This chapter describes the mechanisms used by the Central Bank of Colombia for intervention in the foreign exchange market and explains the rationale and operations. The text characterizes the instruments used to sterilize the monetary effects of US dollar purchases, as well as the effects of the sterilized intervention on the central bank’s balance sheet. At the end, a short survey of the effectiveness of foreign exchange sterilized intervention in Colombia is presented.
Introduction
The policy strategy of Banco de la República, Colombia’s central bank, seeks to maintain a low and stable inflation rate while keeping production levels close to their potential value.1 The central bank’s policy also helps maintain financial stability and adequate functioning of the payments system.
Exchange rate flexibility is fundamental to achieve these objectives for a number of reasons. First, in a flexible exchange rate regime, the exchange rate serves as an adjustment variable against the shocks faced by the economy, which reduces volatility in economic activity.2 Second, exchange rate flexibility allows the interest rate to be used independently as a tool to guide inflation and production to their desired values. Third, exchange rate flexibility reduces incentives for excessive foreign currency risk taking by economic agents, a reduction that is vital for maintaining financial stability.
However, the central bank, as the foreign exchange authority, can intervene in the foreign exchange market. Such intervention does not limit exchange rate flexibility, nor does it aim to achieve any specific exchange rate value. It pursues objectives compatible with the strategy of inflation targeting. The central bank intervenes in the foreign exchange market to (1) increase international reserves to reduce external vulnerabilities and improve access to foreign credit, (2) mitigate fluctuations in the exchange rate that do not clearly reflect the behavior of fundamental economic variables and that may have an adverse impact on inflation and economic activity, and (3) provide foreign liquidity to the market to ensure normal functioning of internal and external payments.
To guarantee the compatibility of foreign exchange intervention with the inflation-targeting strategy, purchases and sales of US dollars are sterilized as needed to stabilize short-term interest rates at a level that the central bank’s board considers coherent with its inflation and output objectives.3,4 This means that monetary expansions or contractions, generated by purchases or sales of US dollars, are compensated to avoid deviations of the overnight banking indicator interest rate (“IBR” in Spanish) from the level prescribed by the board.
An intervention decision considers its benefits and its costs for the country, as well as its effect on the central bank’s financial position. Holding international reserves and sterilizing reserve purchases is costly, especially when the domestic interest rate is higher than the return on reserves, which is the case for most emerging market economies.
This chapter first describes the foreign exchange purchase mechanisms used by the central bank in the period of the flexible exchange rate regime, and it explains their rationale and operation. It then characterizes the instruments used to sterilize the monetary effects of those purchases, as well as the effects of the sterilized intervention on the central bank’s balance sheet. Next, the current decision-making process to buy reserves is described, and the criteria for both the timing and the intensity of the purchases is specified. The next section turns to sales of international reserves, and it focuses on both the mechanisms used and the current decision-making process. A short survey of the effectiveness of foreign exchange sterilized intervention in Colombia is presented at the end of the chapter.
Purchases of Reserves and Volatility Options
On September 25, 1999, Colombia shifted from a crawling band to a flexible exchange rate regime, and in December of that year, it committed to a stabilization program with the IMF, amid a difficult macroeconomic environment. Part of that program was to restore international reserves to an adequate level. The central bank then started to offer put options to satisfy this requirement.
Put Options to Accumulate Reserves
Put options to accumulate reserves was adopted based on Mexico’s experience. The Colombian central bank auctioned these options at the end of each month (from November 1999 to September 2002, March to August 2003, March to May 2008) to banks, financial corporations, and the Ministry of Finance (this group is known as intermediaries of exchange rate options).5 These were US options with one-month maturity and a strike price equal to the representative market Colombian peso/US dollar exchange rate (TRM in Spanish).6 Given that the objective was solely to accumulate reserves, agents could only exercise the options when the TRM was below its 20-day average. The central bank could then avoid buying US dollars when the Colombian peso was weaker than the previous 20 days.
This is a suitable mechanism for countries that need to accumulate reserves and want to minimize the effect of reserve accumulation on the exchange rate. Through these auctions, the central bank bought $3.4 billion.7 All 49 auctions were oversubscribed, and the minimum and maximum amounts offered were $30 million and $250 million, respectively. During the first years, the amount auctioned was guided by the goals established in the agreement with the IMF. In October 2002, the central bank suspended this mechanism to avoid potential pressure on the exchange rate, given the depreciation of the peso (23 percent from April to October 2002).
From December 2002 to January 2004, the central bank carried out five of these auctions (denoted by blue stars in Figure 9.1), but not monthly or at the end of the month. Monthly auctions were reactivated in March 2004, because the appreciation of the peso was seen as transitory. At that moment, additional accumulation of international reserves was considered prudent to face any possible sudden reversals in capital flows or overadjustments in the exchange rate that could affect future inflation behavior.
