Chapter 8 Interventions in Chile
Author:
Catalina Larraín
Search for other papers by Catalina Larraín in
Current site
Google Scholar
PubMedClose
Language:
English
Keywords:
BOOK; central bank; intervention; exchange rate; foreign exchange intervention; spot market; intervention policy; dollar sale; foreign exchange; Exchange rates; Foreign exchange intervention; Currency markets; International reserves; Global; Central America; inflation-targeting regime; intervention program; central bank data; dollar-denominated bond; central bank communication; Inflation targeting; Nominal effective exchange rate; Exchange rate arrangements
Download PDF (381.4 KB)

Abstract

Foreign exchange intervention is widely used as a policy tool, particularly in emerging markets, but many facets of this tool remain limited, especially in the context of flexible exchange rate regimes. The Latin American experience can be informative because some of its largest countries adopted floating exchange rate regimes and inflation targeting while continuing to intervene in foreign exchange markets. This edited volume reviews detailed accounts from several Latin American countries’ central banks, and it provides insight into how and with what aim many interventions were decided and implemented. This book documents the effectiveness of intervention and pays special attention to the role of foreign exchange intervention policy within inflation-targeting monetary frameworks. The main lesson from Latin America’s foreign exchange interventions, in the context of inflation targeting, is that the region has had a considerable degree of success. Transparency and a clear communication policy have been key. For economies that are not highly dollarized, rules-based intervention helped contain financial instability and build international reserves while preserving inflation targets. The Latin American experience can help other countries in the design and implementation of their policies.

In this chapter, we study the four episodes of foreign exchange interventions since the inflation-targeting regime was adopted in Chile, and we present the reasons and the communication strategy that the Central Bank of Chile followed. We ran event studies with day and intraday data to assess the effects of the announcements and actual interventions on the exchange rate level, trend, and volatility. The results suggest that announcements have clearer and more significant effects than the actual interventions. For the announcements, the results in level and trend of the exchange rate go in the expected direction and are bigger in the purchases programs than in the sales programs.

Introduction

Before Chile adopted a full-fledged inflation-targeting regime, in September 1999, its government had tried several exchange rate regimes: flexible, fixed, crawling pegs, and multiple exchange rates. Since 1999, the country has had a flexible exchange regime that, alongside a healthy financial system and responsible macroeconomic policies, is a pillar of inflation targeting. In this context, foreign exchange interventions have been remarkable exceptions—indeed, there were only four of them.

This chapter analyzes what motivated these interventions and their effectiveness; it describes the strategy of the Central Bank of Chile and reveals how it communicated that strategy to the market. To evaluate the drivers of intervention, the analysis relies on the official communiqués of the Central Bank of Chile; and to evaluate the effectiveness, it conducts event studies using daily and intraday data.

The evidence indicates that the announcements themselves, rather than the actual interventions, had an effect. The announcements had short-run effects on the nominal exchange rate and its trend, and that went in the expected direction. The four announcements and the actual sales interventions also tended to increase the volatility of the nominal exchange rate in the few days after the announcement of intervention.

The next section analyzes foreign exchange policies in the 1990s as back-ground on Chile’s move toward inflation targeting. The chapter then describes the four interventions in Chile since 1999 and conducts event study regressions to assess their effectiveness. Central bank communications strategy follows, and the final sections summarize the main lessons and conclusions.

The Road to Inflation Targeting

After decades of experimenting with fixed and multiple exchange rates, as well as crawling pegs, the Central Bank of Chile abandoned the exchange rate as a policy goal in 1999, which left the inflation rate as the only target. Describing in detail the exchange rate regimes before the adoption of the inflation target (and exchange rate flexibility) goes beyond the scope of this chapter. But some exchange rate background of the 1980s and 1990s provides the context in which this adoption took place.

Chile was hit by a financial and banking crisis in 1982 that required the abandonment of the fixed exchange rate adopted in 1979. The economic policy objective in the remaining years of the 1980s was to reconstruct the financial system, which demanded bailouts of the private sector from the central bank and the government. In this context, inflation averaged 20 percent until the end of the decade, and the exchange rate followed a crawling peg to avoid significant real currency appreciation.

In 1989, the Central Bank of Chile became autonomous, and it has used two policy targets since then. First, the exchange rate was allowed to fluctuate within a defined band, and second, an inflation target was announced yearly in September (until 1998). Chile’s inflation was gradually reduced from 23 percent at the beginning of the decade to 3.2 percent in late 1999.

In the early 1990s, amid capital inflows, the Chilean peso appreciated. As Figure 8.1 shows, the nominal exchange rate was often at the lower bound of the target band, and the monetary authority continuously widened the band in response, changed the central parity, and intervened in the foreign exchange market. These actions, and the fact that inflation remained near the target in those years, are anecdotal evidence that the inflation target prevailed over the exchange rate whenever they came into conflict.

