## _wp0826 ## Source details **Canonical URL:** [_wp0826](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp0826.pdf) ## Other formats - [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp0826.pdf.md) - [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp0826.pdf.json) --- ### Research scope and data construction - Sample: - 19 Latin American countries (LACs) and 25 transition economies (TEs) - Period: 1989 to 2004 - Structural reform variables constructed and used: - Financial sector liberalization - Trade liberalization - Privatization - Distinction in measurement: - Reform efforts (policy changes) are distinguished from reform outcomes (financial development, trade openness). - Financial liberalization measures include indicators closer to outcomes and indicators closer to policy efforts. - Trade liberalization: trade openness treated as outcome; tariffs and tariff dispersion treated as reform efforts. ### Theoretical framing and motivation - FDI characterization and motivations: - FDI is an important component of capital flows in financial globalization and is among the least volatile forms of capital flows. - FDI serves as a conduit for technology spillovers that can positively affect growth. - Key research question: - What factors attract FDI, with special focus on structural reforms? - Considerations highlighted: - Structural reforms may signal lower investment risk and generate real benefits that alter parameters underlying foreign investment decisions. - Identical reforms may have heterogeneous effects across countries due to institutional differences, substitutability/complementarity of reforms, or sequencing issues. ### Methodology and empirical approach - Baseline econometric model: - Y_it = X_it β + η_i + γ_t + ε_it - Dependent variable Y_it: FDI as a ratio of GDP (used in log form in estimations). - X_it includes classical determinants (market size, initial income level, natural resource abundance, infrastructure, inflation), structural reforms (overall financial development, bank efficiency, trade liberalization, privatization), and institutional variables (quality of bureaucracy, executive constraints, rule of law). - η_i: country fixed effects; γ_t: time dummies. - Endogeneity and estimator: - Concern for reverse causality (e.g., between FDI and financial liberalization). - System-GMM estimator (Blundell and Bond (1998)) used to address endogeneity and include fixed effects. - Instruments: lagged levels of the dependent variable and lagged differences of strictly exogenous X variables where appropriate. - Instrument validity assessed with Sargan test; second-order serial correlation checked via Arellano-Bond statistics. ### Data and measurement: key variables and construction - FDI: - Source: International Financial Statistics (IFS); defined as capital inflows to acquire at least 10 percent of total ownership. - Financial sector reform measures: - Source: World Bank’s Financial Structure Dataset (February 2006 version) and Abiad, Detragiache, and Tressel (2007) for decomposition measures. - Overall financial development index components: - ratio of liquid liabilities to GDP; - ratio of private sector credit to GDP; - ratio of commercial bank assets to commercial bank assets plus central bank assets. - Bank efficiency index components: - ratio of overhead costs to total bank assets; - net interest margin (interest income minus interest expenses divided by total assets). - Financial liberalization (policy effort) index and components: competition, supervision, privatization_banks, securitiesmarkets, creditceilings, capitalflows, creditcontrols. - Trade liberalization: - Constructed from trade-weighted average tariff rates and tariff dispersion using World Bank-UNCTAD’s WITS system; normalized and averaged. - Privatization: - Based on World Bank privatization proceeds (Kikeri and Kolo, 2005). - Privatization proceeds defined as “all monetary receipts to the government resulting from partial and full divestitures (via asset sales or sale of shares), concessions, leases, and other arrangements.” - Exclusions: management contracts, new green-field investments, investments committed as part of concession agreements; voucher privatization programs not reflected. - A series excluding transactions with foreign buyers was also constructed; main results remain when excluding foreign buyers. - Institutions: - Polity IV: executive constraints (xconst), regime durability (durable). - ICRG: quality of bureaucracy (1 to 4), rule of law (1 to 6). - Other control variables: - Market size: Gross Domestic Product (in PPP terms) from WEO. - Level of development: real per capita GDP from WEO. - Macroeconomic stability: annual average inflation rates from WEO. - Natural resource dependence: percentage of oil and natural gas in total exports (WEO). - Infrastructure: number of main telephone lines (WDI); alternative: number of computers per 1000 people (used in robustness checks). - Normalization: - All reform indices normalized to 0–1 scale using sample maximum and minimum following Lora (1998). ### Principal empirical findings - Main result: - Robust positive relationship from structural reforms to FDI inflows across specifications. - Relative importance of reforms: - Financial sector reforms (bank efficiency, financial liberalization efforts) have a stronger effect on FDI than privatization and trade liberalization. - Privatization is also a significant positive determinant of FDI but generally smaller in magnitude than bank efficiency. - Paradox and interpretation: - “Paradox of finance”: multinational firms that are not clearly financially constrained invest more in countries with more relaxed financial constraints. - Proposed resolution: financial development and reform may be preconditions to maximizing spillover benefits through backward linkages to foreign investors. - Reform efforts (financial liberalization) appear to matter beyond current