## 1. Private Capital Flows Before and After Capital Controls

## Source details

**Canonical URL:** [1. Private Capital Flows Before and After Capital Controls](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp0930.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp0930.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp0930.pdf.json)

---

### I. Introduction and Policy Instruments
- Context: sharp appreciation of currencies in emerging markets through mid-2008 revived interest in capital controls; Colombia imposed market-based controls in May 2007 to slow inflows triggered by higher domestic interest rates.
- Main instruments introduced in May 2007:
  - an unremunerated reserve requirement (URR) on foreign borrowing and portfolio inflows.
  - limits on banks’ currency derivative positions to address systemic risk and circumvention.
- Timeline of key policy changes:
  - May 6, 2007: URR on external borrowing of 40 percent, to be held six months at BdR; applied to private and public sectors.
  - May 23, 2007: URR extended to portfolio inflows by foreign residents; early withdrawal penalties ranged from 9.4 percent (immediate) to 1.6 percent (if held for 5 months).
  - July 4, 2007: ceiling on banks’ gross currency derivative positions set to 500 percent of capital (binding date).
  - December 2007: URR for IPOs of equities eliminated; early-withdrawal penalties reduced.
  - May 2008: URR on portfolio inflows raised from 40 percent to 50 percent; minimum stay of two years imposed on FDI; gross derivatives limit increased from 500 percent to 550 percent of capital.
  - June 2008: early-withdrawal penalty increased.
  - September 2008: minimum stay on FDI revoked; equities purchases exempted from URR.
  - October 2008: capital controls eliminated; derivative position limit of 550 percent of capital remained.
- Exclusions and exemptions: Colombian institutional funds, including pension funds, excluded; June 2007: equities issued abroad (e.g., ADRs) exempted.

### II. Data and Stylized Facts
- Data source: Banco de la República (BdR) balanza cambiaria; weekly coverage; captures capital account transactions intermediated in the foreign exchange market and recorded by BdR.
- Coverage: external borrowing, portfolio investment, foreign direct investment (FDI), and changes in residents’ bank accounts abroad (cuentas corrientes de compensación); excludes capital transactions without an exchange of currency.
- Pre-controls movements and policy backdrop:
  - peso appreciated about 9 percent in nominal effective terms from July through December 2006.
  - BdR began monetary tightening in April 2006; interest rate differentials rose.
  - early 2007: additional 75 basis points increase in policy rates in first four months of year; exchange rate appreciated additional 5 percent in nominal effective terms.
  - sterilized intervention from mid-January 2007 accumulated about 38 percent of base money in first four months of 2007; BdR abandoned discretionary foreign exchange intervention by end-April 2007.
- Features of Colombian private capital flows:
  - Net terms: FDI accounts for vast majority of inflows.
  - Pre-controls era: FDI accounted for 96 percent of net private inflows; net non-FDI flows close to zero.
  - Resident portfolio flows (gross) substantially larger than nonresidents; on net the reverse holds.
  - For resident portfolio flows weekly net movements are between 2 and 8 percent of gross (depending on period).
  - Gross flows much larger than net flows; resident inflows and outflows highly correlated (about 90 percent).

### III. Key Weekly Statistics (pre-controls vs post-controls; US$ millions)
- Table 1 weekly averages (Before; After; Change = After minus Before):
  - Total net inflows: Before 99.6; After 104.9; Change 5.3
  - Total gross inflows: Before 671.1; After 727.7; Change 56.7
  - Total non-FDI net inflows: Before 3.7; After -32.8; Change -36.4
  - Total non-FDI gross inflows: Before 565.6; After 580.6; Change 15.0
  - Borrowing net inflows: Before -10.8; After -17.1; Change -6.3
  - Borrowing gross inflows: Before 41.1; After 17.1; Change -24.1 ***
  - Portfolio (Non-residents) net inflows: Before 25.2; After -3.6; Change -28.8 ***
  - Portfolio (Non-residents) gross inflows: Before 31.9; After 9.9; Change -22.0 ***
  - Portfolio (Colombian residents) net inflows: Before -4.7; After -18.1; Change -13.3
  - Portfolio (Colombian residents) gross inflows: Before 155.9; After 227.5; Change 71.5 ***
  - Drawdown of bank accounts abroad net inflows: Before -0.9; After 0.3; Change 1.2
  - Drawdown of bank accounts abroad gross inflows: Before 347.4; After 320.6; Change -26.9
  - FDI net inflows: Before 95.9; After 137.7; Change 41.8 **
  - FDI gross inflows: Before 105.5; After 147.6; Change 42.2 **
- Memorandum (average weekly flows, annualized as share of GDP): Total net flows 2.8; After 2.4. Total non-FDI gross inflows 16.0; After 13.5.
- Note on significance: ***, **, * denote 1%, 5%, and 10% levels respectively (as used in source).
- Pre-controls period: Average flow for July 6th, 2006 to May 6th (for portfolio May 23rd), 2007.
- Post-controls period: Average flow for May 7th (for portfolio May 24th), 2007 to Jul 4th, 2008.