Colombian Central Bank Foreign Exchange Intervention, 1999–2017
(Millions of US dollars, left scale; US dollar/Colombian peso official exchange rate, right scale)
Source: Central Bank of Colombia.Note: Data are from September of each year.Colombian Central Bank Foreign Exchange Intervention, 1999–2017
(Millions of US dollars, left scale; US dollar/Colombian peso official exchange rate, right scale)
Source: Central Bank of Colombia.Note: Data are from September of each year.Colombian Central Bank Foreign Exchange Intervention, 1999–2017
(Millions of US dollars, left scale; US dollar/Colombian peso official exchange rate, right scale)
Source: Central Bank of Colombia.Note: Data are from September of each year.When the central bank decided to shift to a flexible exchange rate regime, it introduced an intervention mechanism, known as volatility options, to provide coverage mechanisms to agents against extreme exchange rate movements. At that moment, the central bank considered the market unprepared to offer hedges under extreme circumstances.
Volatility Options
The central bank publicly committed itself to offer, through auctions, call (put) options with one-month maturity to intermediaries of exchange rate options, whenever the TRM was higher (lower) than its 20-day moving average plus (or minus) 5 percent.8 The threshold took into account a low probability of activation. The strike price was the TRM, and the condition for being able to exercise the option was the same as for the activation of the auction.
Considering that the objective was to mitigate an extreme fluctuation, the amount auctioned was large in comparison to the market. From 1999 until October 2001, it was $120 million.9 The central bank changed the offered amount to $180 million, given the increase in the volume of the foreign exchange market, and the threshold to 4 percent, since 5 percent seemed too extreme, given the behavior of the peso under the flexible exchange rate regime.
The first volatility auction (call options) was on July 29, 2002. Agents exercised all options in two days. On August 1, 2002, the condition for activation was met, and the central bank offered $180 million through call options.10 The first put volatility auction was held in December 2004.
The central bank again changed the threshold to 2 percent in December 2005. With such a small threshold, it had to do 5 call and 11 put option auctions. Discretionally, when the condition was reached but the maturity of the central bank’s options had not expired, the central bank did four additional call and put option auctions.
This mechanism was suspended from June to October 2008 and from October 2009 to October 2011, because the central bank was buying US dollars daily (as explained later). In February 2012, when the central bank resumed buying US dollars daily, it put off the mechanism again, which at that moment had a threshold of 4 percent (Table 9.1) and an amount of $200 million (20 percent of daily traded volume).
Volatility Auctions
The central bank was buying reserves daily.
Volatility Auctions
| Period | Trigger (%) | No. of Obliged Call Auctions | No. of Obliged Put Options | No. of Discretionary Call Options | No. of Discretionary Put Options |
|---|---|---|---|---|---|
| Nov. 1999–Oct. 2001 | 5 | 0 | 0 | n/a | n/a |
| Oct. 2001–Dec. 2005 | 4 | 2 | 1 | 1 | 0 |
| Dec. 2005–Jun. 2008 | 2 | 5 | 11 | 4 | 4 |
| Jun. 2005-Oct. 20081 | n/a | n/a | n/a | n/a | n/a |
| Oct. 2008–Oct. 2009 | 5 | 2 | 5 | 3 | 0 |
| Oct. 2009–Oct. 20111 | n/a | n/a | n/a | n/a | n/a |
| Oct. 2011–Feb. 2012 | 4 | 0 | 0 | n/a | n/a |
The central bank was buying reserves daily.
Volatility Auctions
| Period | Trigger (%) | No. of Obliged Call Auctions | No. of Obliged Put Options | No. of Discretionary Call Options | No. of Discretionary Put Options |
|---|---|---|---|---|---|
| Nov. 1999–Oct. 2001 | 5 | 0 | 0 | n/a | n/a |
| Oct. 2001–Dec. 2005 | 4 | 2 | 1 | 1 | 0 |
| Dec. 2005–Jun. 2008 | 2 | 5 | 11 | 4 | 4 |
| Jun. 2005-Oct. 20081 | n/a | n/a | n/a | n/a | n/a |
| Oct. 2008–Oct. 2009 | 5 | 2 | 5 | 3 | 0 |
| Oct. 2009–Oct. 20111 | n/a | n/a | n/a | n/a | n/a |
| Oct. 2011–Feb. 2012 | 4 | 0 | 0 | n/a | n/a |
The central bank was buying reserves daily.
The last volatility option was conducted under a slightly different mechanism: only the volatility call option instrument was turned on.11 The Colombian peso was at its weakest value historically, and therefore the central bank did not want to have to buy reserves at all. This variant was implemented in November 2015, after the peso had depreciated substantially in response to the collapse in the oil price.12 The purpose of the mechanism was to mitigate expectations of depreciation and inflation, as the latter significantly exceeded the target, and to provide liquidity to the market under an extreme depreciation.13,14
The central bank established the threshold at 7 percent. To impact expectations, the amount offered was $500 million (50 percent of daily traded volume). Later, the threshold changed to 5 percent and then to 3 percent. At this last threshold the condition was achieved. On May 20, 2016, the central bank offered $500 million in call options. Agents demanded $411 million and exercised $256 million. The central bank deactivated the mechanism at the end of May, when liquidity in the market was adequate.
