 0.020 -0.327 0.024 -0.021 0.114* -0.044 0.044 -0.063 0.006 -0.082* 0.029 0.004 0.049* -0.076* 0.032 -0.028 se (0.042) (0.198) (0.108) (0.039) (0.049) (0.045) (0.105) (0.078) (0.035) (0.043) (0.041) (0.078) (0.026) (0.032) (0.039) (0.044) PRIV = β MDB (4) 0.053 -0.063 0.014 -0.167 0.218* -0.038 0.054 -0.158 0.010 -0.115* 0.055 0.006 0.099* -0.068* 0.046 -0.064 se (0.103) (0.054) (0.058) (0.330) (0.067) (0.041) (0.132) (0.176) (0.063) (0.064) (0.075) (0.124) (0.047) (0.031) (0.053) (0.098) Country dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Country-trends Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No. of countries 23 22 21 19 31 41 38 36 30 40 41 41 84 103 100 96 No. of observations 185 162 151 86 258 323 326 202 258 335 371 228 701 820 848 516 R -squared (3) 0.419 0.325 0.402 0.462 0.404 0.357 0.384 0.265 0.360 0.419 0.163 0.256 0.388 0.370 0.299 0.275 R -squared (4) 0.333 0.219 0.069 0.366 0.440 0.295 0.321 0.280 0.350 0.318 0.248 0.337 0.378 0.303 0.254 0.321 Low-income Lower-middle-income Upper-middle-income All MDB RDB WB Others Private Crisis ( β )0.129** 0.127*** 0.111** -0.002 -0.231*** se (0.059) (0.048) (0.049) (0.049) (0.056) Country FE Yes Yes Yes Yes Yes Country-trends Yes Yes Yes Yes Yes No. of countries 108 98 108 107 106 No. of observations 3,411 2,967 3,323 3,233 2,944 R-squared 0.202 0.151 0.221 0.121 0.087 Fiscal Crises 16 Note: Net flows are scaled by trend GDP, de-meaned, and standardized by country standard deviations. The sample period is from 1980 to 2015. Standard errors in parentheses are clustered at the country-level; *** p<0.01, ** p<0.05, * p<0.1. MDB: multilateral development banks; RDB: regional development banks; WB: World Bank; Others: other multilateral organizations. But it is important to differentiate among types of fiscal crises, as not doing so might hide major heterogeneities. The fiscal crisis with the highest frequency throughout the period is related to credit events when countries experience a significant decrease of net flows from the private sector but an increase of support from RDBs. In events of exceptionally large official financing, that is, when the IMF gives its support, the MDBs join efforts to provide financing to governments, while this is not the case for other multilateral organizations (Table 6). The third type of crisis examined is implicit defaults. Implicit defaults signal that the government either resorted to seigniorage to finance the fiscal deficit and/or accumulated domestic arrears. When countries default implicitly, all the MDBs significantly decrease their lending, apart from other multilateral institutions. Table 6. Net Flows Dynamics in Fiscal Crises, by Type Source: Authors’ calculations. Note: Net flows are scaled by trend GDP, de-meaned, and standardized by country standard deviations. The sample period is from 1980 to 2015. Standard errors in parentheses are clustered at the country-level; *** p<0.01, ** p<0.05, * p<0.1. MDB: multilateral development banks; RDB: regional development banks; WB: World Bank; Others: other multilateral organizations. Finally, as expected, when a country experiences a loss of market confidence, private net flows decrease; but under these circumstances, MDBs do not change their net flows to the country affected. MDB RDB WB Others Private MDB RDB WB Others Private Crisis ( β )0.033 0.094* 0.048 -0.072 -0.141** 0.412*** 0.188*** 0.320*** 0.0705 0.023 se (0.059) (0.051) (0.056) (0.050) (0.061) (0.078) (0.068) (0.074) (0.069) (0.083) Country FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Country-trends Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No. of countries 108 98 108 107 106 108 98 108 107 106 No. of observations 3,411 2,967 3,323 3,233 2,944 3,411 2,967 3,323 3,233 2,944 R-squared 0.200 0.149 0.219 0.122 0.082 0.218 0.153 0.233 0.122 0.080 MDB RDB WB Others Private MDB RDB WB Others Private Crisis ( β )-0.580*** -0.327** -0.546*** 0.194 0.021 -0.004 0.043 -0.083 -0.039 -0.660*** se (0.159) (0.138) (0.164) (0.143) (0.170) (0.100) (0.088) (0.099) (0.077) (0.116) Country FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Country-trends Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No. of countries 108 98 108 107 106 70 67 70 69 69 No. of observations 3,409 2,965 3,321 3,231 2,942 1,387 1,268 1,357 1,311 1,310 R-squared 0.207 0.150 0.227 0.123 0.080 0.250 0.168 0.277 0.133 0.163 Credit Event Exceptionally Large Official Financing Implicit Domestic Public Default Loss of Market Confidence 17 So far, sovereign lending co-moves with government expenditure, irrespective of the primary fiscal balance and the type of creditor. At times of fiscal crises, on the other hand, the direction of net flows changes according to the creditor, with private investors decreasing their lending in crises due to credit events or a loss of market confidence. Moreover, net flows from MDBs do not seem to finance governments in fiscal crises, unless it is a joint effort with the IMF. However, the relationship between fiscal crises and sovereign borrowing might be capturing a relevant feature of the MDB business model, for which net flows co-move with government expenditure. For example, if government expenditure decreases during implicit defaults and MDB lending consequently decreases, the estimated β in Table 6 might be simply capturing the relationship between net flows and government expenditure. The next section will test whether this is the case. 4.4. Net Flows and Government Expenditure in Fiscal Crises This section studies what happens to the relationship between net flows and government expenditure in times of fiscal crises and to the relationship between net flows and fiscal crises once we control for government expenditure. Table 7 estimates equation (6) for the different types of fiscal crises. The co-movement between net flows to the public sector and government expenditure is confirmed when there are no fiscal crises: β1 is always positive and significantly different from zero. 21 Panel A shows that credit crises do not change the relationship previously found between government expenditure and net flows: β3 is not significantly different from zero and the sum of β1 + β3 is always positive and significantly different from zero. The relationship between RDB lending and credit crises is still positive once we control for government expenditure (β2), but not significantly different from zero anymore. The positive relationship between crises due to exceptionally large official financing (that is, when the IMF intervenes) and multilateral lending is maintained even when government expenditure is controlled for: in Panel B, β2 is positive and significantly different from zero for MDBs, RDBs, and the World Bank. Controlling for exceptionally large official financing, the co-movement between net flows and government expenditure is also maintained (β1). It is interesting to note that once this type of crisis hits, RDBs significantly reverse their behavior (β3) until they no longer co-move with government expenditure: the sum of β1 + β3 is no longer significantly different from zero. However, the positive relationship between net flows and government expenditure is maintained also in fiscal crises for all MDBs together, for the World Bank, for other multilateral organizations and for private creditors. It can also be seen in Panel C that once government expenditure is controlled for, net flows from MDBs, RDBs, and the World Bank do not retrench any further during implicit public domestic defaults (β2), that is, multilateral lending does not seem to decrease in crises either. Moreover, the positive association between government expenditure and MDBs in the aggregate remains significant (β1 + β3). The relationship between government expenditure and net flows from RDBs, 21 The relationship with MDBs and the World Bank under the loss of market confidence scenario is not shown because this fiscal crisis variable has many missing values and, together with the missing values of government expenditure, more than half of the sample is lost. Results are available upon request. 