### IV. Market Responses and Secondary Effects
- Offshore derivatives and circumvention:
  - Cross-border foreign currency derivative contracts between domestic pension funds and offshore entities rose from an average monthly value of US$6 million in pre-control period to about US$405 million since controls — over a 60-fold increase (reached US$800 million in July 2008).
  - Average monthly value of contracts traded between off-shores and local banks increased by 250 percent.
  - Increase likely due to the derivative position ceiling on local banks, shifting hedging to offshore counterparties; imposition of higher reserve requirements on new deposits in May 2007 (encaje marginal) also constrained local banks’ liquidity and forward provision.
- Exchange rate evolution during controls:
  - Controls were not associated with an immediate halt to appreciation; peso continued to appreciate until brief global turbulence in June–August 2007.
  - From start of controls to end-July 2008, bilateral rate appreciated by 16 percent.
  - Later in second half of 2008 peso depreciated sharply following weakening commodity prices and global financial turbulence.

### V. Empirical Framework and Identification
- Capital flows regressions:
  - OLS regressions on weekly balanza cambiaria data for net and gross flows by category (foreign borrowing; portfolio investment; residents’ bank accounts abroad; total non-FDI).
  - Key regressors: BdR discretionary intervention (INT, US$ millions), interest rate differential (ri−*, percent per year), EMBI spread (percent per year), unconditional volatility of daily nominal exchange rate over previous three months (VOL), expected appreciation proxy (EApp: spot vs 15–35 day forward NDF percentage difference).
  - Capital controls proxied by dummy DCC = 1 in weeks after controls were imposed, 0 otherwise. Parameter of interest δ6 measures impact of controls on weekly flows.
- Exchange rate model:
  - GARCH(1,1) on daily peso-dollar returns with mean equation including INT, interest rate differential, EMBI spread (first differences), VIX, daily change in commodity export price index (Commis), day-of-week dummies, and DCC. Volatility equation includes absolute unexpected returns and same financial controls. Parameter of interest β6 measures impact of controls on exchange rate returns.
- Identification considerations:
  - High-frequency weekly/daily data over short period reduces endogeneity and unit-root concerns.
  - Under full sterilization and inflation-targeting, two-way causality between flows and interest differentials argued to be limited.
  - Challenges: concurrent derivative-position limit, staged imposition and subsequent modifications of controls, inability to fully separate URR and derivative-limit effects.