To sum up, volatility options seek to mitigate extreme short-term exchange rate movements, but they are not meant to moderate trends. If policymakers activate this mechanism with the aim of intervening in the market, they will have the tendency to lower the threshold when the intervention has not been activated.
From 1999 to 2003, Colombia’s de facto regime was classified as independently floating, according to the IMF.15 Since 2004, it has been in the floating category. 16 One of the IMF’s reasons to change Colombia’s de facto classification was likely the shift from rules-based mechanisms to discretionary purchases of US dollars.
Discretionary Purchases
Despite the accumulation of reserves between December 2003 and September 2004 ($1.4 billion), the Colombian peso appreciated 8 percent. The central bank therefore opted for another mechanism, discretionary intervention, to accumulate reserves to mitigate the appreciation trend. On September 17, 2004, the central bank announced that it was going to buy up to $1.0 billion by the end of the year, directly in the market (as another participant) and through put options for accumulation of reserves, considering the negative effects that a stronger peso could have on some tradable sectors.17 On December 21, 2004, after an extraordinary board meeting, the central bank ratified its discretionary purchases without mentioning any limit. From the first announcement until the end of 2004, the central bank bought $1.3 billion directly in the market (it did not auction put options). The central bank continued to buy $5.9 billion using this mechanism until March 1, 2006, when the peso started to depreciate rapidly after the US Federal Reserve increased interest rates. In press releases from the meetings of the board, until March 2006, the central bank reiterated its decision to continue intervening discretionally in the foreign exchange market.
The central bank resumed its discretionary purchases on January 15, 2007 (Figure 9.1), after an appreciation of the peso of 16 percent from July 2006.18 At the end of January, the central bank announced that it was going to intervene massively in the exchange market to confront the transitory pressures arising from the sale of public sector assets. It bought $4.5 billion until April 2007, while the peso appreciated 5 percent. In a single day, purchases were $732 million (89 percent of the daily average traded volume in March 2007).
After all this active discretionary intervention, the central bank decided to return to announced mechanisms. It concluded that trying to defend an exchange rate level was not feasible, and the mechanism gave a sense that that was precisely what the central bank was doing.19 On June 24, 2008, the central bank started to accumulate reserves through daily auctions of spot purchases of small amounts in comparison to the market.
Daily discretionary purchases were publicly disclosed on April 2013.20 During the discretionary purchases, starting in February 2007, the public had access, with a 1-month lag, to the information on the quantity bought by the central bank during each month. This mechanism demanded a lot of time from staff and board.
Auctions of Spot Purchases
This mechanism was adopted based on Chile’s experience. On June 20, 2008, when the Colombian peso was at its highest value historically, the central bank announced that it was changing its mechanism of accumulating reserves, from monthly put options of $150 million (12 auctions starting in March 2008) to daily auctions of purchases of $20 million.21 The central bank said the change reinforced the policy of accumulation of reserves to face an eventual deterioration of the international environment, and that the measure took advantage of an appreciated peso that seemed unsustainable.
Given the latter, with its intervention, the central bank wanted to have an effect on the exchange rate level. This seemed more likely with the daily auctions of spot purchases than with the put options. As already mentioned, agents could only exercise these put options when the peso was appreciating in comparison to the previous 20 days. Therefore, under a depreciation trend, the central bank would not be able to buy US dollars with the put option mechanism with the spot auctions it was buying every day.
On October 6, 2008, the central bank stopped purchasing US dollars through spot auctions when the peso depreciated after the financial turmoil in the United States. These auctions were reactivated in March 2010 to limit the country’s risk profile by increasing international reserves. From March 2010 to December 2014, the central bank accumulated $22,450 million (47 percent of actual reserves; see Figure 9.2, panel 1) through this mechanism.22 Daily purchases were small in comparison to the traded volume (on average $942 million). The maximum and minimum amounts were $50 million and $5 million, respectively. Usually, the central bank announced the approximate amount it was going to buy. For example, in September 2010, the board meeting press release said that the central bank was going to purchase at least $20 million daily, for at least the next 4 months. Purchases were suspended on December 11 of 2014, given the decline in the price of oil and the ensuing large depreciation of the peso (Figure 9.1).