18 the World Bank and other multilateral organizations is still positive during this type of crisis, but no longer significant. The relationship between net flows and government expenditure in crises of loss of market confidence is not assessed because there are too few observations to draw any conclusions. Heterogeneous results by income group are presented in the Appendix. 19 Table 7. Net Flows to the Government and Government Expenditure during Fiscal Crises Source: Authors’ calculations. Note: The table shows the correlations between net flows and government expenditure (G) in times of fiscal crises. Net flows and G are scaled by trend GDP, de-meaned, and standardized by country standard deviations. The sample period is from 1980 to 2015. Standard errors in parentheses are clustered at the country-level; *** p<0.01, ** p<0.05, * p<0.1. FE: fixed effects; MDB: multilateral development banks; RDB: regional development banks; WB: World Bank; Others: other multilateral organizations. 20 5. Robustness Checks This section constructs net flows by scaling them by trend GDP and by standardizing, demeaning, and dividing by the country standard deviation. To check whether results are driven by the manipulation of the dependent variable, the analysis is performed again by scaling net flows by simply dividing by GDP. Trend GDP is used to isolate the cyclical component of GDP away from the analysis. However, even when net flows are re-scaled only by GDP, the same results are obtained. 6. Conclusion The literature on fiscal policy and international private capital flows has found evidence of procyclicality. This paper has explored whether multilateral sovereign lending moves together with private capital flows and fiscal policy, amplifying economic fluctuations. The analysis finds that sovereign borrowing tracks government expenditure, irrespective of the type of creditor. This result holds whether the government runs a deficit or not. These findings show that the co-movement of net flows with public expenditure holds regardless of the stance of fiscal policy. The study then explores sovereign net flows dynamics during fiscal crises. The evidence is that multilateral lending and private lending exhibit very different behavior during fiscal crises, with private creditors mostly decreasing their exposure, but MDBs only doing so during implicit domestic public defaults. However, once we control for government expenditure, results show that multilateral lending does no longer change in implicit defaults. Finally, countries going through a macroeconomic crisis might seek help from the IMF. The analysis finds that MDBs increase their net flows when there is an IMF program in place, and this finding holds also when controlling for the co-movement between net flows and government expenditure. This is evidence of coordination of MDBs and the IMF during fiscal crises, in line with the mandate of working as a system within the International Financial Architecture. 21 Appendix 1. Definition of Fiscal Crises A credit event occurs when the government reduces the present value of its debt owed to official or other creditors (de facto, mainly defaults on external debt). Exceptionally large official financing refers to any year under an IMF financial arrangement with access above 100 percent of quota and fiscal adjustment as a program objective. Financial support from the IMF is an alternative to outright default, usually for countries that are unable to pay their international bills and have associated balance of payment problems. Implicit domestic public debt default happens when countries default implicitly on domestic debt or their payment obligations by running domestic payment arrears or printing money to finance their budget (high inflation). The inflation rate threshold is 35 percent per year for advanced markets (the average haircut of their public debt) and small developing states. The threshold for emerging markets and low-income developing countries is 100 percent. Finally, a loss of market confidence crisis occurs in years of extreme market pressures, when either the country loses market access 22 or the price of market access surpasses a threshold of 1,000 basis points for the spreads, which is widely seen as market participants’ psychological barrier (Gerling et al., 2017). 