### VI. Main Empirical Findings
- Aggregate non-FDI flows:
  - Capital controls had no statistically significant impact on gross or net total non-FDI capital flows once other drivers were accounted for.
- Foreign borrowing:
  - Controls contributed to a significant decline in foreign borrowing in both gross and net terms.
  - Point estimate: reduction in weekly gross foreign currency borrowing of approximately US$44 million (about equal to average pre-controls level).
  - Net reduction significant at 10 percent confidence level in some specifications.
- Portfolio inflows:
  - No statistical evidence that controls reduced aggregate portfolio inflows or bank account transfers of Colombian residents (majority of total non-FDI flows).
  - For non-resident portfolio flows: effect on gross inflows apparent, but not on net inflows.
  - For resident portfolio flows: no significant effect in gross or net terms (consistent with exclusion of domestic institutional funds from controls and continued high resident inflows).
- Exchange rate and volatility:
  - No statistical evidence that capital controls resulted in a more depreciated exchange rate or reduced the sensitivity of the exchange rate to interest rate differentials.
  - Interaction terms (DCC * Interest Rate Spread, DCC * Change in EMBI, DCC * VIX, DCC * Commodity Price Change) are reported as statistically insignificant for reducing sensitivity to global developments.
  - Capital controls are associated with an increase in conditional volatility of exchange rate returns.
  - Selected GARCH results (examples):
    - Dummy Capital Controls (DCC): -0.07 (Mean) and 0.16 ** (Variance) in baseline specification.
    - Change in EMBI Sovereign Spread: 2.96 *** (Mean) and 0.73 * (Variance).
    - Commodity Price Change: -0.04 *** (Mean) and -0.02 ** (Variance).
- Behavioral/market structure effects:
  - Pension funds’ average weekly trading in foreign exchange markets (proxied by sum of gross portfolio inflows and outflows by residents in balanza cambiaria) increased 52 percent across periods.
  - Minimum return rules for pension funds may encourage herd behavior and large sudden movements between peso- and foreign-currency denominated assets.
- Model priors confirmed:
  - An increase in interest rate spreads, a reduction in exchange rate volatility, and an increase in expected appreciation are estimated to lead to higher net inflows.
  - Exchange rate intervention is estimated to lead to higher non-FDI net inflows.

### VII. Conclusions and Policy Implications
- Effectiveness:
  - Capital controls appear successful in reducing foreign borrowing but have not affected more significant inflow categories such as portfolio inflows.
  - For non-FDI net flows as a whole, controls have no statistically significant impact.
- Exchange rate:
  - Controls have not affected the level of the exchange rate nor reduced its sensitivity to global financial developments; instead, controls have been associated with increased exchange rate volatility.
- Monetary policy independence:
  - Controls do not appear to have affected the independence of monetary policy (no statistically significant relationship between capital controls and the interface between interest rate differentials and the exchange rate).
- Economic significance:
  - Given the modest share of foreign borrowing in total flows, and statistically insignificant relationships for other flows, the effect of the controls on total non-FDI flows is economically and statistically insignificant.
- Observed secondary effects to monitor:
  - Shift of hedging and transactions to offshore counterparties.
  - Increased role of excluded domestic institutional investors (pension funds) in FX markets.

### VIII. Recommendations for Further Research (authors' suggestions)
- Study channels through which carry trades are executed to assess effectiveness of capital controls.
- Examine the role of the offshore sector and pension funds in the exchange rate market.
- Conduct further empirical work on the effects of capital controls on exchange rate volatility.
- Use different measures of capital controls in empirical work, including:
  - A “tax equivalent” measure quantifying the implicit premium required on local returns to compensate investors for financial costs associated with capital controls.
  - Measures derived from banks’ gross derivatives positions, particularly how closely these come to the ceiling of 550 percent of gross capital.
- Broader assessment of effects of controls on financial stability, including their role in reducing vulnerabilities associated with rapid increases in capital inflows and buildup of derivative positions in the banking system.

*Source: _wp0930 - 1. Private Capital Flows Before and After Capital Controls (PDF chapter).*

### 1. Private Capital Flows Before and After Capital Controls .............................................8

### 1. Private Capital Flows Before and After Capital Controls

### Major sections (chapter structure)
- 1. Private Capital Flows Before and After Capital Controls .............................................8
- 2. Impact of Capital Controls on Non-FDI Private Capital Flows ..................................18
- 3. Impact of Capital Controls on Foreign Borrowing and Drawdown of Residents’ Bank Accounts Abroad ................................................................................................19
- 4. Impact of Capital Controls on Portfolio Inflows .........................................................19
- 5. Impact of Capital Controls on the Exchange Rate.......................................................21

### Figures
- Figure 1. Private Capital Flows Before and After Capital Controls .............................................9

*Source: _wp0930 - 1. Private Capital Flows Before and After Capital Controls, PDF chapter.*

### References..............................................................................................................