Central Bank’s Indicators, 1999–2017
Note: IBR = Indicador Bancario de Referencia (Banking Reference Interest Rate).Central Bank’s Indicators, 1999–2017
Note: IBR = Indicador Bancario de Referencia (Banking Reference Interest Rate).Central Bank’s Indicators, 1999–2017
Note: IBR = Indicador Bancario de Referencia (Banking Reference Interest Rate).Purchases of Reserves: Sterilization Instruments and the Effect on the Central Bank’s Balance Sheet
Sterilization mechanisms to contract the monetary base are a key element of the central bank’s inflation targeting, so that it is able to buy international reserves, as needed, without inducing deviation of the short-term interest rate from the policy rate. The central bank might otherwise compromise the inflation target with the purchases of foreign currency. Indeed, it faced this challenge in 2007, because it was buying reserves and did not have enough securities to be sold in the market (see Figure 9.2, panel 2, on holdings of public bonds), and the law prohibited it to issue its own bonds. As a sterilization mechanism, the central bank thus offered short-term deposits, which paid an interest rate close to the policy rate. These term deposits did not attract as many resources as needed because of the lack of a secondary market. As a result, the short-term interbank overnight rate was below the policy rate during some periods (Figure 9.2, panel 3).
In 2009, the law allowed the central bank to issue its own bonds, and it added a new sterilization instrument to the toolkit. So far, it has not used it, because in 2012 it signed a memorandum of understanding with the Ministry of Finance by which the latter would issue short-term government bonds and would deposit the proceeds in the central bank, to reduce the monetary base.23 The central bank would remunerate the Ministry of Finance’s account and, therefore, would carry with the cost of sterilization. The advantage of this mechanism is that the country has only one risk-free issuer; that is, the central bank and the Ministry of Finance are not both setting prices for public bonds. The agreement also states that sterilization bonds will have, at most, a 3-year maturity. This was a concern for the central bank, considering that shorter bonds (lower market risk) could attract more foreign investors. Time has shown that foreign investors are more averse to liquidity than to market risk, and sterilization bonds are not as liquid as longer-term bonds.
Sterilization with bonds affects banks’ balance sheets. According to Vargas, González, and Rodriguez (2013), if sterilized foreign exchange intervention is effective because of portfolio balance channel effects, it may also have an expansionary effect on credit supply and aggregate demand. The authors assume that banks have an optimal composition between public bonds and loans.
The sterilization of the purchases of reserves has costs that affect the central bank’s balance sheet. From 2012 to 2015, the central bank produced losses as a result of the low return on international reserves and the higher cost of sterilization (Figure 9.2, panel 4). Given that, by law, the government has to assume the central bank’s losses, the central bank feared that its independence might be affected. However, this did not happen.
How to Decide the Amount and Timing of Purchases
Until 2012, there was an important debate among the central bank’s board and its staff on how many reserves to buy and when. Models of optimal reserves were not very useful, considering their high sensitivity to model parameter values (see Técnica 2014). The central bank decided to establish a framework as a basis for the discussion.
Given that there is no agreement on an adequate level of reserves, the framework takes into account the most common indicators in this regard: (1) reserves/imports, (2) reserves/(1-year external amortizations and the current account deficit), (3) reserves/(M3 and private agents’ holdings of public bonds and equities), and (4) reserves/GDP.24 The third indicator was modified 3 years ago, to take into account that residents and nonresidents (private agents) can leave the country through the liquidation of their financial investments (previously it was reserves/M3).25 The fourth indicator is included because it is widely used for comparisons across countries. It is less relevant, however, because the vulnerability to external shocks does not depend as much on the size of the economy as it does on a country’s trade and financial integration with the rest of the world.26
Once a set of indicators is defined, how should their adequate values be determined? This is a question for which there are no clear-cut answers. The central bank decided to take two target levels for each indicator. First, its average from 2003 onward (starting when the floating exchange rate regime was consolidated—a period in which Colombia has not experienced an external crisis). Second, the 2008 average, the year of the global financial crisis, in which Colombia did well. The central bank takes the simple average of these eight targets to determine the adequate level of reserves (Figure 9.3).
International Reserves Indicators
Source: Central Bank of Colombia.Note: TES = títulos de tesorería (Treasury bills in Colombia).International Reserves Indicators
Source: Central Bank of Colombia.Note: TES = títulos de tesorería (Treasury bills in Colombia).International Reserves Indicators
Source: Central Bank of Colombia.Note: TES = títulos de tesorería (Treasury bills in Colombia).If the result of the framework’s calculation suggests that the central bank should accumulate reserves, and authorities desire to confront the possibly unsustainable appreciation dynamics, then purchases should be carried out during periods when the Colombian peso/US dollar exchange rate is below an estimate of its equilibrium level. The further it is from this level, the faster the pace of the purchases.
To estimate the equilibrium of the real exchange rate for intervention, the central bank uses behavioral equilibrium exchange rate models. These relate the multilateral real exchange rate with its short- and long-term fundamentals. Significant deviations of the real exchange rate from its fundamentals may reflect speculative behavior in the exchange rate market, which the sterilized intervention of the central bank might be able to correct.27
Sales of Reserves
In addition to the sale of US dollars though volatility options, the central bank has also used call options as a mechanism to deaccumulate reserves. Call options were used from February to April 2003, when the peso was depreciating, to complement interest rate hikes and contain rising inflation expectations, after a substantial depreciation of the currency had taken place.28 These options work in the same way as the put options for accumulation but can only be exercised when the TRM is above its 20-day average. Similar to put options for accumulation, call options to deaccumulate are not suitable to face an exchange rate trend.
