## 4. Drivers of Capital Flows and the Role of the Investor Base in Latin America

## Source details

**Canonical URL:** [4. Drivers of Capital Flows and the Role of the Investor Base in Latin America](https://www.imf.org/-/media/files/publications/reo/whd/2017/may/wreo0517-chp4.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/reo/whd/2017/may/wreo0517-chp4.pdf.md)
- [Structured JSON version](/-/media/files/publications/reo/whd/2017/may/wreo0517-chp4.pdf.json)

---

### Common cyclical behavior of capital flows
- Gross inflows and gross outflows exhibit a strong positive correlation over time; changes tend to be in the same direction and partially offset each other.
- Gross inflows are generally significantly larger than gross outflows; net inflows tend to be driven by gross inflows.
- Capital inflow cycles are synchronized across countries, particularly for the LA5 (Brazil, Chile, Colombia, Mexico, Peru).
- LA7 (Argentina, Brazil, Chile, Colombia, Mexico, Peru, Uruguay) patterns:
  - Brazil and Mexico inflows tend to follow each other and the LA7 median closely.
  - Chile, Colombia, and Peru broadly exhibit similar cyclical behavior; Chile’s inflows relative to economy size have been larger on average.
  - Argentina and Uruguay show higher volatility; both had a significant fall in inflows in the early 2000s, Uruguay rebounded strongly while Argentina remained subdued.

### Structural differences across countries
- Cross-country averages of gross inflows (since 2000):
  - Albania, Bulgaria, Kazakhstan, Vietnam: gross inflows averaged more than 12 percent of GDP.
  - Argentina, Egypt, Indonesia: gross inflows averaged less than 2 percent of GDP.
  - Chile (LA7): gross inflows averaged 7½ percent of GDP.
  - Argentina (LA7): gross gross inflows averaged ¾ percent of GDP.
- Variation decomposition (emerging markets example):
  - 43 percent of variation in gross inflows attributable to variation “within” countries (over time).
  - 36 percent attributable to variation “between” countries (across countries).
- The relative importance of “between” variation is driven largely by other emerging market regions; cyclical variables may play a more important role in Latin America.

### Subcomponents of capital flows: FDI, portfolio, other
- LA7 averages since 2000:
  - FDI inflows averaged 3¾ percent of GDP.
  - Portfolio inflows averaged 1¼ percent of GDP.
  - Other investment inflows averaged ¼ percent of GDP.
- Characteristics:
  - FDI: larger on average; more “between” country variation (FDI inflows: 19 percent “within”, 46 percent “between” for emerging markets).
  - Portfolio: more volatile; more “within” country variation (portfolio inflows: 61 percent “within”, 18 percent “between”).
- Country heterogeneity:
  - Gross FDI inflows averaged more than 9 percent of GDP in Bulgaria, Hungary, Vietnam since 2000.
  - Gross FDI inflows averaged less than 1½ percent of GDP in Egypt and Indonesia.
  - Within LA7, Chile and Uruguay have been the largest recipients of both gross FDI and portfolio inflows since 2000; Argentina recorded the lowest amounts for both.

### Drivers of capital flows: global “push” and country “pull” factors (model setup)
- Core global variables: VIX, global output growth, U.S. interest rates, index of global commodity prices.
- Country-specific variables:
  - Cyclical: domestic growth minus global growth; domestic interest rates minus U.S. interest rates.
  - Structural: governance, regulatory quality, business climate, political risk, corporate tax rate, credit rating (included one at a time due to multicollinearity).
- Estimation framework:
  - Cyclical model: Yi,t = a + b1 Gt + b2 Ci,t + mi + ei,t, estimated with quarterly data 2000:Q1–2016:Q2 using country fixed effects; country-specific variables lagged.
  - Structural model: mi = g + ρ Zi,t + ui,t, where mi are fixed effects from the cyclical model and Zi,t are structural factors.
- Fixed effects mitigate omitted country-level variation; structural factors introduced one-at-a-time because of multicollinearity.

### Empirical findings — macro regressions and robustness
- Gross inflows drive net inflows; gross and net regressions deliver similar results.
- Global commodity prices:
  - Strong positive association with higher capital inflows to all emerging markets.
  - Capital flow cyclicality closely follows the global commodity price cycle.
  - Robust to alternative indices, individual series (oil, copper, iron ore), and commodity price residuals controlling for global factors.
  - Replacing the commodity price index with Gruss (2014) commodity terms-of-trade yields similar results.
- Global growth:
  - Positive relationship with capital inflows; statistically significant mainly for Latin American economies.
- VIX and U.S. interest rates:
  - In full specifications, VIX and U.S. interest rates do not appear strongly associated with capital flows once commodity prices are included.
  - VIX is significant when included alone but loses significance when commodity prices are added.
- Gross outflows:
  - Global and country-specific factors explain less of gross outflow variation; R-squared in gross outflow regressions is less than half that of gross inflow regressions.