financial development (outcomes), suggesting signaling effects of policy changes. - Other robust correlates of FDI inflows: - More stable macroeconomic environment (lower inflation) - Higher levels of economic development (log(GDP per capita) positive) - Better infrastructure (telephone lines or computers) - Regional and sectoral differences: - Trade liberalization associated with higher FDI inflows in LACs but not in TEs. - Rule of law matters more for LACs; quality of bureaucracy matters more for TEs. - Financial liberalization index significant for countries with large foreign presence in the financial sector (FinFDI), but not for countries without such presence (NoFinFDI). - Robustness: - Results hold across different measures of FDI, reform indices, estimators, split samples, and checks for measurement error, endogeneity, and omitted variable bias. - Inclusion of institutional quality controls does not overturn main findings. - Granger-causality tests and IV estimations (reported by authors) indicate financial reforms drive FDI inflows but not vice versa; bank efficiency remains significant in IV results. ### Selected empirical estimates and key statistics (preserve reported figures exactly) - Baseline structural coefficients (examples from Table 1): - Bank efficiency: 4.485*** (column 1); 3.048** (column 2); 4.132** (column 3); 9.180*** (subgroup columns). - Privatization: 1.133*** (column 5); 1.410***; 1.512*** across specifications. - log(GDP): -0.167*** (column 1). - log(GDP per capita): 0.499*** (column 1). - log(inflation): -0.237*** (column 1). - log(telephone lines): 0.167** (ALL, Table 5); log (computer): 0.224*** (ALL alternative, Table 5). - Decomposition of financial liberalization (Table 2 examples): - creditceilings: 0.580* (Table 2). - supervision: 0.150** (Table 2). - securitiesmarkets: 0.152* (Table 2). - Financial liberalization index: 0.733* (ALL, Table 2 columns); -0.499 (LACs example); 2.009*** (TEs example). - Nonfinancial vs Financial FDI (Table 3 examples): - Financial liberalization index: 1.708*** (FinFDI, ALL); -0.6611 (NoFinFDI, ALL). - Overall fin. development: -5.977*** (FinFDI, TE in Table 3). - Robustness and institutions (Table 4 examples): - Durability: 0.00586** (Table 4, column 6). - Quality of bureaucracy: 0.225*** (Table 4, column 1). - Summary statistics (Appendix 1 selected): - log(FDI/GDP): Obs 582, Mean -4.73, Std.Dev. 1.12, Min -8.3, Max -2.2 - log(GDP): Obs 704, Mean 10.62, Std.Dev. 1.39, Min 8.3, Max 14.2 - log(GDP per worker): Obs 704, Mean 7.51, Std.Dev. 1.00, Min 3.4, Max 9.7 - log(inflation): Obs 563, Mean 2.70, Std.Dev. 1.65, Min -3.0, Max 8.9 - log(telephone): Obs 638, Mean 13.84, Std.Dev. 1.38, Min 10.6, Max 17.6 - log(fuel): Obs 523, Mean 1.39, Std.Dev. 1.94, Min -4.6, Max 4.5 - Executive constraints: Obs 543, Mean 5.44, Std.Dev. 1.80, Min 1, Max 7 - Qual. of Bureaucracy: Obs 498, Mean 1.93, Std.Dev. 0.86, Min 0, Max 4 - Rule of Law: Obs 498, Mean 3.51, Std.Dev. 1.19, Min 1, Max 6 - Overall fin. development: Obs 672, Mean 0.52, Std.Dev. 0.08, Min 0, Max 1 - Bank efficiency: Obs 672, Mean 0.78, Std.Dev. 0.09, Min 0, Max 1 - Trade liberalization: Obs 571, Mean 0.93, Std.Dev. 0.07, Min 0, Max 1 - Privatization: Obs 704, Mean 0.05, Std.Dev. 0.11, Min 0, Max 1 - Financial liberalization Index: Obs 473, Mean 0.58, Std.Dev. 0.23, Min 0, Max 1 ### Conclusions and suggested extensions - Synthesis: - Structural reforms, especially financial sector reforms and privatization, are robust positive determinants of FDI inflows in LACs and TEs for 1989–2004. - Reform efforts (financial liberalization) tend to have stronger effects on FDI than reform outcomes (overall financial development), implying signaling and policy-change effects matter. - Policy-relevant controls—macroeconomic stability, level of development, and infrastructure—are robust correlates of higher FDI. - Suggested extensions for future research: - Re-investigate long-term implications of FDI on growth while accounting for structural reforms (financial sector reform as a key enhancer). - Test whether the findings hold for developed countries and other developing regions (Africa, Middle East, Asia), noting substantial data collection would be required. - Broaden structural reform coverage to include labor market and product market liberalization, tax policy, and regulatory changes. *Source: _wp0826 (IMF working paper content provided).* ### References ............................................................................................................. ### _wp0826 - References ### Tables and Figures inventory - Tables: - 1. Determinants of FDI:Baseline Model: [Dependent variable = log(FDI/GDP)] — page 26 - 2. Determinants of FDI: Decomposition of Financial Liberalization [Dependent variable = log(FDI/GDP)] — page 27 - 3. Determinants of FDI: Nonfinancial vs Financial FDI [Dependent variable = log(FDI/GDP)] — page 28 - 4. Determinants of FDI:Robustness Check -Institutions [Dependent variable = log(FDI/GDP)] — page 29 - 5. Determinants of FDI:Robustness Check - Infrastructure [Dependent variable = log(FDI/GDP)] — page 30 - Figures: - 1. Foreign Direct Investment Inflows over GDP,Latin American and Transition Economies, 1989–2004 — page 23 - 2. Foreign Direct Investment Inflows per Worker, Latin American and Transition Economies, 1989–2004 — page 23 - 3. Overall Financial Reform Index — page 24 - 4. Index of Efficiency of Financial Intermediation — page 24 - 5. Trade Liberalization Index — page 25 - 6. Privatization Index — page 25 ### Research scope and data construction - Sample: - 19 Latin American countries (LACs) and 25 transition economies (TEs) - Period: 1989 to 2004 - Structural reform variables constructed and used: - Financial sector liberalization - Trade liberalization - Privatization - Distinction in measurement: - The paper differentiates reform efforts (policy changes) from reform outcomes (financial development, trade