Figure 8.1.
Figure 8.1.

Exchange Rate Band, 1990–99

(Chilean pesos per US dollar, observed Central Bank of Chile)

Source: Central Bank of Chile.

During the Asian crisis in 1997, pressure to appreciate the local currency vanished, and depreciation pressures were met with foreign exchange interventions and tight monetary policy. Despite depreciation after the Asian crisis, the exchange rate was still at the lower bound of its band in early 1998. The central bank feared that the sharp depreciation of the peso threatened the inflation target set for that year, so it intervened in the foreign exchange market and raised its benchmark interest rate.

In June 1998, the exchange rate was still in the lower part of the band, despite a 10.8 percent depreciation since October 1997. The central bank decided to narrow the band—from 12.5 percent above and below the center, to 3 percent above and 2.5 percent below the center—to curb depreciation expectations. At the same time, the central bank confirmed its commitment to the inflation target for the end of the year. The central bank coped with a narrower band and the external shock coming from the Russian financial crisis (which increased the cost of borrowing and reduced Chilean terms of trade), using an increase in interest rates, which exacerbated the contractionary effect of the foreign shock. It became more difficult to remain inside the band, and the central bank decided to widen it again to 7 percent (total width) and to continue widening it in subsequent months. The tightening of monetary policy was then gradually relaxed in the following months, because the economy was entering a recession, and no inflationary pressures were coming from peso depreciation.

In September 1999, when uncertainty declined, the central bank announced the end of the exchange rate band and adopted an inflation-targeting regime. During the exchange rate band regime, the central bank intervened whenever it deemed it necessary, without any notice.

The foreign exchange interventions implied an increase in foreign reserves until late 1997, and it reached its maximum value in October of that year. In June 1998, after narrowing the band, the national currency was depreciated, and the central bank sold reserves to keep the exchange rate within the band. These interventions lasted until September 1999, when the band regime was ended. Figure 8.2 presents the reserves-to-GDP ratio, which fluctuated around 20 percent during the 1990s, since both international reserves and GDP grew during that decade.

Figure 8.2.
Figure 8.2.

Monthly Reserves over Yearly GDP, 1990–2011

(Reserves/GDP)

Source: Central Bank of Chile.

During the last period of the bands (1998–99), as the Chilean economy was recovering, and the central bank widened the band, interventions became less frequent and were concentrated between January 1998 and March 1999. Total foreign currency sold by the central bank reached $4.83 billion during 43 days of intervention (equivalent to 10 percent of the days of this period). The daily average amount sold was $109 million in 1998, and $39 million in 1999. These amounts varied considerably, with small interventions of $5 million and larger, to above $350 million. Also, the central bank did intervene with central bank dollar-denominated bonds (BCDs), when its sales concentrated between July 1998 and April 1999, and reached $1,440 million on 60 different days (Morandé and Tapia 2002).

On reaching full inflation targeting, the monetary authority announced annual inflation targets in September of each year, until 1998. In September 1999, the central bank announced the annual target for the following year, along with an indefinite target that ranged from 2 percent to 4 percent starting in 2001.1 Since then, keeping inflation at its target has been the sole commitment of the central bank.

The inflation-targeting regime adopted free flotation for the exchange rate and a comprehensive communication strategy to enhance policy transparency and credibility. Since April 2001, the main policy instrument has been the short-term nominal interest rate, and the target has been to keep inflation between 2 percent and 4 percent on a 12–24 month horizon. As the central bank recognized in its release of September 1999, a floating currency should give the central bank more autonomy to manage monetary policy and to increase its capacity to deal with external shocks. Since then, the Central Bank of Chile has been strongly committed to exchange rate flexibility and to the inflation-targeting regime.

Nevertheless, including in its announcement of September 1999, the central bank explicitly stated that it reserved the option to intervene in the foreign exchange market, but only in exceptional and qualified cases, and that it would report such decisions. In that communiqué, the central bank stated that it would provide biweekly (but with some delay) disaggregated information on the holdings of international reserves.

Foreign Exchange Interventions Since the Adoption of Inflation Targeting

During inflation targeting, inflation expectations have been anchored around the target most of the time, and foreign exchange interventions have been exceptional episodes.

As noted, since September 1999 the central bank has intervened only four times: in 2001, 2002–03, 2008, and 2011. In 2001 and 2002–03, it intervened to provide foreign currency liquidity (by selling US dollars and dollar-denominated bonds) as a result of overreactions in the foreign exchange market, mainly because of instabilities in neighboring countries. In 2008 and early 2011, the central bank announced foreign exchange programs to increase the availability of international liquidity (by purchases of US dollars).

Details of the four interventions are discussed in the next few pages.