2. Further Results by Income This section presents the relationship between net flows and government expenditure by income. Net Flows and Government Expenditure The co-movement between net flows from MDBs and expenditure does not change significantly by income group (Appendix Table 1). Net flows from the World Bank also always co-move with expenditure, but this co-movement is significantly stronger in low-income countries and loses significance in middle-income countries. When flows are divided by income, RDBs and private net flows no longer co-move with government expenditure in low-income and low-middle-income countries, probably due to the loss of power when adding interaction terms in the estimation. 22 The authors define loss of market access as the inability to “tap international capital markets on a sustained basis through the contracting of loans and/or issuance of securities across a range of maturities” (Gerling et al. 2017, p. 11). 28 Appendix Table 5 (continued). Net Flows and Government Expenditure in Times of Fiscal Crises, by Income Panel B MDB RDB WB Others Private G (β1L)0.0878*** 0.0062 0.0561*** 0.0293** 0.0380 (0.021) (0.011) (0.016) (0.012) (0.027) G # I_LM (β1LM)-0.0258 0.0293* -0.0404** -0.0132 0.0039 (0.037) (0.017) (0.020) (0.014) (0.030) G # I_UM (β1UM)-0.0391 -0.0028 -0.0241 0.0053 -0.0004 (0.031) (0.015) (0.019) (0.022) (0.035) Exceptionally large official financing (β2L) 0.0782 0.2864 -0.1196 -0.0260 0.0365 (0.186) (0.202) (0.152) (0.179) (0.122) Exceptionally large official financing # I_LM (β2LM) 0.2107 -0.0613 0.4564** 0.0083 0.0880 (0.233) (0.241) (0.217) (0.244) (0.197) Exceptionally large official financing # I_UM (β2UM) 0.5800** 0.1225 0.4945** 0.4119 -0.1233 (0.229) (0.285) (0.201) (0.283) (0.195) Exceptionally large official financing # G (β3L) 0.0231* -0.0051 0.0373*** -0.0055 -0.0098 (0.014) (0.011) (0.014) (0.009) (0.006) Exceptionally large official financing # G # I_LM (β3LM) -0.0286* -0.0062 -0.0449*** 0.0094 0.0003 (0.016) (0.012) (0.015) (0.011) (0.015) Exceptionally large official financing # G # I_UM (β3UM) -0.0349** -0.0037 -0.0468*** -0.0035 0.0069 (0.016) (0.013) (0.015) (0.013) (0.012) Observations 2,371 2,054 2,324 2,258 1,989 Number of countries 0.325 0.239 0.345 0.156 0.132 Country FE 108 98 108 107 106 Country-trends Yes Yes Yes Yes Yes Average NFLs Yes Yes Yes Yes Yes Average G 0.203 0.0218 0.0527 0.00833 0.0481 β1L+ β1LM 0.0619** 0.0355 0.0157** 0.0161** 0.0420*** β1L+ β1UM 0.0487** 0.00344* 0.0320 0.0346 0.0377** β1L+ β1LM + β3L+β3LM 0.0565* 0.0242 0.00813* 0.0200* 0.0324 β1L+ β1UM + β3L+β3UM 0.0369* -0.00530*** 0.0225 0.0256** 0.0347* R-squared 7.869 7.655 7.774 7.871 7.536 29 Appendix Table 6 (continued). Net Flows and Government Expenditure in Times of Fiscal Crises, by Income Source: Authors’ calculations. Note: The table shows the correlations between net flows and government expenditure (G) in times of fiscal crises. Net flows and G are scaled by trend GDP, de-meaned, and standardized by country standard deviations. The sample period is from 1980 to 2015. Standard errors in parentheses are clustered at the country-level; *** p<0.01, ** p<0.05, * p<0.1. FE: fixed effects; MDB: multilateral development banks; RDB: regional development banks; WB: World Bank; Others: other multilateral organizations. Panel C MDB RDB WB Others Private G (β1L)0.1100*** 0.0123 