openness, etc.). - Example: For financial liberalization, measures range from indicators of financial development (outcomes) to measures closer to government policy changes (efforts). - Example: For trade liberalization, trade openness is treated as outcome; actual tariffs and their dispersion are treated as reform efforts. ### Theoretical framing and motivation - FDI is framed as: - An important component of capital flows in financial globalization. - One of the least volatile forms of capital flows, reducing vulnerability to sudden stops or reversals. - A conduit for technology spillovers that can positively affect growth. - Key research question: - What are the factors that attract FDI, with a special focus on the role of structural reforms? - Considerations highlighted: - Structural reforms may act as a positive signal to foreign investors by implying less investment risk. - Structural reforms may generate real benefits by affecting parameters underlying foreign investment decisions. - Identical reforms may have different effects across countries due to institutional differences, substitutability/complementarity of reforms, or sequencing issues. ### Principal empirical findings - Main result: - Strong relationship from structural reforms to FDI. - Relative importance of reforms: - Financial sector reforms have a stronger effect on FDI than privatization and trade liberalization. - Interpretation and puzzle: - Identified a “paradox of finance”: multinational firms that are not clearly financially constrained invest more in countries with the most relaxed financial constraints. - Proposed explanation: financial development may be a precondition to maximizing the benefits of spillovers via backward linkages to foreign investors. - Supporting literature: - Findings align with Alfaro and others (2004) and Prasad and others (2007): better developed financial markets and absorptive capacity (financial development) enhance the benefits of FDI. - The paper suggests financial reform is more important than financial development and more important than other structural reforms for attracting FDI. - Other robust correlates of FDI inflows: - More stable macroeconomic environment - Higher levels of economic development - Better infrastructure - Robustness: - Results robust to different measures of FDI, reform, different estimators, split samples, measurement error, endogeneity, and potential omitted variables bias. - Results hold after inclusion of institutional quality controls. ### Methodology and organization - Dataset: - A unique panel dataset constructed for comparability across countries and regions and consistent over time (1989–2004). - Paper structure: - Section II: recent developments in the two regions during the period - Section III: detailed construction of structural reform variables - Section IV: theoretical framework on determinants of FDI and relation to empirical literature - Section V: main econometric results and sensitivity checks *Source: _wp0826 - References (content from the referenced PDF).* ### Section VI concludes the paper and outlines directions for future research. ### Section VI concludes the paper and outlines directions for future research. ### II. FDI in Transition Economies and Latin America - Period covered: 1989–2004. - Transition from centrally planned to market economy began in nearly thirty countries with similar inherited institutions, initial conditions, and income levels. - Outcomes of reform diverged markedly across countries; transition experience characterized by an unexpectedly severe fall of per capita GDP and “a decade of disappointment.” - Comparison of FDI flows: - TEs received less FDI than LACs throughout the 1990s, reversing the trend only after 2002. - According to UNCTAD, LACs received about 10 percent of global FDI inflows between 1990 and 1994 while Asia received almost twice as much. - In the second half of the 1990s, the share of FDI into LACs increased to about 12 per cent; Asia accounted for 16 percent. - FDI inflows to LACs were disproportionately concentrated in a handful of countries such as Argentina, Brazil, Chile, and Mexico. - Macroeconomic context in Latin America: - 1990s marked by intense structural reform (Singh et al. 2005); macro stabilization programs largely successful except Brazil (stabilization succeeded only in 1994 with Plano Real). - Despite aggressive reforms (trade liberalization, privatization, labor and credit market liberalization), growth pay-offs were low and accompanied by severe financial crises. - Example statistic: average yearly per capita GDP growth rate was only 2.1 percent in the 1990s compared with more than 3 percent for the 1960s and 1970s. - Regional dynamics around crises: - East Asian crisis (1997) spilled over to Brazil and Russia; in Latin America FDI inflows halted and only recovered in per-worker terms by 2004, while TEs saw milder effects and recovery two years after the crisis. - A relatively small dip in 2002 in Latin America coincides with the Argentinean Crisis. ### III. Modeling the Determinants of FDI A. Brief overview of the literature - Main theoretical determinants: ownership advantages, location advantages, and benefits of internalization (Dunnings, 1974; Hymer, 1960; Caves, 1982). - Two strands of empirical work: - Determinants endogenous to firms (firm size, R&D intensity). - Location advantages of host countries (market size, labor cost). - Classical locational determinants include market size, relative factor prices (natural resources, labor cost, human capital), infrastructure, past stock of FDI, and economic and political risk. - Specific findings in the literature: - Wheeler and Mody (1992): infrastructure availability and past stock of foreign investment important. - Bevan and Estrin (2000): macroeconomic and political stability attract FDI. - Studies of TEs highlight EU membership prospects, institutions, and natural resource abundance as