### Global cyclical drivers — cross-correlations and principal components
- Cross-correlation strengths:
  - Principal component of capital inflows in EMs with Global commodity prices (log) = 0.82***.
  - Principal component of stock prices in EMs (log) with Global commodity prices (log) = 0.93***.
  - Principal component of capital inflows in EMs with principal component of stock prices in EMs (log) = 0.86***.
  - U.S. nominal effective exchange rate (log) correlation with principal component of capital inflows in EMs = 20.75***.
  - vIX (log) correlation with principal component of capital inflows in EMs = 20.38***.
- VIX shows mixed results; often not statistically significant.
- G7 real GDP growth (y/y) generally positively associated with capital inflows for Latin America in many specifications.

### Country-specific structural determinants (slow-moving factors)
- Institutional and political improvements:
  - Improvements in government effectiveness, regulatory quality, control of corruption, or rule of law from current LA7 levels to average advanced-economy levels → sustained increase in inflows of about 1½–1¾ percent of GDP.
  - Improving Brazil, Colombia, Peru indicators to Chile levels → raise inflows by about 1½–2 percent of GDP.
  - For Argentina, improving to Chile levels could raise inflows by up to 3 percent of GDP.
- Lower corporate tax rates and higher credit ratings are associated with higher inflows; credit ratings reflect policy credibility and asset eligibility for institutional investors.

### Robustness: Sala-i-Martin–style model-selection exercise
- N = 15 potential explanatory variables → >32,000 possible models; histograms of coefficients used to assess robustness.
- Global commodity prices:
  - Latin America: all estimated coefficients positive and significant across all model variants.
  - Other emerging markets: two-thirds of estimated coefficients positive and significant.
- Table 4.5 (selected shares of significant coefficients, 2000–16):
  - Global commodity prices: LA5 = 100 (expected sign); OEM = 67 (expected sign).
  - G7 real GDP growth: LA5 = 95 (expected sign); OEM = 1 (expected sign).
  - Real GDP growth differential: LA5 = 79 (expected sign); OEM = 2 (expected sign).
  - vIX: LA5 = 8 (expected sign) and 8 (opposite sign); OEM = 0 (expected sign) and 26 (opposite sign).

### Historical decompositions and regional contrasts
- Two episodes decomposed:
  1. Global financial crisis: most of the fall in inflows to Latin America driven by global factors—mainly global output growth and commodity prices; for other emerging markets, domestic growth differential accounted for >25 percent of explained variation.
  2. End of the commodity super-cycle: sharp decline in commodity prices was the largest contributor to reduced inflows across EMs; domestic growth differential accounted for:
     - 19 percent of variation in inflows for other emerging markets.
     - 9 percent of variation in inflows for Latin America.
- Once commodity prices are accounted for, Latin American inflows are not strongly linked to domestic growth, unlike other emerging markets where domestic growth remains important.

### IPVAR methodology and impulse-response quantitative findings
- IPVAR used to assess how responses of capital flows to external shocks depend on investor base and domestic financial-system characteristics; shocks include VIX, global commodity prices, G7 real GDP growth, and an identified U.S. monetary shock (Osorio Buitron and Vesperoni (2015)).
- Key impulse-response magnitudes:
  - A sustained 20 percent increase in commodity prices → average increase in capital inflows of almost 2 percent of GDP to Latin America and other EMs.
  - An increase in the VIX of some 10 points → fall in capital inflows of about the same magnitude (about the decline implied by the 20 percent commodity shock in absolute terms).
  - An unanticipated U.S. monetary tightening of about 50 basis points → fall in capital inflows in EMs of close to 1 percent of GDP.
- IPVAR benchmark: shocks to VIX or U.S. monetary policy rates lead to declines in inflows; positive shocks to commodity prices and global growth boost inflows.

### Heterogeneity by investor base and domestic market characteristics
- Higher foreign participation in domestic debt markets → greater sensitivity of capital inflows to external shocks.
- Larger stock market capitalization and deeper domestic financial markets → lower sensitivity of capital flows to external shocks.
- Larger share of pension funds → decreases sensitivity of inflows to global factors (pension funds favor long-term stable investments).
- Fixed exchange rate regimes → greater sensitivity to external shocks, particularly U.S. monetary shocks, than flexible regimes.
- Higher capital account openness → inflows more vulnerable to external conditions.
- Countries combining deeper markets and higher foreign participation tend to have:
  - Lower inflation and inflation volatility
  - Higher domestic growth and lower growth volatility
  - Lower sovereign spreads
  - More favorable credit ratings
  - Better governance indicators

### Representative regression estimates (selected coefficients, 2000–16)
- Total gross capital inflows (column (1)):
  - vIX (log): 20.830 (0.813)
  - G7 real GDP growth (y/y): 20.776* (0.406)
  - U.S. short-term interest rates: 0.341 (0.383)
  - Real GDP growth differential (lagged): 0.435*** (0.107)
  - Demand component of commodity price (log): 0.252*** (0.046)
  - Supply component of commodity price (log): 2.989 (2.343)
  - Observations: 1,312; R-squared: 0.135; Number of countries: 22
- Portfolio inflows (column (3)):
  - vIX (log): 21.295*** (0.257)
  - G7 real GDP growth: 0.122* (0.067)
  - U.S. short-term interest rates: 20.159* (0.088)
  - Demand component of commodity price (log): 0.050** (0.022)
  - Observations: 1,312; R-squared: 0.168; Number of countries: 22
- Other inflows (column (4)):
  - Demand component of commodity price (log): 0.119*** (0.020)
  - Observations: 281; R-squared: 0.169; Number of countries: 22
- Note: Robust standard errors in parentheses; significance: * p < 0.10; ** p < 0.05; *** p < 0.01.