### _wp0930 - References..............................................................................................................

### I. Introduction
- Context: sharp appreciation of currencies in emerging markets through mid-2008 revived interest in capital controls, especially for inflation-targeting countries facing overheating and currency appreciation pressures.
- Case: Colombia imposed market-based controls in May 2007 to slow inflows triggered by higher domestic interest rates.
- Main instruments in May 2007:
  - an unremunerated reserve requirement (URR) on foreign borrowing and portfolio inflows.
  - limits on banks’ currency derivative positions to address systemic risk and circumvention.
- Paper scope:
  - assesses effects of recent controls on capital flows and the exchange rate.
  - focuses on disaggregated categories of capital inflows and both gross and net flows.
  - does not analyze microeconomic costs, market-development effects, or full financial-stability benefits.

### II. Database and Stylized Facts on Capital Flows and Exchange Rates
- Data source: Banco de la República (BdR) balanza cambiaria; captures all capital account transactions by the private sector intermediated in the foreign exchange market and recorded by BdR; available weekly.
- Coverage: external borrowing, portfolio investment, foreign direct investment (FDI), and changes in residents’ bank accounts abroad (cuentas corrientes de compensación); excludes capital transactions without an exchange of currency.
- Pre-controls developments:
  - peso appreciated about 9 percent in nominal effective terms from July through December 2006.
  - BdR began monetary tightening in April 2006; interest rate differentials rose.
  - net FDI inflows strong; net non-FDI flows muted.
  - early 2007: additional 75 basis points increase in policy rates in first four months of year; exchange rate appreciated additional 5 percent in nominal effective terms.
  - sterilized intervention from mid-January 2007 accumulated about 38 percent of base money in first four months of 2007; BdR abandoned discretionary foreign exchange intervention by end-April 2007.
- Controls-era overview:
  - May 6, 2007: URR on external borrowing of 40 percent, to be held six months at BdR; applied to private and public sectors.
  - July 4, 2007: ceiling on banks’ gross currency derivative positions set to 500 percent of capital (binding date).
  - May 23, 2007: URR extended to portfolio inflows by foreign residents; early withdrawal penalties ranged from 9.4 percent (immediate) to 1.6 percent (if held for 5 months).
  - Exclusions and exemptions:
    - Colombian institutional funds, including pension funds, excluded.
    - June 2007: equities issued abroad (e.g., ADRs) exempted.
  - Subsequent changes:
    - December 2007: URR for IPOs of equities eliminated; early-withdrawal penalties reduced.
    - May 2008: URR on portfolio inflows raised from 40 percent to 50 percent; minimum stay of two years imposed on FDI; gross derivatives limit increased from 500 percent to 550 percent of capital.
    - June 2008: early-withdrawal penalty increased.
    - September 2008: minimum stay on FDI revoked; equities purchases exempted from URR.
    - October 2008: capital controls eliminated; derivative position limit of 550 percent of capital remained.
- Features of Colombian private capital flows:
  - Net terms: FDI accounts for vast majority of inflows.
  - Pre-controls era: FDI accounted for 96 percent of net private inflows; net non-FDI flows close to zero.
  - Resident portfolio flows (gross) substantially larger than nonresidents; on net the reverse holds.
  - For resident portfolio flows weekly net movements are between 2 and 8 percent of gross (depending on period).
  - Gross flows much larger than net flows, inflated by transactions between residents; resident inflows and outflows highly correlated (about 90 percent).