important (Resmini 2000; Bevan and Estrin 2000; Campos and Kinoshita 2003). - Latin America studies: De Gregorio (1992) finds FDI three to six times more efficient than total investment; Bengoa and Sanchez-Robles (2003) emphasize economic freedom, stability, and human capital; Trevino et al. (2002) find GDP level, privatization, and CPI inflation significant. B. Econometric model - Baseline specification: FDI as a function of classical determinants, structural reforms, and institutional quality. - Model presented (equation (1)): Y_it = X_it β + η_i + γ_t + ε_it - Y_it measured as FDI as a ratio of GDP in country i at year t. - X_it includes: - (i) classical factors: market size, initial income level, natural resource abundance, infrastructure, inflation; - (ii) structural reform variables: overall financial market development, bank efficiency, trade liberalization, privatization; - (iii) institutional variables: quality of bureaucracy, executive constraints, rule of law. - η_i: unobservable country-specific attributes; γ_t: time-specific effects (time dummies). - Concern for endogeneity of regressors (e.g., reverse causality between FDI and financial liberalization). - Estimation approach: - System-GMM estimator from Blundell and Bond (1998) chosen to address endogeneity and include fixed effects. - Advantages over Arellano-Bond difference-GMM: efficiency gains when time window is relatively short and ability to identify coefficients of time-invariant variables in the level equation. - Valid instruments: lagged levels of dependent variable Y_{t−s} where 2≥s and s = 2,...,4,3,T; if X_it strictly exogenous, lagged differences of X can be additional instruments. - Instrument validity checked by the Sargan test; second-order correlation assessed using Arellano-Bond statistics (asymptotically N(0,1)). - Note: small sample bias is a risk when number of observations is small relative to parameter estimates. - Motive-based hypotheses for FDI: - Market-seeking FDI: positive relation with host country market size (log of GDP). - Resource-seeking FDI: positive relation with natural resources (proxy: log of percentage of fuel and natural gas in total exports). - Other classical variable definitions: - Log GDP per capita: captures level of development/initial conditions. - Inflation: proxy for macroeconomic stability; expected negative coefficient on (log) inflation. - Infrastructure: (log of) number of main telephone lines used as proxy for overall infrastructure quality. - Institutional hypotheses: - Institutional quality variables: rule of law, quality of bureaucracy, executive constraints. - Theory implies institutions may increase or decrease FDI inflows depending on sector and FDI type (e.g., hold-up problem may make FDI more attractive under poor institutional protection). ### IV. Data and Measurement: FDI, Structural Reforms and Institutions - Sample coverage: 19 Latin American and 25 transition economies from 1989 to 2004. - LACs: Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Dominican Republic, Ecuador, El Salvador, Guatemala, Honduras, Jamaica, México, Nicaragua, Paraguay, Perú, Uruguay, Venezuela, and Trinidad and Tobago. - TEs: Albania, Armenia, Azerbaijan, Belarus, Bulgaria, Croatia, Czech Republic, Estonia, Georgia, Hungary, Kazakhstan, Kyrgyz Republic, Latvia, Lithuania, Macedonia, Moldova, Poland, Romania, Russia, Slovak Republic, Slovenia, Tajikistan, Turkmenistan, Ukraine, and Uzbekistan. A. Foreign Direct Investment - FDI data source: International Financial Statistics (IFS); defined as capital inflows to acquire a lasting management interest (lasting interest defined as acquiring at least 10 percent of total ownership). - Figures (1 and 2): throughout the 1990s average FDI inflows (over GDP and per worker) to LACs were substantially larger than to TEs, reversing only for two years in the analysis period. - Regional crisis responses: East Asia crisis (1997) affected both regions; Latin America saw halted FDI inflows in GDP terms with recovery in per-worker terms by 2004; transition economies recovered two years after the crisis. B. Measures of Structural Reform - Focused reform areas due to data comparability constraints: financial sector reform, privatization, and trade liberalization. - Emphasis on distinguishing reform inputs (efforts) from reform outcomes. - Financial sector reform measures: - Data source: World Bank’s Financial Structure Dataset (February 2006 version). - First indicator: overall financial development (three underlying variables): - ratio of liquid liabilities to GDP (liquid liabilities of the financial system: currency plus demand and interest-bearing liabilities of bank and non-bank intermediaries); - ratio to GDP of credit issued to the private sector by banks and other financial intermediaries; - ratio of commercial bank assets to the sum of commercial bank assets and central bank assets. - Two versions generated: (a) arithmetic average of normalized values; (b) based solely on ratio of commercial bank assets to commercial bank assets plus central bank assets. - Second indicator: efficiency of financial intermediation (two variables): - ratio of overhead costs to total bank assets; - net interest margin (bank interest income minus interest expenses, divided by total assets). - Because larger values indicate more inefficiency, normalization adjusted (subtract actual value from minimum) so larger normalized values indicate more efficient intermediation. - A third stock-market development index was constructed (stock market capitalization to GDP, total value traded to GDP, turnover ratio) but results similar to bank-based index and not reported for space. - Figures 3 and 4: overall financial development was relatively more intense in TEs than in LACs from 1989 to mid-1990s, reversing after that (except post-2001 years); efficiency index shows TEs catching up and surpassing LACs after mid-decade. - Normalization