### Micro evidence: commodity prices and firm-level investment (LA6 firms)
- Firm-level investment comoves with country-specific commodity export price changes for publicly listed firms in Argentina, Brazil, Chile, Colombia, Mexico, Peru.
- Investment growth tracks export-related commodity price growth for median agricultural and mining firms; noncommodity firms show similar patterns—spillovers beyond direct commodity sectors.
- Microeconometric specification (preserved notation) links I/K(-1) to Q, profits, leverage, interest expense, debt ratios, commodity export prices P^x_i,t, and interaction P^x_i,t × X_i.
- Key micro estimation results (Table 4.2.1):
  - a (Q_i,t): 1.56*** (0.310) and 1.55*** (0.310)
  - λ (D_i,t / E_i,t): 24.43*** (0.706) and 24.42*** (0.705)
  - k (P^x_i,t): 0.05*** (0.013) and 0.04*** (0.014)
  - m (P^x_i,t × X_i): restricted m = 0 in column (1); 0.04 (0.040) in column (2) (not statistically significant)
  - Number of observations: 4,651 (column (1)); 4,650 (column (2))
  - Number of firms: 763 (column (1)); 762 (column (2))
- Inference:
  - Tobin’s q (Q) strongly positive and significant.
  - k > 0 indicates positive association between commodity prices and investment.
  - Lack of significance for m implies commodity-price-driven investment increases occur for both commodity and noncommodity firms (spillovers).

### Annex macro findings (selected coefficients and fits)
- Core gross inflows specification, 2000–16 (Annex Table 4.1) — global commodity price (log) coefficients:
  - LA5 core model: 4.182*** (0.387)
  - LA7: 4.458** (1.354)
  - OEM: 4.918*** (1.261)
  - EMs: 4.434*** (0.973)
- G7 real GDP growth (y/y): 0.509** (0.171) for LA5 in first stage.
- First-stage observations and R-squared (first stage):
  - Observations: 322 (LA5), 440 (LA7), 872 (OEM), 1,312 (EMs).
  - R-squared (first stage): 0.480 (LA5), 0.385 (LA7), 0.141 (OEM), 0.151 (EMs).
- Disaggregated inflows (Annex Table 4.3) — global commodity price (log) for first stage:
  - Gross FDI inflows: 1.493* (0.594) for LA5; 1.508** (0.527) for LA7; 2.100 (1.196) for OEM.
  - Gross portfolio inflows: 1.206* (0.446) for LA5; 2.140** (0.711) for LA7; 1.252*** (0.390) for OEM.
- Institutional variables (second stage) often positive and highly significant (examples):
  - Government effectiveness: 2.214*** (0.087) for LA5 (FDI); 0.400*** (0.025) for LA5 (portfolio).
  - Regulatory quality: 2.670*** (0.064) for LA5 (FDI); 0.521*** (0.021) for LA5 (portfolio).
  - Corporate tax rate: negative and significant (examples: 20.219*** (0.007) for LA5 FDI; 20.047*** (0.002) for LA5 portfolio).

### Policy implications and recommendations
- Monetary policy trade-offs:
  - Commodity-price-driven inflows raise growth and inflationary pressures (suggesting tighter policy) while also inducing exchange rate appreciation and higher inflows, complicating the use of monetary tightening alone to forestall overheating.
- Recommended policy mix:
  - Use fiscal policy and exchange rate flexibility to complement monetary policy in response to surges in inflows.
  - Implement effective macroprudential policies to support monetary objectives and contain financial-stability side effects.
- Financial deepening and investor base:
  - Promote deeper domestic financial markets and stable domestic intermediaries (pension funds, insurance companies) to reduce inflow sensitivity to external shocks.
  - Opening capital accounts to foreign participation to gain market depth is associated, on average, with better macroeconomic performance than remaining relatively closed; pace of financial opening should align with financial stability considerations.
- Exchange rate policy:
  - Greater exchange rate flexibility helps reduce sensitivity of capital flows to adverse external shocks.