### III. Key Statistics (weekly averages and changes; exact figures preserved)
- Table 1 summary (Weekly average flows in US$ millions; pre-controls vs post-controls; change = After minus Before):
  - Total net inflows: Before 99.6; After 104.9; Change 5.3
  - Total gross inflows: Before 671.1; After 727.7; Change 56.7
  - Total non-FDI net inflows: Before 3.7; After -32.8; Change -36.4
  - Total non-FDI gross inflows: Before 565.6; After 580.6; Change 15.0
  - Borrowing net inflows: Before -10.8; After -17.1; Change -6.3
  - Borrowing gross inflows: Before 41.1; After 17.1; Change -24.1 ***
  - Portfolio (Non-residents) net inflows: Before 25.2; After -3.6; Change -28.8 ***
  - Portfolio (Non-residents) gross inflows: Before 31.9; After 9.9; Change -22.0 ***
  - Portfolio (Colombian residents) net inflows: Before -4.7; After -18.1; Change -13.3
  - Portfolio (Colombian residents) gross inflows: Before 155.9; After 227.5; Change 71.5 ***
  - Drawdown of bank accounts abroad net inflows: Before -0.9; After 0.3; Change 1.2
  - Drawdown of bank accounts abroad gross inflows: Before 347.4; After 320.6; Change -26.9
  - FDI net inflows: Before 95.9; After 137.7; Change 41.8 **
  - FDI gross inflows: Before 105.5; After 147.6; Change 42.2 **
- Notes:
  - Statistically significant changes: ***, **, * correspond to 1%, 5%, and 10% levels respectively (as used in source).
  - Pre-controls period: Average flow for July 6th, 2006 to May 6th (for portfolio May 23rd), 2007.
  - Post-controls period: Average flow for May 7th (for portfolio May 24th), 2007 to Jul 4th, 2008.
  - Memorandum items (average weekly flows, annualized as share of GDP): Total net flows 2.8; After 2.4. Total non-FDI gross inflows 16.0; After 13.5.

### IV. Observed Market Responses and Secondary Effects
- Offshore derivative activity:
  - Cross-border foreign currency derivative contracts between domestic pension funds and offshore entities rose from an average monthly value of US$6 million in pre-control period to about US$405 million since controls — over a 60-fold increase (reached US$800 million in July 2008).
  - Average monthly value of contracts traded between off-shores and local banks increased by 250 percent.
  - Increase likely due to the derivative position ceiling on local banks, shifting hedging to offshore counterparties; also affected by imposition of higher reserve requirements on new deposits in May 2007 (encaje marginal) that constrained local banks’ liquidity and forward provision.
- Exchange rate evolution:
  - Capital controls were not associated with an immediate halt to appreciation; peso continued to appreciate until brief global turbulence in June–August 2007.
  - From start of controls to end-July 2008, bilateral rate appreciated by 16 percent.
  - Later in second half of 2008 peso depreciated sharply following weakening commodity prices and global financial turbulence.

### V. Literature Insights on Capital Controls Effectiveness
- Framework for assessment (Magud and Reinhart, 2007): effectiveness judged by ability to:
  1. limit capital inflows;
  2. change composition toward long-term assets;
  3. alleviate real exchange rate pressures;
  4. provide monetary policy independence.
- Empirical consensus in literature:
  - Controls most successful at providing monetary policy autonomy and altering composition of inflows.
  - Mixed success at reducing volume of inflows and relieving exchange rate pressures.
- Colombian 1990s studies:
  - Findings broadly: controls ineffective at reducing inflows or altering real exchange rate, but affected composition and monetary autonomy.
  - Specific results cited: Cárdenas and Barrera (1997), Rocha and Mesa (1998), Ocampo and Tovar (1999, 2003), Rincón (2000), Villar and Rincón (2001) — varied findings on short- vs long-term flows and on real rates.
- Recent studies on 2007 controls:
  - Cardenas (2007): OLS on weekly data for assets > one year maturity — finds no significant effect on capital inflows.
  - Galindo (2007) and Concha and Galindo (2008): VECM with monthly data — find significant effect on short-term capital inflows, but not on real exchange rate; controls reduce real exchange rate volatility.