procedure for reform indices: - Following Lora (1998): transform series to 0–1 scale using sample maximum and minimum so that 1 indicates maximum in-sample level of reform. - Formula provided (equation (2)) used to make reform measures comparable across variables and over time. - Trade liberalization measure: - Constructed from two variables: average tariff rates and tariff dispersion. - Data source: World Bank-UNCTAD’s WITS system for about 6,000 HS-6 digit product groups to calculate trade-weighted average tariff and standard deviations yearly for each country. - Supplemented with Lora (2001) and Heritage Foundation’s Economic Freedom of the World project where UNCTAD data missing. - Normalized variables and took arithmetic average to generate overall trade reform indicator. - Figure 5: trade liberalization dynamics similar in both regions with almost continuous reduction of tariff levels and dispersion; LACs attained somewhat further trade openness than TEs. - Privatization measure: - Based on World Bank data on privatization proceeds (Kikeri and Kolo, 2005) covering privatization transactions between 1990 and [text cut off in source]. *Source: _wp0826 - Section VI concludes the paper and outlines directions for future research.* ### 2003. Privatization proceeds are defined as “all monetary receipts to the government resulting ### _wp0826 - 2003. Privatization proceeds are defined as “all monetary receipts to the government resulting ### Privatization data and index construction - Privatization proceeds are defined as “all monetary receipts to the government resulting from partial and full divestitures (via asset sales or sale of shares), concessions, leases, and other arrangements” (2005, p.2). - Exclusions: management contracts, new green-field investments, and investments committed by new private operators as part of concession agreements are excluded. - Voucher privatization programs are not reflected in these data because they “tend to generate little revenues for the government,” which biases the privatization index downward. The authors note that few Eastern European countries carried out extensive voucher privatization programs (notably, Czech Republic and Russia) and that privatization methods changed over time (for example, Russia moved from voucher privatization to loans-for-shares in the mid-1990s and then to a case-by-case program in the late 1990s). - Data source: World Bank privatization data (Available at http://rru.worldbank.org/Privatization/). The privatization index is constructed by aggregating privatization proceeds from the World Bank data for every country in the sample per year and applying the stated normalization. - The data set contains information on whether the buyer is foreign; the authors also construct a data series of government revenues from privatization that exclude transactions with foreign buyers and report that “All our main results below remain (including that for the role of privatization).” ### Empirical patterns of privatization (summary of Figure 6 observations) - (a) With the exception of 1996 and 1997, the index suggests privatization efforts were broadly comparable across the two regions (Latin America and the Caribbean (LACs) and Transition Economies (TEs)). - (b) Privatization efforts were more intense in Latin America than in Eastern Europe in the first half of the 1990s. - (c) Privatization efforts as measured by proceeds are more volatile in LACs than in TEs. - (d) There is a noticeable slow down of privatization activity after 1998 in the two regions but particularly in Latin America (Kikeri and Kolo, 2005). ### Institutions: data sources and measures - Rationale: Host country institutions influence investment decisions by affecting business operating conditions and non-economic costs (e.g., bribery, time lost in bureaucracy). Two main data sources are used to assess institutional aspects: Polity IV and the International Country Risk Guide (ICRG) (Available at http://www.cidcm.umd.edu/inscr/polity/ and http://www.icrgonline.com/, respectively). - Polity IV measures used: - extent of constraints on the executive power (xconst) - actual number of years the current regime has been in power (durable) - These measures have been used widely in the economics literature (e.g., Acemoglu and others, 2001). - ICRG measures used: - quality of the bureaucracy: a 1 to 4 indicator reflecting “autonomy from political pressure, institutional strength and quality of the bureaucracy” (higher ratings indicate a better bureaucracy). This measure is somewhat close to the corruption measure used by Wei (2000a, 2000b). High values imply good quality of bureaucracy and a lower cost for foreign investors. - rule of law: coded from 1 to 6 with higher ratings reflecting the effectiveness of the legal system (a higher score implies better legal institutions). The indicator is originally called “law and order” in ICRG; the "law" sub-component assesses the strength and impartiality of the legal system and the "order" sub-component assesses popular observance of the law (each sub-component equals half of the total). - Additional ICRG dimensions (e.g., corruption, government stability, political and economic risks) were assessed but not discussed in detail because results are similar to those reported. ### Institutional findings and regional comparisons - Executive constraints: executive constraints seem to have been more effective in LACs than in TEs throughout the period of analysis. LACs’ average is closer to the maximum category (which is seven) and indicates almost perfect parity of the executive with other legislative and judiciary power while the average for TEs is somewhat lower indicating only “substantial limitations on executive authority.” The rapid implementation of political reforms in TEs is visible in a brisk change between 1989 and 1990 for this region. - Rule of law: TEs score better than LACs throughout the