*Source: Chapter 4, "Drivers of Capital Flows and the Role of the Investor Base in Latin America," Regional Economic Outlook: Western Hemisphere, International Monetary Fund, April 2017.*

### 4. Drivers of Capital Flows and the Role of

### 4. Drivers of Capital Flows and the Role of the Investor Base in Latin America

### Common cyclical behavior of capital flows
- Gross inflows and gross outflows exhibit a strong positive correlation over time: gross outflows tend to increase when gross inflows increase, and fall when gross inflows fall, so changes tend to be in the same direction and partially offset each other.
- Overall magnitudes of gross inflows are generally significantly larger than gross outflows; net inflows tend to be driven by gross inflows.
- Capital inflow cycles are synchronized across countries, particularly for the LA5 countries (Brazil, Chile, Colombia, Mexico, Peru).
- For the LA7 countries (Argentina, Brazil, Chile, Colombia, Mexico, Peru, Uruguay):
  - Brazil and Mexico inflows tend to follow each other and the LA7 median closely.
  - Chile, Colombia, and Peru broadly exhibit similar cyclical behavior; Chile’s inflows relative to economy size have been larger on average than other countries in the region.
  - Argentina and Uruguay show higher volatility; both experienced a significant fall in inflows in the early 2000s, but Uruguay rebounded strongly while Argentina remained subdued.

### Structural differences across countries
- Average levels of capital inflows (relative to GDP) vary substantially across countries:
  - Since 2000, gross inflows averaged more than 12 percent of GDP in countries such as Albania, Bulgaria, Kazakhstan, and Vietnam.
  - Gross inflows averaged less than 2 percent of GDP in countries such as Argentina, Egypt, and Indonesia.
  - For the LA7 countries, Chile received the most gross capital inflows, averaging 7½ percent of GDP since 2000, whereas Argentina’s gross capital inflows averaged ¾ percent of GDP over that period.
- Variation decomposition (Table 4.2): for emerging markets, a substantial share of variation is both “within” countries (over time) and “between” countries (across countries):
  - Example aggregate shares reported: 43 percent of variation in gross inflows attributable to variation “within” countries; 36 percent attributable to variation “between” countries (text discussion).
- The relative importance of “between” country variation is driven largely by other emerging market regions in the sample; cyclical variables may play a more important role in explaining capital flows in Latin America than in other emerging markets.

### Subcomponents of capital flows (FDI, portfolio, other)
- Capital flows can be decomposed into foreign direct investment (FDI), portfolio flows, and “other investment” flows (mainly bank loans and deposits).
- In LA7 countries, average inflows since 2000:
  - FDI inflows averaged 3¾ percent of GDP.
  - Portfolio inflows averaged 1¼ percent of GDP.
  - Other investment inflows averaged ¼ percent of GDP.
- Characteristics of subcomponents:
  - FDI inflows are larger and exhibit more “between” country variation: for FDI inflows to emerging markets, 19 percent of variation is attributable to variation “within” countries and 46 percent to variation “between” countries.
  - Portfolio inflows are relatively more volatile and exhibit more “within” country variation: for portfolio inflows, 61 percent of variation is attributable to variation “within” countries and 18 percent to variation “between” countries.
- Country heterogeneity in subcomponents:
  - Gross FDI inflows averaged more than 9 percent of GDP in Bulgaria, Hungary, and Vietnam since 2000.
  - Gross FDI inflows averaged less than 1½ percent of GDP in Egypt and Indonesia.
  - Within LA7, Chile and Uruguay have been the largest recipients of both gross FDI and portfolio inflows since 2000; Argentina recorded the lowest amounts for both.

### Drivers of capital flows: global “push” and country “pull” factors
- Core global variables used in the regression model: VIX (Chicago Board Options Exchange Volatility Index) as global risk aversion proxy, global output growth, U.S. interest rates, and an index of global commodity prices.
- Country-specific variables separated into:
  - Cyclical factors: differential between domestic growth and global growth; differential between domestic interest rates and U.S. interest rates.
  - Structural factors: governance, regulatory quality, business climate, political risk, corporate tax rate, credit rating, and other measures (included one at a time due to multicollinearity).
- Model setup:
  - Cyclical model: Yi,t = a + b1 Gt + b2 Ci,t + mi + ei,t, estimated with quarterly data from 2000:Q1 to 2016:Q2 using country fixed effects; country-specific variables included with a lag to mitigate endogeneity.
  - Structural model: mi = g + ρ Zi,t + ui,t, where mi are the fixed effects from the cyclical model and Zi,t are country-specific structural factors.
- Estimation notes:
  - Fixed effects are used to minimize bias from omitted country-level variation.
  - Structural factors are included one at a time given high multicollinearity.

### Empirical findings from the regression analysis
- Gross inflows and net inflows regressions deliver broadly similar results, consistent with gross inflows driving net inflows.
- Global commodity prices:
  - Higher global commodity prices are strongly associated with higher capital inflows to all emerging markets (Annex Table 4.1).
  - The cyclicality of capital flows tends to follow the global commodity price cycle closely (Figure 4.7).
  - This result is robust to alternative commodity price indices, to using individual series (oil, copper, iron ore), and to using the residual from regressing commodity prices on global factors (including global growth and growth in China).
  - Replacing the commodity price index with the commodity terms-of-trade measure of Gruss (2014) yields similar observations.
- Global growth:
  - An increase in global growth appears to lead to higher capital inflows, although the relationship is only statistically significant for Latin American economies.
- VIX and U.S. interest rates:
  - In the full model specification, VIX and U.S. interest rates do not appear to be strongly associated with capital flows.
  - Interpretation: much of their co-variation with capital inflows is already accounted for by commodity prices; when VIX is included alone as the only explanatory variable, it is statistically significant with the expected sign, but adding commodity prices renders VIX coefficients no longer statistically significant.
- Gross outflows:
  - Global and country-specific factors explain less of the variation in gross outflows than in gross inflows; R-squared in gross outflow regressions is less than half that of gross inflow regressions.