Beyond its possible contribution to reach or maintain the inflation target, sales of international reserves can be used to help preserve financial stability. At present, central bank staff follow the decision tree shown in Figure 9.4 to determine whether the central bank should sell reserves with this aim when the peso is depreciating, and to suggest to the board which mechanism to use.
This decision tree recommends intervening through foreign exchange swaps when financial stability is threatened by closure of external credit lines (the central bank sells US dollars spot and buys US dollars forward). Through this mechanism, intermediaries of exchange rate options (central bank counterparties in the foreign exchange swap) are not changing their foreign exposure. Therefore, the central bank’s foreign exchange swaps will have a limited effect on the exchange rate, but will allow intermediaries to provide credit in foreign currency without exchange rate risk.29, 30 This mechanism was regulated in June 2015 and has not been used so far. Since it should not prevent agents from searching external credit lines, the central bank will auction the forward contract with a maximum accepted price that is lower than the market’s.
The use of the other mechanisms, sales of US dollars through nondeliverable forwards and spot sales, depends on the level of reserves and on the functioning of the derivatives market.31 One of the advantages of nondeliverable forwards is that they do not affect the level of reserves. In many instances, the decision to pursue sterilized intervention depends on its perceived benefits, and specifically, on its effectiveness to affect the exchange rate. This is a key question in any decision tree.
Effectiveness of Intervention
Many papers analyze the effectiveness of the central bank’s foreign exchange intervention to affect the level of the exchange rate or reduce its volatility (Table 9.2). Most studies use a two-stage estimation model to avoid endogeneity issues, in which a first equation describes the behavior of the official foreign interventions, and a second explains the exchange rate percentage changes.32 Others implement techniques based on event studies.
Effectiveness of Sterilized Intervention: Nonexhaustive Evidence for Colombia, Selected Studies, 2008–17
Effectiveness of Sterilized Intervention: Nonexhaustive Evidence for Colombia, Selected Studies, 2008–17
| Literature | Period | Type of Intervention | Method | Was the Intervention Effective? | How Long Was It Effective? |
|---|---|---|---|---|---|
| Kamil 2008 | Sep. 2004-Mar. 2006 Jan.-Apr. 2007 (daily) | Discretionary purchases | Two-stage Tobit-GARCH | Yes, during the first period (the central bank was reducing the interest rate); $30 million generate 0.23 percent depreciation; reduces volatility No, during the second period (the central bank was raising the interest rate) | Contemporary |
| Echavarría, López, and Misas 2009 | Jan. 2000-Aug. 2008 (monthly) | Options, volatility options, and discretionary purchases | SVAR | Yes (no size effect mentioned), given that interventions at the beginning of the sample might have not been sterilized entirely, and that, in some cases, intervention announced an expansive future monetary policy | 1 month |
| Castro and Toro 2010 | 1993–2010 (daily) | All interventions (no differentiation) | Two-stage GARCH | Yes, but only between 2008 and 2010, because of the interaction between interventions and capital controls (75 days); purchases of $1 million depreciated 0.008 percent; in all other periods, volatility increases | Contemporary |
| Murciaand Rojas 2014 | May-Sep. 2012 (intraday) | Daily purchases through auctions | EGARCH | Not in the exchange rate level 3 minutes after intervention; volatility increases | |
| Echavarría, Melo, and Villamizar 2014 | 2000–12 (daily) | Options, volatility options, and discretionary purchases | Event study | Volatility options only (direction, smoothing, reversion, matching) | Up to 10 days |
| Fuentes and others 2014 | May 2007-Nov. 2011 (intraday) | Daily purchases through auctions | Event regressions | No (neither in level nor in volatility) | |
| Durán-Vanegas 2015 | June 2008-Dec 2013 (daily) | Volatility options and purchases through daily auctions | Two-stage ARCH-GARCH | Yes, when the exchange rate is appreciated by more than 5.1 percent against purchasing power parity measures; $1 million generate 0.0002 percent depreciation | Contemporary |
| Villamizar 2015 | 1999–2012 (daily) | Options and discretionary purchases | Two-stage multi-variate Tobit | Level: no; volatility decreases; $1 million decrease volatility by 0.005 percent | 3 weeks |
| Ocampo and Malagón 2015 | 2006–13 (monthly) | All interventions (no differentiation) | OLS and VAR | Yes, in the real exchange rate, when capital controls (unremunerated reserve requirements) are in place. One standard deviation shock in the interaction produces a depreciation of 0.63 percent after 8 months | From the 6th month to the 13th month |
| Kuersteiner, Phillips, and Villamizar 2016 | 2002–12 (daily) | Volatility options | Discontinuous regressions | Put options of $180 million generate 2 percent depreciation; call options of $180 million generate 3 percent appreciation after 2 weeks and a contemporary depreciation of 0.3 percent | 3 weeks 2 weeks (from day 5 to day 15) |