period of analysis. In both regions there is a slight improvement in rule of law between 1992 and 1998 but values at the end of the period return to their 1989 levels. - Quality of the bureaucracy (ICRG): TEs seem to have better bureaucracies than LACs before 1997; from 1998 onwards these indexes overlap. This pattern arguably reflects successful public sector reforms in Latin America and cuts in government expenditures in transition economies. - Overall lesson: different dimensions of institutional development behave differently not only across regions but also over time. *Source: _wp0826 - 2003. Privatization proceeds are defined as “all monetary receipts to the government resulting* ### Appendix Tables 1 and 2 report summary statistics and the correlation matrix, respectively. As it can be ### _wp0826 - Appendix Tables 1 and 2 report summary statistics and the correlation matrix, respectively. As it can be ### Other control variables - Market size proxy: Gross Domestic Product (in PPP terms) from IMF World Economic Outlook (WEO). - Level of development proxy: real per capita GDP from WEO. - Macroeconomic stability proxy: annual average inflation rates from WEO. Expectation: lower average inflation attracts FDI. - Natural resources dependence proxy: percentage of oil and natural gas in total exports (WEO). Note: resource abundance may attract FDI to resource industries and possibly divert manufacturing FDI. - Infrastructure proxy: availability of main telephone lines from World Bank World Development Indicators (WDI). Alternative considered: percentage of paved roads (not used) and number of computers per 1000 people (used in robustness checks). ### V. The results — Objectives and approach - Objective: identify factors explaining distribution of FDI across Latin American (LAC) and transition economies (TEs) for 1989–2004, explicitly introducing structural reforms as determinants of FDI. - Estimation: system-GMM (Blundell and Bond (1998)). Dependent variable = log(FDI/GDP). - Instruments: log(GDP), log(GDP per capita), log(inflation), log(telephone lines), log(fuel), quality of bureaucracy, executive constraints, rule of law, overall financial development, bank efficiency, trade liberalization, and privatization — used in lagged levels for differences and in first differences for level equation. - Specification tests reported: Sargan test and SOC (second-order correlation) test. ### A. Baseline results — key empirical findings - Classical determinants: - Higher log(GDP per capita) positively related to FDI inflows. - Lower log(inflation) positively related to FDI inflows (negative coefficients on log(inflation)). - log(telephone lines) (infrastructure) positively related to FDI inflows. - log(fuel) (resource abundance) positive and significant in some specifications. - log(GDP) (market size) often fails to be statistically significant and carries a negative sign, implying affiliate production may be export-platform rather than market-seeking. - Structural reforms: - Bank efficiency and privatization are significant structural reform determinants of FDI. - Relative magnitudes (column 5 example): bank efficiency is about three times as large as privatization. - Bank efficiency (Table 1): 4.485*** (column 1); 3.048** (column 2); 4.132** (column 3); 9.180*** (columns for subgroups). - Privatization (Table 1): 1.133*** (column 5); 1.410***; 1.512*** across specifications. - Institutions: - Quality of bureaucracy consistently positive impact on FDI inflows for all countries; significance diminishes in region-wise regressions. - Executive constraints and rule of law coefficients positive but often statistically insignificant. - Rule of law important for LACs (column 6); quality of bureaucracy most important for TEs (column 7). - Trade liberalization: - Progress in trade liberalization is associated with higher FDI inflows only in LACs, not in TEs. - Robustness of instruments: - Sargan and SOC tests indicate instruments are valid across regressions reported. ### B. Decomposition of the financial sector reform — findings - Purpose: distinguish reform efforts from reform outcomes (current level of financial development). - Additional financial reform variables and composite indices (Abiad, Detragiache, and Tressel (2007) data) used; sample size reduced. - Components associated with higher FDI inflows when included separately (Table 2): supervision, creditceilings, securitiesmarkets. - creditceilings coefficient: 0.580* (Table 2, column shown). - supervision coefficient: 0.150** (Table 2). - securitiesmarkets coefficient: 0.152* (Table 2). - Financial liberalization index (composite): - Positive and significant for ALL and for TEs in columns 8–10: Financial liberalization index coefficients include 0.733* (ALL), -0.499, and 2.009*** (TE reported in Table 2 columns). - For LACs, financial liberalization index is not significant; for LACs privatization is a stronger driver. - Interpretation: reform efforts (financial liberalization) encourage more FDI inflows even after controlling for overall financial development; reform signals matter beyond current depth of financial markets. - Reverse causality addressed: - Split-sample by foreign presence in financial sector (share of foreign ownership > 20% classified as financialFDI). - Table 3: financial liberalization index significant only for countries with large foreign presence in financial sector (FinFDI columns). - Table 3 example coefficients: Financial liberalization index 1.708*** (FinFDI, ALL); -0.6611 (NoFinFDI, ALL). - Overall fin. development sometimes insignificant; when significant negative for financial subgroup: -5.977*** (FinFDI, TE in Table 3). - Additional checks: Granger-causality tests and IV estimation reported (results available from authors): tests show financial reforms drive FDI inflows but not vice versa; bank efficiency remains significant in IV results. ### C. Robustness checks — institutions