### Summary of stylized facts presented
- Gross inflows are the main driver of net inflows in emerging markets and Latin America.
- Capital inflows show both strong cyclical synchronization across countries and significant structural heterogeneity across countries.
- FDI is the largest inflow subcomponent by average size and is relatively more driven by structural, “between”-country factors; portfolio inflows are smaller on average but more volatile and driven more by “within”-country (cyclical) variation.
- Global commodity prices are a key global push factor associated with capital inflows, particularly for Latin America; global growth also matters for Latin American inflows.
- Traditional global risk and interest-rate indicators (VIX, U.S. interest rates) have limited explanatory power once commodity prices are included.

*Source: Chapter 4, "Drivers of Capital Flows and the Role of the Investor Base in Latin America," Regional Economic Outlook: Western Hemisphere, International Monetary Fund, April 2017.*

### 4. DRIvERs Of CApITAL fLOWs AND THE ROLE Of THE INvEsTOR BAsE IN LATIN AMERICA

### 4. DRIvERs Of CApITAL fLOWs AND THE ROLE Of THE INvEsTOR BAsE IN LATIN AMERICA

### Global cyclical drivers and empirical findings
- Global commodity prices exhibit the strongest association with the principal component of capital inflows and of asset prices in emerging markets:
  - Cross-correlation: principal component of capital inflows in EMs with Global commodity prices (log) = 0.82***.
  - Principal component of stock prices in EMs (log) with Global commodity prices (log) = 0.93***.
- Other global factor correlations (Table 4.4):
  - Principal component of capital inflows in EMs with principal component of stock prices in EMs (log) = 0.86***.
  - U.S. nominal effective exchange rate (log) is negatively correlated with the principal component of capital inflows in EMs = 20.75***.
  - vIX (log) correlation with principal component of capital inflows in EMs = 20.38***.
- VIX shows mixed results:
  - The VIX can have a negative association with capital flows in some models and a positive association in others; most relationships are not statistically significant.
- G7 real GDP growth (year over year) generally shows a positive association with capital inflows for Latin American countries in many specifications (see Table 4.3 coefficients and Figure 4.8 panel 3).

### Country-specific structural determinants (structural/slow-moving factors)
- Institutional and political factors matter strongly:
  - Improvements in government effectiveness, regulatory quality, control of corruption, or rule of law from current LA7 levels to the average among advanced economies would lead to a sustained increase in capital inflows of about 1½–1¾ percent of GDP.
  - Improving these indicators from current levels in Brazil, Colombia, and Peru to the levels observed in Chile would raise their capital inflow levels by about 1½–2 percent of GDP.
  - For Argentina, improving to Chile levels could raise capital inflow levels by up to 3 percent of GDP.
- Lower domestic corporate tax rates and higher credit ratings are associated with higher capital inflows.
  - Credit ratings reflect policy credibility and are used by institutional investors in asset eligibility.

### Subcomponents of capital flows: FDI versus portfolio and other investment
- Portfolio inflows are strongly associated with global and cyclical factors.
- FDI inflows are less associated with global cyclical factors, except:
  - Commodity prices have a positive association with FDI inflows in Latin American economies.
- Country-specific structural variables are important for explaining slower-moving changes in FDI, portfolio, and other investment.

### Robustness and model-selection checks
- Robustness exercise following a Sala-i-Martin–style approach:
  - N = 15 potential explanatory variables were compiled, yielding more than 32,000 possible models.
  - Histograms of estimated coefficients across all model combinations were used to assess robustness.
  - For global commodity prices:
    - For Latin American countries, all estimated coefficients related to global commodity prices are positive and significant across all model variants.
    - For other emerging markets, two-thirds of estimated coefficients related to commodity prices are positive and significant.
- Table 4.5: Share of statistically significant coefficients across models, 2000–16 (selected entries)
  - Global commodity prices: LA5 = 100 (expected sign); OEM = 67 (expected sign).
  - G7 real GDP growth: LA5 = 95 (expected sign); OEM = 1 (expected sign).
  - Real GDP growth differential: LA5 = 79 (expected sign); OEM = 2 (expected sign).
  - vIX: LA5 = 8 (expected sign) and 8 (opposite sign); OEM = 0 (expected sign) and 26 (opposite sign).