| Echavarría, Melo, and Villamizar 2017 | 2004–12 (daily) | Discretionary purchases and through daily auctions | Two-stage Tobit-GARCH | $1 million through daily auctions (discretionary purchases) depreciate 0.004 percent (0.001 percent) | Contemporary |
Effectiveness of Sterilized Intervention: Nonexhaustive Evidence for Colombia, Selected Studies, 2008–17
| Literature | Period | Type of Intervention | Method | Was the Intervention Effective? | How Long Was It Effective? |
|---|---|---|---|---|---|
| Kamil 2008 | Sep. 2004-Mar. 2006 Jan.-Apr. 2007 (daily) | Discretionary purchases | Two-stage Tobit-GARCH | Yes, during the first period (the central bank was reducing the interest rate); $30 million generate 0.23 percent depreciation; reduces volatility No, during the second period (the central bank was raising the interest rate) | Contemporary |
| Echavarría, López, and Misas 2009 | Jan. 2000-Aug. 2008 (monthly) | Options, volatility options, and discretionary purchases | SVAR | Yes (no size effect mentioned), given that interventions at the beginning of the sample might have not been sterilized entirely, and that, in some cases, intervention announced an expansive future monetary policy | 1 month |
| Castro and Toro 2010 | 1993–2010 (daily) | All interventions (no differentiation) | Two-stage GARCH | Yes, but only between 2008 and 2010, because of the interaction between interventions and capital controls (75 days); purchases of $1 million depreciated 0.008 percent; in all other periods, volatility increases | Contemporary |
| Murciaand Rojas 2014 | May-Sep. 2012 (intraday) | Daily purchases through auctions | EGARCH | Not in the exchange rate level 3 minutes after intervention; volatility increases | |
| Echavarría, Melo, and Villamizar 2014 | 2000–12 (daily) | Options, volatility options, and discretionary purchases | Event study | Volatility options only (direction, smoothing, reversion, matching) | Up to 10 days |
| Fuentes and others 2014 | May 2007-Nov. 2011 (intraday) | Daily purchases through auctions | Event regressions | No (neither in level nor in volatility) | |
| Durán-Vanegas 2015 | June 2008-Dec 2013 (daily) | Volatility options and purchases through daily auctions | Two-stage ARCH-GARCH | Yes, when the exchange rate is appreciated by more than 5.1 percent against purchasing power parity measures; $1 million generate 0.0002 percent depreciation | Contemporary |
| Villamizar 2015 | 1999–2012 (daily) | Options and discretionary purchases | Two-stage multi-variate Tobit | Level: no; volatility decreases; $1 million decrease volatility by 0.005 percent | 3 weeks |
| Ocampo and Malagón 2015 | 2006–13 (monthly) | All interventions (no differentiation) | OLS and VAR | Yes, in the real exchange rate, when capital controls (unremunerated reserve requirements) are in place. One standard deviation shock in the interaction produces a depreciation of 0.63 percent after 8 months | From the 6th month to the 13th month |
| Kuersteiner, Phillips, and Villamizar 2016 | 2002–12 (daily) | Volatility options | Discontinuous regressions | Put options of $180 million generate 2 percent depreciation; call options of $180 million generate 3 percent appreciation after 2 weeks and a contemporary depreciation of 0.3 percent | 3 weeks 2 weeks (from day 5 to day 15) |
| Echavarría, Melo, and Villamizar 2017 | 2004–12 (daily) | Discretionary purchases and through daily auctions | Two-stage Tobit-GARCH | $1 million through daily auctions (discretionary purchases) depreciate 0.004 percent (0.001 percent) | Contemporary |
The majority of these papers use daily data. Conclusions vary. In purchases through daily auctions, Echavarría, Melo, and Villamizar (2017) find a bigger depreciation effect compared to discretionary purchases, but Fuentes and others (2014) report that daily auctions do not affect the level of the exchange rate after some minutes have passed. Murcia and Rojas (2014) reached a similar conclusion. The biggest effect is found with volatility put options: purchases of $1 million depreciated the currency by 0.01 percent for 3 weeks. Most likely, this small and short-lived effect did not have a significant impact on key economic sectors. This result is not surprising; after all, the possibility of Colombian authorities stabilizing the exchange rate is limited, given that the country chose an independent monetary policy and a relative open capital account (the impossible trinity).
Conclusion
Exchange rate flexibility is fundamental for an economy to maintain low and stable inflation and production levels close to their potential value. However, the Colombian central bank intervenes in the foreign exchange market without aiming to achieve any specific exchange rate value to (1) accumulate reserves to reduce external vulnerability and improve access conditions to foreign credit, (2) mitigate fluctuations in the exchange rate that do not clearly reflect the behavior of fundamental economic variables, and that may have an adverse impact on inflation and economic activity, and (3) provide foreign liquidity to the market to ensure the normal functioning of internal and external payments. Regarding the second motivation, the literature that has studied the central bank’s intervention finds that the effect on the exchange rate is small, if any. Therefore, the probability of impacting inflation and economic activity is also small.