and infrastructure - Institutions one-at-a-time (Table 4): - Bureaucracy and executive constraints closely related to FDI inflows. - Durability significant in some specifications: Durability coefficient 0.00586** (Table 4, column 6). - Rule of law remains often statistically insignificant in pooled regressions; regionally, rule of law matters for LACs and quality of bureaucracy for TEs. - Infrastructure alternative (Table 5): - Replacing telephone lines with log (computer) preserves main result: infrastructure provision attracts FDI. - Table 5 coefficients: log(telephone lines) 0.167** (ALL); log (computer) 0.224*** (ALL alternative). - Financial liberalization index remains significant in infrastructure robustness: 0.578** (ALL); 0.503* (ALL with computers). - Overall robustness conclusion: - Main findings on structural reforms and institutions withstand checks: banking sector efficiency and privatization are consistently important; quality of bureaucracy and rule of law matter regionally (TEs and LACs, respectively). ### Key regression summary statistics (selected) - Table 1 (Baseline): - Observations: 355, 315, 315, 315, 298, 182, 116 (across columns 1–7). - Number of ccode: 33, 33, 33, 33, 11, 5, 16 (reported). - Sample significant coefficients (examples): log(GDP) -0.167*** (col 1); log(GDP per capita) 0.499*** (col 1); log(inflation) -0.237*** (col 1). - Table 2 (Decomposition of Financial Liberalization): - Observations: 245, 245, 245, 245, 253, 113, 245, 245, 145, 100 (columns 1–10). - Number of ccode: 24, 24, 24, 24, 25, 13, 24, 24, 11, 13. - Selected coefficients: Privatization 1.010*** to 1.671*** across columns; Overall fin. development -0.672 to -4.761*** in some columns. - Table 3 (Nonfinancial vs. Financial FDI): - Observations: 72, 173, 73, 61, 93, 66 (columns across subsamples). - Number of ccode: 17, 22, 8, 11, 9, 11. - Selected coefficients: Privatization 1.930** (NoFinFDI ALL); 1.074** (FinFDI ALL); Financial liberalization index 1.708*** (FinFDI ALL). - Table 4 (Institutions robustness): - Observations: 298 (columns 1–3), 314, 244, 244 (columns 4–6). - Number of ccode: 31, 31, 31, 35, 30, 30. - Selected coefficients: Bank efficiency 2.703* to 5.676***; Qual. of Bureaucracy 0.225*** (column 1). - Table 5 (Infrastructure robustness): - Observations: 245; 210 (columns). - Number of ccode: 24; 21. - Selected coefficients: log (computer) 0.224***; Privatization 1.069*** and 1.208***. ### VI. Conclusions — synthesis and extensions - Main empirical conclusion: robust positive relationship from structural reforms to FDI; financial sector reforms and privatization exert stronger effects than trade liberalization. - Reform efforts vs. outcomes: reform efforts (e.g., financial liberalization) tend to have more powerful effects on FDI inflows than reform outcomes measured by overall financial development. - Robust determinant set: financial reform, privatization, level of development, and quality of infrastructure robust to different measures, estimators, split samples, and endogeneity/omitted-variable checks. - Suggested extensions: - Re-investigate long-term implications of FDI on growth accounting for structural reforms (financial sector reform as a key enhancer of foreign capital benefits). - Assess whether findings hold for developed countries and other developing regions (Africa, Middle East, Asia) — would require substantial data collection. - Broaden structural reform coverage to labor market and product market liberalization, tax policy, and regulatory changes. *Source: IMF working paper appendix and results (content unit provided).* ### References ### _wp0826 - References ### References (selected) - Abed, G.T. and Davoodi, H.R., 2000, "Corruption, Structural Reforms, and Economic Performance in the Transition Economies," IMF Working Paper 00/132 - Abiad, A., E. Detragiache, and T. Tressel, 2007, “A New Database of Financial Reforms,” unpublished, International Monetary Fund. - Acemoglu, D. and J. Robinson, 2001, “A Theory of Political Transitions,” American Economic Review, 91, pp. 938-63. - Acemoglu, D., S. Johnson, J. Robinson, and Y Thaicharoen, 2003, “Institutional Causes, Macroeconomic Symptoms: Volatility, Crises and Growth,” Journal of Monetary Economics, 50, pp. 49-123. - Alfaro, L., A. Chanda, S. Kalemli-Ozcan, and S. Sayek, 2004, “FDI and economic growth: the role of local financial markets,” Journal of International Economics, 64, pp. 89–112. - Antras, P., 2003, “Firms, Contracts, and Trade Structure,” Quarterly Journal of Economics, 118, No. 4, pp. 1375-1418. - Arellano, M., and S. Bond, 1991, “Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equation,” Review of Economic Studies, 58, pp. 277–97. - Beck, Thorsten, Asli Demirgüç-Kunt and Ross Levine, 2000, "A New Database on Financial Development and Structure," World Bank Economic Review, 14, pp. 597-605. - Borensztein, E., J. DeGregorio, and J-W. Lee, 1998, “How Does Foreign Direct Investment Affect Economic Growth?” Journal of International Economics, 45, pp. 115–35. - Campos, N. F., and F. Coricelli, 2002, “Growth in Transition: What We Know, What We Don’t, and What We Should,” Journal of Economic Literature, Vol. XL (3), pp. 793–836. - De Melo, M., C. Denizer, A. Gelb, and S. Tenev, 1997, “Circumstance and Choice: The Role of Initial Conditions Policies in Transition Economies,” World Bank WPS 1866. - Detragiache, E., P. Gupta, and T. Tressel, 2005, “Finance in Lower-Income Countries: An Empirical Exploration,” IMF Working Paper 05/167. - La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R. Vishny, 1998, “Law and Finance,” Journal of Political Economy, 106(6), pp. 1113–55. - Lucas, Jr., R.E., 1988, “The Mechanics of Economic Development,” Journal of Monetary Economics, 22, pp. 3–42. - Sachs, J., and A. Warner, 1995, “Economic Reform and the Process of Global Integration” Brookings Papers on Economic Activity, 1, pp. 1-118. - Wei, S. J., 2000a, “How Taxing