### Historical decompositions and regional differences
- Two periods analyzed for drivers of changes in capital inflows:
  1. Global financial crisis: most of the fall in capital inflows to Latin America was driven by global factors—mainly global output growth and commodity prices; for other emerging markets, the domestic growth differential accounted for more than a quarter of the explained variation.
  2. End of the commodity super-cycle: the sharp decline in commodity prices was the largest contributor to the reduction in capital inflows across EMs; the domestic growth differential accounted for:
     - 19 percent of the variation in inflows for other emerging market economies.
     - 9 percent of the variation in inflows for Latin America.
- Key regional contrast:
  - Once commodity prices are taken into account, capital inflows to Latin America are not strongly linked to domestic growth, while they remain highly linked for other emerging market economies.
  - Global factors play a predominant role in driving the cyclical behavior of capital flows to Latin American countries relative to other emerging market economies.

### Implications and forward-looking observations
- Because external demand and commodity prices are expected to remain low for the foreseeable future, downward pressure on capital inflows to Latin America and other emerging markets is likely to persist compared with the period following the global financial crisis.
- Despite this, inflows to Latin America are expected to remain relatively more resilient than in other emerging market economies whose weaker domestic growth prospects are also weighing down on capital inflows.

*International Monetary Fund | April 2017.*

### 4. DRIvERs Of CApITAL fLOWs AND THE ROLE Of THE INvEsTOR BAsE IN LATIN AMERICA

### 4. DRIvERs Of CApITAL fLOWs AND THE ROLE Of THE INvEsTOR BAsE IN LATIN AMERICA

### Methodology
- An interacted panel vector autoregression (IPVAR) estimation approach is used to assess how dynamic responses of capital flows to external shocks are affected by policy choices and characteristics of the investor base and the domestic financial system.
- Shocks analyzed include the VIX, global commodity prices, global GDP growth (G7 real GDP growth), and an identified U.S. monetary shock that increases U.S. interest rates.
- The model setup follows Towbin and Weber (2013). The U.S. monetary shock series used is the “identified” series estimated by Osorio Buitron and Vesperoni (2015).

### Key Quantitative Impulse-Response Findings
- A sustained 20 percent increase in commodity prices would be accompanied by an average increase in capital inflows of almost 2 percent of GDP to Latin America and other emerging market economies.
- An increase in the VIX of some 10 points would lead to a fall in capital inflows of the same magnitude (about the decline implied by the 20 percent commodity shock in absolute terms).
- An unanticipated U.S. monetary policy tightening of about 50 basis points would lead to a fall in capital inflows in emerging markets of close to 1 percent of GDP.
- In the IPVAR benchmark (unrestricted sample), shocks to the VIX or U.S. monetary policy rates lead to declines in capital inflows; positive shocks to global commodity prices and global growth boost capital inflows.

### Heterogeneity by Investor Base and Domestic Market Characteristics
- Foreign participation in domestic debt markets:
  - Countries with higher foreign participation are found to be more sensitive to external factors (greater sensitivity of capital inflows to shocks).
- Domestic market depth and liquidity:
  - Larger stock market capitalization relative to the economy and deeper domestic financial markets lower the sensitivity of capital flows to external shocks.
- Presence of pension funds:
  - A larger share of pension funds decreases the sensitivity of capital inflows to global factors (pension funds allocate to long-term stable investments).
- Exchange rate regime:
  - Capital inflows in countries with fixed exchange rate regimes tend to exhibit greater sensitivity to external shocks—particularly to U.S. monetary shocks—than those in countries with more flexible exchange rate arrangements.
- Capital account openness:
  - Countries with higher degrees of capital account openness show capital inflows that are more vulnerable to external conditions.

### Cross-Country Associations and Structural Factors
- Countries with both deeper markets and higher foreign participation tend to exhibit better macroeconomic and financial fundamentals, including:
  - Lower inflation rates and inflation volatility
  - Higher domestic growth and lower growth volatility
  - Lower sovereign spreads
  - More favorable credit ratings
  - Better governance indicators
- Country-specific structural “pull” factors (better governance, more efficient public institutions, stronger regulatory and legal frameworks, higher political stability and accountability) tend to attract higher average capital inflows.
- The cyclical component of capital inflows across emerging markets, particularly in Latin America, is driven importantly by global factors; commodity prices play a predominant role.

### Box 4.1 — Commodity Prices and Underlying Global Forces: Findings
- In the 2000s, commodity prices correlated more strongly with the principal component of gross capital inflows and stock prices in emerging markets than either the VIX (log) or U.S. short-term interest rates, suggesting commodity prices are a better proxy for the “global financial cycle” in that period.
- In the 1990s, commodity prices were less correlated with capital flows and stock prices than the VIX.
- Decomposition of commodity prices into demand and supply components indicates:
  - The demand component of commodity prices is positive and statistically significantly associated with gross capital inflows.
  - Increases in commodity prices attributable to increases in global aggregate demand tend to drive capital flows to emerging market economies.
  - The demand component shows its strongest role for “other” inflows (primarily cross-border bank lending), while portfolio flows show stronger sensitivity to global risk aversion and U.S. monetary policy.