### VI. Empirical Methodology (models and identification)
- Capital flows model:
  - OLS regressions on weekly balanza cambiaria data for net and gross flows by category.
  - Four categories: foreign borrowing; portfolio investment (bonds and equities); movements in residents’ bank accounts abroad (cuentas corrientes de compensación); total non-FDI flows.
  - Key regressors: BdR discretionary intervention (INT, US$ millions), interest rate differential (ri−*, percent per year), EMBI spread (percent per year), unconditional volatility of daily nominal exchange rate over previous three months (VOL), expected appreciation proxy (EApp: spot vs 15–35 day forward NDF percentage difference).
  - Capital controls proxied by dummy DCC = 1 in weeks after controls were imposed, 0 otherwise.
  - Parameter of interest: δ6 (impact of capital controls on weekly flows).
- Exchange rate model:
  - GARCH(1,1) estimated on daily peso-dollar exchange rate returns.
  - Mean equation regressors: BdR intervention (INT), interest rate differential, EMBI spread (in first differences), VIX (implied volatility of U.S. stock market), daily change in commodity export price index (Commis), day-of-week dummies, capital controls dummy (DCC).
  - Volatility equation includes absolute unexpected returns and same financial controls; parameter of interest β6 (impact of controls on exchange rate returns).
- Identification and econometric considerations:
  - High-frequency weekly/daily data over short period reduces endogeneity and unit-root concerns.
  - Under Colombia’s inflation-targeting and full sterilization of inflows, two-way causality between capital flows and interest differentials is argued to be limited.
  - Challenges remain: concurrent derivative-position limit, staged imposition and subsequent modifications of controls, inability to separate URR and derivative-limit effects fully.

### VII. Empirical Results (summary of main findings)
- Aggregate non-FDI flows:
  - Econometric estimates indicate capital controls had no statistically significant impact on gross or net total non-FDI capital flows once other drivers were accounted for.
- Foreign borrowing:
  - Controls contributed to a significant decline in foreign borrowing in both gross and net terms.
  - Point estimate: reduction in weekly gross foreign currency borrowing of approximately US$44 million (about equal to average pre-controls level).
  - Net reduction significant at 10 percent confidence level.
- Portfolio inflows:
  - No statistical evidence that controls reduced aggregate portfolio inflows or bank account transfers of Colombian residents (majority of total non-FDI flows).
  - For non-resident portfolio flows: effect on gross inflows apparent, but not on net inflows.
  - For resident portfolio flows: no significant effect in gross or net terms (consistent with exclusion of domestic institutional funds from controls and continued high resident inflows).
- Exchange rate:
  - Controls were not associated with an immediate effect on exchange rate appreciation; peso continued appreciating through mid-2007 and overall appreciated 16 percent from start of controls to end-July 2008.
  - Proper assessment requires models that account for changing domestic and external determinants (monetary policy stance, cessation of discretionary intervention, rising interest differentials, EMBI spreads, VIX, commodity prices, FDI flows, security improvements).

*Italic source attribution: Content summarized from _wp0930 - References (PDF chapter/section)._

### 2008. Results are available upon request.

### _wp0930 - 2008. Results are available upon request.

### Impact of capital controls on capital flows — summary findings
- During the peak period of appreciation pressures—during the first four months of 2007—these inflows, in net terms, equaled ¾ percentage point of annualized GDP.
- Non-resident gross portfolio inflows averaged about 5 percent of total gross capital inflows in the pre-controls era.
- Overall evidence on the effects of the controls on capital flows is mixed:
  - Foreign borrowing appears to have declined in net terms, but only at the 10 percent confidence level.
  - Given the modest share of foreign borrowing in total flows, and statistically insignificant relationships for other flows, the effect of the controls on total non-FDI flows is economically and statistically insignificant.
- Model priors confirmed: an increase in interest rate spreads, a reduction in exchange rate volatility, and an increase in expected appreciation are estimated to lead to higher net inflows.
- Exchange rate intervention is estimated to lead to higher non-FDI net inflows.