is Corruption on International Investors?” Review of Economics and Statistics, 82 (1), pp. 1–11. - Wheeler, D., and A. Mody, 1992, “International Investment Location Decisions: The Case of U.S. firms,” Journal of International Economics, 33, pp. 57–76. (References list continues in source.) ### Appendix 1 — Summary Statistics (table key figures) - Variable observations and summary statistics: - log(FDI/GDP): Obs 582, Mean -4.73, Std.Dev. 1.12, Min -8.3, Max -2.2 - log(GDP): Obs 704, Mean 10.62, Std.Dev. 1.39, Min 8.3, Max 14.2 - log(GDP per worker): Obs 704, Mean 7.51, Std.Dev. 1.00, Min 3.4, Max 9.7 - log(inflation): Obs 563, Mean 2.70, Std.Dev. 1.65, Min -3.0, Max 8.9 - log(telephone): Obs 638, Mean 13.84, Std.Dev. 1.38, Min 10.6, Max 17.6 - log(fuel): Obs 523, Mean 1.39, Std.Dev. 1.94, Min -4.6, Max 4.5 - Executive constraints: Obs 543, Mean 5.44, Std.Dev. 1.80, Min 1, Max 7 - Qual. of Bureaucracy: Obs 498, Mean 1.93, Std.Dev. 0.86, Min 0, Max 4 - Rule of Law: Obs 498, Mean 3.51, Std.Dev. 1.19, Min 1, Max 6 - Overall fin. development: Obs 672, Mean 0.52, Std.Dev. 0.08, Min 0, Max 1 - Bank efficiency: Obs 672, Mean 0.78, Std.Dev. 0.09, Min 0, Max 1 - Trade liberalization: Obs 571, Mean 0.93, Std.Dev. 0.07, Min 0, Max 1 - Privatization: Obs 704, Mean 0.05, Std.Dev. 0.11, Min 0, Max 1 - Financial liberalization Index: Obs 473, Mean 0.58, Std.Dev. 0.23, Min 0, Max 1 - competition: Obs 478, Mean 2.28, Std.Dev. 0.97, Min 0, Max 3 - supervision: Obs 478, Mean 1.02, Std.Dev. 0.90, Min 0, Max 3 - privatization_banks: Obs 478, Mean 1.23, Std.Dev. 1.18, Min 0, Max 3 - securitiesmarkets: Obs 478, Mean 1.62, Std.Dev. 1.00, Min 0, Max 3 - creditceilings: Obs 286, Mean 0.81, Std.Dev. 0.39, Min 0, Max 1 - capitalflows: Obs 476, Mean 1.72, Std.Dev. 1.06, Min 0, Max 3 - creditcontrols: Obs 474, Mean 1.87, Std.Dev. 1.00, Min 0, Max 3 ### Appendix 2 — Data Description and Sources (variable definitions and sources) - log(FDI/GDP): Log of FDI to GDP ratio — Source: WEO - log(GDP): Log of GDP — Source: WEO - log(GDP per worker): Log of GDP per capita — Source: WEO - log(inflation): Log of annual inflation — Source: WEO - log(telephone): Log of number of main telephone lines — Source: World Development Indicators - log(computer): Log of number of computers — Source: World Development Indicators - log(fuel): Log of fuel exports as % of total exports — Source: World Development Indicators - Executive constraints: Executive constraints, operational de facto independence of chief executives — Source: polity IV - Qual. of Bureaucracy: Source: ICRG - Rule of Law: Source: ICRG - Overall fin. development / Bank efficiency: Index of bank efficiency based on the ratio of overhead cost to total bank assets and net interest margin — Source: Author's calculations - Trade liberalization: Arithmetic average of normalized average tariff rate and tariff dispersion — Source: Author's calculations - Privatization: Government's privatization proceeds — Source: Kikeri and Kolo (2005) - Corruption: Corruption indicator. Component of the political risk rating, counting for 6 of the 100 points. — Source: ICRG - Political risk: Political risk rating measuring political stability from 0 (most risky) to 100 (least risky). Composed of 12 weighed variables — Source: ICRG - Durability: The number of years the dictator is in power — Source: polity IV - competition, supervision, privatization_bank, securitiesmarkets, creditceilings, capitalflows, Financial liberalization index: Source: Detragiache and others (forthcoming) or Detragiache and others (2007) as indicated - creditceilings: Ceilings on expansion of bank credit imposed by the CB; =0 if yes, =1 if no. — Source: Detragiache and others (2007) - capitalflows: Capital account restrictions: (i) ex rate unified?, (ii) restrict capital inflow?, (iii) restrict capital outflows? — Source: Detragiache and others (2007) - Additional measurement notes (from source text): - Overall financial development index based on the ratio of liquid liabilities to GDP, the ratio of private sector credit to GDP, and the ratio of commercial bank assets to the total bank assets — Author's calculations - Privatization of banks index coding: =3(FL) if no state banks exist, =2(LL) if most banks are privately owned, =1(PR) if major banks are still state-owned, =0(FR) if major banks are all state owned - Policies to enhance banking sector supervision include adoption of capital adequacy ratio based on Basle standard, supervisory agency independence, effective supervision, and coverage of all financial institutions - Policies to lower banking sector entry barriers include foreign banks allowed, new domestic banks allowed, restrictions on branching, breadth of bank activities - Securities market development measures: a country took measures to develop securities markets; stock market open to foreigners - Aggregate governance indicator measuring the quality of contract enforcement, police, courts, and likelihood of crime and violence ### Appendix 3 — Correlation Matrix (selected pairwise correlations) - fdi/gdp with: - fdi/wk: 0.8048 - gdp_ppp_: -0.0898 - gdp_pc_: 0.2622 - inflation: -0.1222 - telephone: -0.0801 - fuel: 0.0414 - bureacracy: 0.3672 - xconst: 0.2952 - ruleoflaw: 0.2135 - fd1b: 0.1158 - fd2: 0.1982 - TradeLib: 0.1354 - privat: 0.2395 - fdi/wk with: - gdp_ppp_: 0.0048 - gdp_pc_: 0.5852 - inflation: -0.1001 - telephone: -0.0015 - bureacracy: 0.5057 - xconst: 0.2810 - ruleoflaw: 0.3965 - fd1b: 0.2134 - fd2: 0.2359 - TradeLib: 0.1154 - privat: 0.2334 - gdp_ppp_ with: - telephone: 0.9384 - gdp_pc_: 0.1506 - bureacracy: 0.1363 - telephone with: - gdp_pc_: 0.0771 - inflation: 0.0279 - bureacracy with: - xconst: 0.4742 - gdp_pc_: 0.5564 - privat: 0.2060 - fd1b with: - fd2: 0.0713 - gdp_pc_: 0.3136 - TradeLib with: - fd2: 0.1318 - privat: -0.0048 *Source: _wp0826 - References (source PDF content provided).* --- _Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp0826.pdf_