### Representative Regression Estimates (Table 4.1.2) — Selected Coefficients (2000–16)
- Dependent variable: Total gross capital inflows (column (1))
  - vIX (log): 20.830 (0.813)
  - G7 real GDP growth (year over year): 20.776* (0.406)
  - U.S. short-term interest rates: 0.341 (0.383)
  - Real GDP growth differential (lagged): 0.435*** (0.107)
  - Demand component of commodity price (log): 0.252*** (0.046)
  - Supply component of commodity price (log): 2.989 (2.343)
  - Observations: 1,312; R-squared: 0.135; Number of countries: 22
- Dependent variable: Portfolio inflows (column (3))
  - vIX (log): 21.295*** (0.257)
  - G7 real GDP growth: 0.122* (0.067)
  - U.S. short-term interest rates: 20.159* (0.088)
  - Demand component of commodity price (log): 0.050** (0.022)
  - Observations: 1,312; R-squared: 0.168; Number of countries: 22
- Dependent variable: Other inflows (column (4))
  - Demand component of commodity price (log): 0.119*** (0.020)
  - Observations: 281; R-squared: 0.169; Number of countries: 22
- Note: Robust standard errors in parentheses; significance indicated by * p < 0.10; ** p < 0.05; *** p < 0.01.

### Policy Implications and Recommendations
- Monetary policy trade-offs:
  - Commodity-price-driven inflows raise growth and inflationary pressures (suggesting tighter policy) while also inducing exchange rate appreciation and higher inflows, complicating monetary tightening as a tool to forestall overheating.
- Recommended policy mix:
  - Use fiscal policy and exchange rate flexibility to complement monetary policy in response to surges in capital inflows.
  - Implement effective macroprudential policies to support monetary policy objectives and contain potential financial-stability side effects.
- Financial deepening and investor base:
  - Promoting deeper domestic financial markets and stable domestic financial intermediation (pension funds, insurance companies) can reduce vulnerabilities of capital flows to external shocks.
  - Opening capital accounts to foreign participation to gain market depth is associated, on average, with better macroeconomic performance than remaining relatively closed with shallower markets; however, the pace of financial opening should be aligned with financial stability considerations to avoid rapid risk build-up.
- Exchange rate policy:
  - Allowing for more exchange rate flexibility is effective in reducing the sensitivity of capital flows to adverse external shocks.

_International Monetary Fund | April 2017_

### Box 4.1 (continued)

### Box 4.1 (continued)

### Commodity prices and firm-level investment: empirical patterns
- Firm-level investment growth comoves with country-specific commodity export price changes for publicly listed firms in Argentina, Brazil, Chile, Colombia, Mexico, and Peru.
- Panel evidence (Figure 4.2.1):
  - Investment growth tracks the growth rate in export-related commodity prices for the median agricultural and mining firm.
  - Investment growth for all other firms shows a similar, albeit slightly less volatile, pattern.
  - Broad conclusion: investment in commodity-producing firms and in other firms responds similarly to changes in commodity prices, indicating important spillover effects beyond direct commodity sectors.

### Microeconometric investment specification (equation 4.2.1)
- Estimated specification (notation preserved):
  - I_i,t / K_i,t-1 = 100 * [ a Q_i,t + b (π_i,t / K_i,t-1) + λ (D_i,t / E_i,t) + ρ (IE_i,t / D_i,t-1) + τ ((D_i,t / K_i,t-1)) + k P^x_i,t + m P^x_i,t × X_i + X_i controls ] + u_t + c_i + u_i,t
- Interpretations:
  - I = investment in fixed capital; K = capital stock; Q = market capitalization to book value of equity (proxy for Tobin’s q); π = net profit; D = book value of debt; E = book value of equity; IE = interest expense; P^x = commodity export price in firm’s domicile country; X = indicator for agriculture or mining industries; k measures comovement of investment with commodity prices; m measures difference in comovement between commodity-producing firms and other firms.
- Estimation approach: fixed effects on firm-year observations for firms domiciled in the LA6 countries.

### Key micro estimation results (Table 4.2.1)
- Coefficient estimates (standard errors in parentheses):
  - a (Q_i,t): 1.56*** (0.310) in column (1); 1.55*** (0.310) in column (2).
  - b (π_i,t / K_i,t-1): 0.58 (0.675) in both columns.
  - λ (D_i,t / E_i,t): 24.43*** (0.706) in column (1); 24.42*** (0.705) in column (2).
  - ρ (IE_i,t / D_i,t-1): 3.88 (4.724) in column (1); 3.98 (4.723) in column (2).
  - τ ((D_i,t / K_i,t-1)): 0.16 (0.142) in column (1); 0.15 (0.141) in column (2).
  - year t (u): 20.17*** (0.060) in column (1); 20.16*** (0.059) in column (2).
  - k (P^x_i,t): 0.05*** (0.013) in column (1); 0.04*** (0.014) in column (2).
  - m (P^x_i,t × X_i): (restricted m = 0 in column (1)); 0.04 (0.040) in column (2).
- Sample sizes:
  - Number of observations: 4,651 (column (1)); 4,650 (column (2)).
  - Number of firms: 763 (column (1)); 762 (column (2)).
- Inference:
  - Strong role for Tobin’s q (a positive and highly significant coefficient).
  - Positive and statistically significant k (k > 0) indicates a positive association between commodity prices and investment.
  - m is not statistically significant in column (2), indicating that higher commodity prices raise investment by both commodity producers and other firms in a similar fashion (spillovers matter).