0.0795*** 0.0268** 0.0405 (0.019) (0.011) (0.014) (0.013) (0.029) G # I_LM (β1LM)-0.0478 0.0216 -0.0638*** -0.0097 0.0001 (0.036) (0.017) (0.018) (0.014) (0.031) G # I_UM (β1UM)-0.0581* -0.0064 -0.0461*** 0.0110 -0.0051 (0.031) (0.016) (0.017) (0.023) (0.037) Implicit domestic public default (β2L) 1.5546*** 0.1392 1.1463*** 0.3187* 0.6949 (0.291) (0.135) (0.195) (0.188) (0.438) Implicit domestic public default # I_LM (β2LM) -2.9058*** -0.9328 -3.0340*** 2.4870*** -0.0256 (0.392) (0.743) (0.508) (0.189) (0.526) Implicit domestic public default # I_UM (β2UM) -0.6621** 0.2393 -0.2391 -0.4398 -1.9397*** (0.330) (0.357) (0.334) (0.952) (0.447) Implicit domestic public default # G (β3L) -0.1272*** -0.0146 -0.0937*** -0.0290** -0.0527 (0.022) (0.013) (0.016) (0.014) (0.033) Implicit domestic public default # G # I_LM (β3LM) 0.2600*** 0.1128** -0.0856 1.8363*** -0.2049 (0.027) (0.045) (0.150) (0.022) (0.274) Implicit domestic public default # G # I_UM (β3UM) 0.1040*** 0.0009 0.0687*** 0.0385 0.1034** (0.023) (0.016) (0.017) (0.031) (0.041) Observations 2,369 2,052 2,322 2,256 1,987 Number of countries 0.315 0.229 0.339 0.153 0.136 Country FE 108 98 108 107 106 Country-trends Yes Yes Yes Yes Yes Average NFLs Yes Yes Yes Yes Yes Average G 0.204 0.0224 0.0531 0.00831 0.0483 β1L+ β1LM 0.0622** 0.0339*** 0.0157 0.0171*** 0.0406** β1L+ β1UM 0.0519** 0.00586** 0.0334*** 0.0378*** 0.0355 β1L+ β1LM + β3L+β3LM 0.195*** 0.132 -0.164 1.824 -0.217*** β1L+ β1UM + β3L+β3UM 0.0287 -0.00785 0.00839 0.0474* 0.0862 R-squared 7.875 7.662 7.779 7.877 7.542 30 Appendix Table 7. List of Countries in the Sample, by Income Group Source: Prepared by the authors. Low-income Lower-middle-income Upper-middle-income Burundi Armenia Pakistan Angola Kazakhstan Benin Bangladesh Philippines Albania Lebanon Burkina Faso Bolivia Papua New Guinea Argentina St. Lucia African Republic Côte d'Ivoire Sudan Azerbaijan Maldives Comoros Cameroon Solomon Islands Bulgaria Mexico Ethiopia Republic of Congo El Salvador Bosnia and Herzegovina FYR Macedonia Guinea Cabo Verde Syria Belarus Mauritius The Gambia Djibouti Tajikistan Belize Malaysia Guinea-Bissau Egypt Tonga Brazil Panama Haiti Ghana Tunisia China Peru Liberia Guatemala Ukraine Colombia Paraguay Madagascar Honduras Uzbekistan Costa Rica Russia Mali Indonesia Vietnam Dominica Thailand Mozambique India Vanuatu Dominican Republic Turkmenistan Malawi Kenya Samoa Algeria Turkey Niger Kyrgyz Republic Yemen Ecuador St. Vincent and the Grenadines Nepal Cambodia Zambia Fiji Venezuela Rwanda Lao P.D.R. Gabon Senegal Sri Lanka Georgia Sierra Leone Morocco Equatorial Guinea Chad Moldova Grenada Togo Mongolia Guyana Tanzania Mauritania Islamic Republic of Iran Uganda Nigeria Jamaica Zimbabwe Nicaragua Jordan 31 References Alesina, A., F.R. Campante, and G. Tabellini. 2008. Why Is Fiscal Policy Often Procyclical. Journal of the European Economic Association 6: 1006–036. Araujo, J.D., A.C. David, C. van Hombeeck, and C. Papageorgiou. 2017. Joining the Club? Procyclicality of Private Capital Inflows in Lower Income Developing Economies. Journal of International Money and Finance 70: 157–82. Avellán, L., and G. Vuletin. 2015. Fiscal Procyclicality and Output Forecast Errors. Journal of International Money and Finance 55: 193–204. Bova, M.E., N. Carcenac, and M.M. Guerguil. 2014. Fiscal Rules and the Procyclicality of Fiscal Policy in the Developing World. IMF Working Paper No. 14/122. International Monetary Fund, Washington, DC. Broner, F., T. Didier, A. Erce, and S.L. Schmukler. 2013. Gross Capital Flows: Dynamics and Crises. Journal of Monetary Economics 60(1): 113–33. Cuddington, J. 1989. Commodity Export Booms in Developing Countries. World Bank Research Observer 4: 143–65. Dasgupta, D., and D. Ratha. 2000. What Factors Appear to Drive Private Capital Flows to Developing Countries? And How Does Official Lending Respond? Policy Research Working Paper 2392. World Bank, Washington, DC. De la Torre, A., T. Didier, and S. Pienknagura. 