The central bank has used different intervention mechanisms, and it always sterilizes its foreign exchange interventions to keep the short-term interest rate aligned with the policy rate. Sterilization mechanisms are key elements for central banks under an inflation-targeting regime, allowing them to buy and sell reserves as needed without compromising the inflation target. Sterilization of purchases of reserves is costly.
The use of put options is a suitable mechanism for countries that need to accumulate reserves and that want to minimize the effect of reserve accumulation on exchange rates. Daily purchases of reserves might be more adequate when central banks want to influence the exchange rate. In the Colombian central bank’s experience, daily purchases did not generate a sense that it was defending an exchange rate level, but discretionary purchases did. This occurs because with the daily purchases, the amounts of the intervention were the same regardless of the value of the currency. If a perceived exchange rate goal is judged as nonattainable by market participants, additional capital inflows may be attracted, rendering the foreign exchange intervention ineffective, possibly introducing unwarranted volatility to the exchange rate, and imposing greater costs on the central bank, especially if the latter reacts by increasing intervention.
Volatility options should be designed to provide a hedge under extreme circumstances, and not to control the normal volatility of the exchange rate. Volatility options should therefore have large thresholds and might be appropriate for countries with shallow derivatives markets. The disadvantage is that they might discourage the development of the derivatives market. Overall, the central bank used options to implement trigger rules for spot intervention (there was no underlying motive to use derivatives other than to “wire” the rule for the spot intervention). Foreign exchange swaps are the right mechanism to use when external credit lines have been closed. To decide how many reserves to buy or sell, it is useful to have an established framework, at least as a starting point for the discussion.
References
Durán-Vanegas, Juan D. 2015. “Do Foreign Exchange Interventions Work as Coordinating Signals in Colombia?” Ensayos Sobre Política Económica 33 (78): 169–75.
Echavarría, Juan J., Enrique López, and Martha Misas. 2009. “Intervenciones cambiarias y política monetaria en Colombia. Un análisis de var estructural.” Borradores de Economía 580, Banco de la República, Bogotá.
Echavarría, Juan J., Luis F. Melo, and Mauricio Villamizar. 2014. “The Impact of Foreign Exchange Intervention in Colombia: An Event Study Approach.” Revista Desarrollo y Sociedad 73.
Echavarría, Juan J., Luis F. Melo, and Mauricio Villamizar. 2017. “The Impact of Pre-announced Day-to-Day Interventions on the Colombian Exchange Rate.” Empirical Economics 55 (3), 1319–36.
Echavarría, Juan J., Diego Vásquez, and Mauricio Villamizar. 2010. “Impacto de las intervenciones cambiarias sobre el nivel y la volatilidad de la tasa de cambio en Colombia.” Ensayos sobre Política Económica 28 (62).
Fuentes, Miguel, Pablo Pincheira, Juan M. Julio, Hernán Rincón-Castro, Santiago García-Verdú, Miguel Zerecero, Marco Vega, Erick Lahura, and Ramon Moreno. 2014. “The Effects of Intraday Foreign Exchange Market Operations in Latin America: Results for Chile, Colombia, Mexico, and Peru.” Borradores de Economía 849, Banco de la República, Bogotá.
Inter-American Development Bank. 2016. Time to Act: Latin America and the Caribbean Facing Strong Challenges. 2016 Latin American and Caribbean Macroeconomic Report, p. 19.
Kamil, Herman. 2008. “Is Central Bank Intervention Effective under Inflation Targeting Regimes? The Case of Colombia.” IMF Working Paper 08/88, Washington, DC: International Monetary Fund, Washington, DC.
Kuersteiner, Guido M., David C. Phillips, and Mauricio Villamizar. 2016. “The Effects of Foreign Exchange Intervention: Evidence from a Rule-Based Policy in Colombia.” Borradores de Economía 964, Banco de la República, Bogotá.
Mandeng, Ousmene J. 2003. “Central Bank Foreign Exchange Market Intervention and Option Contract Specification: The Case of Colombia.” IMF Working Papers 03/135, International Monetary Fund, Washington, DC.
Murcia, Andrés, and Diego Rojas. 2014. “Determinantes de la tasa de cambio en Colombia: un enfoque de microestructura de mercados.” Ensayos sobre Política Económica 32 (74): 52–67.
Ocampo, A José., and Jonathan Malagón. 2015. “Colombian Monetary and Exchange Rate Policy over the Past Decade.” Comparative Economic Studies 57 (3): 454–82.
Rincón, Hernán, and Jorge Toro. 2010. “Are Capital Controls and Central Bank Intervention Effective?” Borradores de Economía 625, Banco de la República, Bogotá.
Técnica, Gerencia. 2014. “Optimum and Adequate Level of International Reserves.” Borradores de Economía 727, Banco de la República, Bogotá.