### Selected econometric results (high-level)
- Table 2 (Total Non-FDI, OLS weekly flows, July 2006 - July 2008, Number of Observations 101):
  - R-Squared: 0.36 for Net Inflows.
  - Coefficients for Net Inflows (selected):
    - α1 (EMBI Sovereign Spread): -28.3
    - α2 (Interest Rate Spread): 42.9 ***
    - α3 (Exchange Rate Volatility): -126.2 *
    - α4 (Expected Appreciation): 37.8 ***
    - α5 (Discretionary Intervention): 0.31 ***
    - α6 (Dummy for period 6th May 2007 onwards): 14.2
- Table 3 (Foreign Borrowing and Drawdown of Residents' Bank Accounts Abroad, OLS weekly flows, July 2006 - July 2008):
  - R-Squared reported range across columns: 0.28, 0.42, 0.20, 0.14.
  - Effect of capital controls (α6 Dummy for Period May 6th 2007 onwards) examples:
    - Drawdown of Residents' Bank Accounts Abroad: α6 = -32.5 *
    - Foreign Borrowing (one column): α6 = -44.2 ***
- Table 4 (Portfolio Inflows, OLS weekly flows, July 2006 - July 2008):
  - R-Squared reported range across columns: 0.28, 0.33, 0.07, 0.35, 0.45, 0.66.
  - Example coefficient: Effect of Capital Controls (α6 Dummy for Period May 23rd 2007 onwards) on one series: -13.3-34.6 *** (table reports multi-column values).
- Note: Asterisks in tables denote significance with ***, ** and * indicating significance at the 1%, 5% and 10% level, respectively. Source indicated as "Authors' estimates."

### Impact on exchange rate and volatility
- Using GARCH(1,1) models for daily percentage changes in the US dollar–peso exchange rate (July 2006–July 2008):
  - No statistical evidence that capital controls resulted in a more depreciated exchange rate or reduced the sensitivity of the exchange rate to interest rate differentials.
  - Interaction terms (DCC * Interest Rate Spread, DCC * Change in EMBI, DCC * VIX, DCC * Commodity Price Change) are reported as statistically insignificant for reducing sensitivity to global developments.
  - Capital controls are associated with an increase in conditional volatility of exchange rate returns.
- Selected coefficients from Table 5 (Baseline and Augmented Interaction Model; Number of Observations 477):
  - Lagged Exchange Rate: 0.10 ** (Mean, baseline)
  - Central Bank Intervention: 0.03 (Mean) and 0.04 (Variance) in one specification
  - Change in EMBI Sovereign Spread: 2.96 *** (Mean) and 0.73 * (Variance)
  - Commodity Price Change: -0.04 *** (Mean) and -0.02 ** (Variance)
  - Dummy Capital Controls (DCC): -0.07 (Mean) and 0.16 ** (Variance) in baseline; DCC reported as 0.16 ** increasing volatility in one specification
  - Adjusted R-Squared: 0.15 (mean), 0.15 (variance) in baseline specifications
- Proposed mechanisms for increased volatility include:
  - Quantity controls shifting adjustment burden to prices.
  - Further segmentation of foreign exchange markets increasing conditional volatility.
  - Pension funds being excluded from controls, increasing their relative trading importance:
    - Average weekly trading in foreign exchange markets by pension funds (proxied by sum of gross portfolio inflows and outflows by residents in balanza cambiaria) increased 52 percent across periods.
    - Minimum return rules may encourage herd behavior and large sudden movements between peso- and foreign-currency denominated assets.

### Conclusions and policy implications
- Capital controls appear successful in reducing foreign borrowing but have not affected more significant inflow categories such as portfolio inflows.
- For non-FDI net flows as a whole, controls have no statistically significant impact.
- Controls have not affected the level of the exchange rate nor reduced its sensitivity to global financial developments; instead, controls have been associated with increased exchange rate volatility.
- Controls do not appear to have affected the independence of monetary policy (no statistically significant relationship between capital controls and the interface between interest rate differentials and the exchange rate).

### Suggestions for future research (authors' recommendations)
- Study channels through which carry trades are executed to assess effectiveness of capital controls.
- Examine the role of the offshore sector and pension funds in the exchange rate market.
- Conduct further empirical work on the effects of capital controls on exchange rate volatility.
- Use different measures of capital controls in empirical work, including:
  - A “tax equivalent” measure quantifying the implicit premium required on local returns to compensate investors for financial costs associated with capital controls.
  - Measures derived from banks’ gross derivatives positions, particularly how closely these come to the ceiling of 550 percent of gross capital.
- Broader assessment of effects of controls on financial stability, including their role in reducing vulnerabilities associated with rapid increases in capital inflows and buildup of derivative positions in the banking system.

*Source: Authors' estimates.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp0930.pdf_