2012. Latin America Copes with Volatility: The Dark Side of Globalization. World Bank, Washington, DC. Dreher, A., Klasen, S., Vreeland, J., & Werker, E. 2010. The costs of favoritism: Is politicallydriven aid less effective? e 4820. Dreher, A., Sturm, J., & Vreeland, J. 2009a. Development aid and international politics: Does membership on the UN Security Council influence World Bank decisions?. e 88: 1–18. Dreher, A., Sturm, J., & Vreeland, J. 2009b. Global horse trading: IMF loans for votes in the United Nations Security Council. European Economic Review 53: 742–757. Frankel, J.A., C.A. Végh, and G. Vuletin. 2013. On Graduation from Fiscal Procyclicality. Journal of Development Economics 100(1): 32–47. Galindo, A.J., and U. Panizza. 2018. The Cyclicality of International Public Sector Borrowing in Developing Countries: Does the Lender Matter? World Development 112: 119-135. Gavin, M., and R. Perotti. 1997. Fiscal Policy in Latin America. NBER Macroeconomics Annual 12: 11–61. 32 Gerling, M.K., M.P.A. Medas, M.T. Poghosyan, J. Farah-Yacoub, and Y. Xu. 2017. Fiscal Crises. IMF Working Paper No. 17/86. International Monetary Fund, Washington, DC. Humphrey, C., and K. Michaelowa. 2011. The Business of Development: Trends in Lending by Multilateral Development Banks to Latin America, 1980-2009. Proceedings of the German Development Economics Conference, Verein für Socialpolitik, Research Committee Development Economics. Humphrey, C. and Michaelowa, K., 2013. Shopping for development: Multilateral lending, shareholder composition and borrower preferences. World Development 44:142-155. Humphrey, C. 2014. The politics of loan pricing in multilateral development banks. Review of International Political Economy 21:3, 611-639. Humphrey, C., 2017. He who pays the piper calls the tune: Credit rating agencies and multilateral development banks. The Review of International Organizations 12:2, 281-306. Ilzetzki, E., and C.A. Végh. 2008. Procyclical Fiscal Policy in Developing Countries: Truth or Fiction? NBER Working Paper No. 14191. National Bureau of Economic Research, Cambridge, MA. Kaminsky, G.L., C.M. Reinhart, and C.A. Végh. 2005. When It Rains, It Pours: Procyclical Capital Flows and Macroeconomic Policies. NBER Macroeconomics Annual 2004, Volume 19. National Bureau of Economic Research, Cambridge, MA. Kilby, C. 2006. Donor influence in multilateral development banks: The case of the Asian development bank. Review of International Organizations 1(2): 173–195. Kilby, C. 2011. Informal influence in the Asian development bank. Review of International Organizations 6(3–4): 223–257. Levy Yeyati, E., 2009. Optimal Debt? On the Insurance Value of International Debt Flows to Developing Countries. Open Economies Review 20(4): 489–507. Levy Yeyati, E., and J. Zúñiga. 2015. Varieties of Capital Flows: What Do We Know?. CID Working Paper 296. Center for International Development at Harvard University. McDowell, D. 2017. Need for speed: The lending responsiveness of the IMF. The Review of International Organizations, Springer, 12(1): 39-73. Mody, A., and Saravia, D. 2013. The response speed of the international monetary fund. International Finance, 15: 189–211. Ratha, D. 2005. Demand for World Bank lending. Economic Systems. 29, 408–421. Reinhart, C.M., and V.R. Reinhart. 2008. Capital Flow Bonanzas: An Encompassing View of the Past and Present. NBER Working Paper No. 14321 National Bureau of Economic Research, Cambridge, MA. 33 Talvi, E., and C.A. Végh. 2005. Tax Base Variability and Procyclical Fiscal Policy in Developing Countries. Journal of Development Economics 78(1): 156–90. Tornell, A., and P.R. Lane. 1999. The Voracity Effect. American Economic Review 89(1): 22–46. Végh, C.A., D. Lederman, and F. Bennett. 2017. Leaning Against the Wind: Fiscal Policy in Latin America and the Caribbean in a Historical Perspective. LAC Semiannual Report (April). World Bank, Washington, DC.