### Investor base, domestic market measures, and IPVAR methodology (Annex 4.1)
- Investor base and domestic market measures (definitions preserved):
  - Foreign participation in domestic debt markets: share of domestic debt instruments held by nonresidents out of total domestic debt instruments (Arslanalp and Tsuda (2014)).
  - Domestic market capitalization: ratio of total domestic market capitalization to nominal GDP (World Bank GFDD, following Cihak and others (2012)).
  - Presence of domestic pension funds: ratio of total assets under management of domestic pension funds to the country’s total financial sector assets (GFDD).
  - Exchange rate flexibility index: Aizenman, Chinn, and Ito (2010) index; high score = fixed exchange rate regime; low score = more flexible arrangement.
  - Capital account openness indicator: Aizenman, Chinn, and Ito (2010) indicator; value of one for relatively open countries and zero for relatively closed countries.
- Interacted Panel Vector Autoregression (IPVAR):
  - Purpose: explore how impulse response of capital inflows to external shocks depends on investor base and domestic financial market characteristics.
  - General panel VAR form shown, with y (country-specific variables) and X (global variables). X includes global commodity prices, VIX, G7 real GDP growth, and identified monetary shock to U.S. interest rates (Osorio Buitron and Vesperoni (2015)); X treated as block-exogenous.
  - In IPVAR, coefficients in A_j are functions of country-specific characteristics F_i,t: a_i,t = c + Γ' F_i,t. This allows coefficients to vary by country and over time according to investor-base measures.

### Macro cross-country findings (Annex tables)
- Core specification for gross inflows, 2000–16 (Annex Table 4.1) — selected coefficients and significance:
  - Global commodity price (log): 4.182*** (0.387) for LA5 core model; 4.458** (1.354) for LA7; 4.918*** (1.261) for OEM; 4.434*** (0.973) for EMs.
  - G7 real GDP growth (year over year): 0.509** (0.171) for LA5 in first stage.
  - vIX (log): 1.230 (0.862) for LA5 first stage.
  - Country-specific real GDP growth differential (lagged): 0.055 (0.116) for LA5 first stage.
  - Observations and fit:
    - Observations: 322 (LA5), 440 (LA7), 872 (OEM), 1,312 (EMs).
    - R-squared (first stage): 0.480 (LA5), 0.385 (LA7), 0.141 (OEM), 0.151 (EMs).
- Alternative specifications and splits (Annex Table 4.2) show robustness of global commodity price coefficient in several specifications for LA5 and OEM samples (examples: 4.182*** (0.387); 4.918*** (1.261); 4.024*** (0.567); 2.566** (0.919); 3.989*** (0.410); 2.641*** (0.832)).
- Disaggregated inflows (Annex Table 4.3) — selected results for Gross FDI and Gross portfolio inflows, 2000–16:
  - Global commodity price (log) in the first stage:
    - Gross FDI inflows: 1.493* (0.594) for LA5; 1.508** (0.527) for LA7; 2.100 (1.196) for OEM.
    - Gross portfolio inflows: 1.206* (0.446) for LA5; 2.140** (0.711) for LA7; 1.252*** (0.390) for OEM.
  - vIX (log) for gross portfolio inflows: negative and significant for OEM in first stage: 20.938*** (0.227).
  - Country-specific institutional variables (second stage) show consistently positive and often highly significant coefficients (examples):
    - Government effectiveness: 2.214*** (0.087) for LA5 (FDI); 0.400*** (0.025) for LA5 (portfolio).
    - Regulatory quality: 2.670*** (0.064) for LA5 (FDI); 0.521*** (0.021) for LA5 (portfolio).
    - Control of corruption, rule of law, voice and accountability, political stability, political risk, polity synthetic index, credit rating — many show positive and significant relationships with inflows.
  - Corporate tax rate shows negative and significant coefficients in both FDI and portfolio second-stage estimates (examples: 20.219*** (0.007) for LA5 FDI; 20.047*** (0.002) for LA5 portfolio).

### Synthesis of empirical implications
- Commodity prices play a strong role in determining capital flows to emerging markets in Latin America, with a pronounced and robust positive association between global or country-specific commodity price indices and gross capital inflows.
- Micro evidence indicates that higher commodity prices raise investment not only in commodity-producing firms but also in noncommodity firms, pointing to spillover effects beyond direct profitability channels.
- Investor-base and institutional characteristics matter for the sensitivity of inflows; stronger institutions, higher credit ratings, and better governance measures are associated with larger inflows across specifications.

*Source: IMF staff calculations, Box 4.1 (continued) and Annex 4.1, Regional Economic Outlook: Western Hemisphere, April 2017.*

---


_Source: https://www.imf.org/-/media/files/publications/reo/whd/2017/may/wreo0517-chp4.pdf_