Toro, Jorge, and Juan M. Julio. 2005. “Efectividad de la intervención discrecional del Banco de la República en el mercado cambiario.” Borradores de Economía 336, Banco de la República, Bogotá.
Uribe, D José., and Jorge Toro. 2005. “Foreign Exchange Market Intervention in Colombia.” In Foreign Exchange Market Intervention in Emerging Markets: Motives, Techniques and Implications. BIS Papers No. 24 Bank for International Settlements, Basel.
Vargas, Hernando, Andrés González, and Diego Rodriguez. 2013. “Foreign Exchange Intervention in Colombia.” Borradores de Economía 783, Banco de la República, Bogotá.
Villamizar, Mauricio. 2015. “Identifying the Effects of Simultaneous Monetary Policy Shocks.” Contemporary Economic Policy, 34: 2.
See http://www.banrep.gov.co/es/politica-intervencion-cambiaria (similar text available in Spanish).
As an illustration, the Inter-American Development Bank (2016) compares the response of Ecuador, which is a dollarized economy, and Colombia to the drop in oil prices. Tis commodity is very important for both countries.
The strategy has been in place since 1999.
Purchases and sales of US dollars were not fully sterilized during the first years of the inflation-targeting regime.
This group also includes finance companies and financial cooperatives that have larger-than-required technical equity to establish a financial corporation.
The TRM is calculated as the weighted (by quantity) average of Colombian peso/US dollar spot operations of the previous day.
The central bank announced exercised options daily.
The central bank was committed only if the maturity of the last options had expired.
The objective was to intervene with 50 percent of the market volume on the days the options were “in the money.” The data showed that past options would have been, on average, 4 days in the money.
This auction was under the central bank’s discretion, given that the maturity of the previous options had not expired.
For simplicity, the central bank presented this auction as a call option for the deaccumulation of reserves.
From June 2014 to October 2015, the peso depreciated 53 percent in response to the decline of the price of oil (56 percent).
Expectations stood at 4.12 percent, the highest since February 2012, while the target was 3 percent.
Some liquidity indicators (such as market depth and bid-ask spreads—not the traded amount) showed an important deterioration.
“The exchange rate is market-determined, with any official foreign exchange market intervention aimed at moderating the rate of change and preventing undue fluctuations in the exchange rate, rather than at establishing a level for it. In these regimes, in principle, the authorities may pursue an independent monetary policy.” See https://www.imf.org/external/np/mfd/er/2003/eng/0603.htm.
“A floating exchange rate is largely market determined, without an ascertainable or predictable path for the rate. Foreign exchange market intervention may be either direct or indirect, and such intervention serves to moderate the rate of change and prevent undue fluctuations in the exchange rate, but policies targeting a specific level of the exchange rate are incompatible with floating. Indicators for managing the rate are broadly judgmental (for example, balance of payments position, international reserves, parallel market developments). Floating arrangements may exhibit more or less exchange rate volatility, depending on the size of the shocks affecting the economy.”
The transactional market is semiblind (customers know their counterparty only when the transaction is executed).
In the press release of December 2006, the central bank reiterated its commitment to intervene in the foreign exchange market without affecting the achievement of the inflation target.
Internally, the central bank decided the exchange rate levels at which intervention should be done. They changed over time.
The auction lasts 3 minutes and is competitive. The details of the auction can be found at http://www.banrep.gov.co/sites/default/fles/reglamentacion/archivos/Compendio_CRE_DODM_143_27_may_2016.pdf (in Spanish). All of the central bank’s auctions are a uniform price.
International reserves increased $39.6 billion, and the central bank’s net purchases were $38.4 billion between October 1999 and November 2017.
The issuance of these bonds was authorized in Colombia’s National Development Plans for 2010–14 and for 2014–18. The authorization must be included in each National Development Plan.
This is known as TES.
Since 2014, foreign portfolio investment has been an important source of Colombia’s capital account.
The central bank applied the IMF method to the Colombian case to define the country’s reserve adequacy. Tis indicator is used for comparison purposes only, because of the short length of the time series used for estimation. The estimation period starts in 2003, when the flexible exchange rate regime was consolidated. It covers the global financial crisis but does not include a crisis in Colombia.
The methods for estimating the equilibrium multilateral real exchange rate, based on purchasing power parity, show parity measures in the very long term, beyond the horizon and effectiveness of the exchange rate policy. The fundamental equilibrium exchange rate models, on the other hand, relate the equilibrium real exchange rate to measures of the “sustainable” current account deficit. This hinders its use as a tool to detect short-term misalignments, since they reflect imbalances of savings and investment that are not likely to be corrected by sterilized intervention.
There were three auctions of $200 million each. All options were exercised.
On the asset side of the balance sheet, they have US dollars, and on the liabilities side, the sale of US dollar forwards.
The credit will be hedged with the sale of US dollar forwards to the central bank.
These have not been regulated.
Foreign interventions might react to exchange rate movements as long as they influence its behavior.