First Intervention: August 16, 2001

From January to the beginning of August 2001, the currency depreciated a cumulative 20 percent—from Ch$574 to Ch$689, per dollar—in part because of the situation in Argentina and then exacerbated by the events of 9/11 in the United States. The central bank intervened, and its communiqué stated that “the complex financial situations experienced by some economies in the region […] have caused instability in financial markets. This has been manifested in the exchange rate, without affecting, however, the perception of risk of the rest of the world on the Chilean economy […] the succession of movements in the same direction have exacerbated the depreciation and volatility of the peso, concerning financial markets.”2

The intervention program consisted of selling foreign currency in the spot market and increasing the sales of BCDs. The communiqué only specified the maximum amount of sales, stating that sales could occur up to $2 billion for the spot market and up to the same amount for the US dollar-denominated debt.

Spot market interventions totaled $803 million, less than half the maximum announced, and 5.2 percent of the stock of foreign reserves in December. Instead, the total sales of BCD reached $2.58 billion, from August 16 to the end of 2001.3

Second Intervention: October 10, 2002

The exchange rate depreciated 5.2 percent in September 2002, but with no major deterioration in its fundamentals, except for commotion from Brazil around the presidential elections of 2002. In the communiqué, the central bank stated that “during the last weeks, there has been an additional deterioration in the financial situation faced by some economies in the region […] the depreciation of the exchange rate in recent months is the expected reaction […] However, part of the recently observed depreciation reflects movements in a market that has operated with low liquidity and high volatility, and, consequently, with some degree of overreaction […] the Board of the central bank deems it prudent to provide a greater supply of currency hedging instruments and liquidity in foreign currency, in order to facilitate the orderly operation of the foreign exchange market.”4

The central bank announced a period of sales of foreign currency and an increase in the supply of BCDs, with $2 billion in spot market and $2 billion more of BCDs, but it also announced that the duration of the program would be four months (until February 10, 2003).

In the end, however, the central bank did not intervene in the spot market, and only issued US dollar-denominated debt. Between October 10 and February 10, the central bank issued debt equivalent to $1,440 million (less than the amount previously announced). It issued $360 million in October, $480 million in November, and $300 million in December and January (until February 10).5

This second episode involved much less intervention than the first one. Reserves did not change (because there was no intervention in the spot market), and sales of BCDs were 55 percent of the sales in the former intervention.

Third Intervention: April 10, 2008

The third intervention program was intended to increase the stock of international reserves, which the central bank considered insufficient to face a risky external scenario. It decided to intervene once the peso had appreciated 6.9 percent since the beginning of 2008.

The communiqué stated the main reason for the intervention: “This increase in the stock of international reserves will leave the Chilean economy better prepared to face the eventuality of an additional, severe and abrupt deterioration of the external environment […] [This measure] is also consistent with the evaluation that the real exchange rate is currently below the level that should prevail once real and financial conditions in the world economy return to normal.”6

On this occasion, the central bank gave more details of the program. It announced that it would increase the stock of international reserves by $8 billion. The program would start on April 14 and end on December 12. It also stated that it would announce the specific purchase program, which could be modified according to market conditions, starting with daily purchases of $50 million of foreign currency, which was maintained, unmodified, until the end.

The program ended earlier than planned, on September 29, because of the collapse of Lehman Brothers, when the central bank had increased its international liquidity position by $5.75 billion (equivalent to 70 percent of the initial announced amount).7 Figure 8.2 shows that the reserves-to-GDP ratio increased by 4.4 percentage points (from 9 to 13 percent).

Fourth Intervention: January 3, 2011

At the time of the fourth intervention, the currency had appreciated by 14.4 percent in the last six months of 2010. In this scenario, the central bank stated: “The international economy shows an unusual configuration, characterized by high commodity prices, low interest rates, slow recovery of developed economies, and depreciation of the US dollar. On the other hand, emerging economies are growing vigorously and their currencies are appreciating […] There is also the risk that the difference in the growth rate between advanced and emerging economies will be maintained beyond expectations, which could intensify the exchange pressures.”8 To strengthen their international liquidity position, the central bank board decided to initiate another foreign currency purchase program. In its announcement, it explicitly said that the objective was to bring international reserves to a level compatible with the values observed in similar economies, for precautionary reasons. It stated: “The availability of additional reserves will allow us to better face the eventuality of a significant deterioration of the external environment […] This program would lead the international liquidity position to 17% of GDP.”

The central bank announced that it would increase the level of reserves by $12 billion, with daily purchases of $50 million of foreign currency during the rest of the year. The announcement explained that the first stage of the program would consist of daily purchases of $50 million, which could be modified. However, the strategy was kept unmodified until the end of the program. As Figure 8.2 shows, when the program ended, these reserves had up to 17 percent of GDP, equivalent to an increase of $12 billion (daily purchases over 240 days).

There are other episodes that could arguably be termed interventions—in late 2008 and late 2017. The first corresponds to the aftermath of Lehman Brothers’ bankruptcy, when a series of US dollar swaps was offered to provide short-term liquidity in foreign currency: “The Central Bank of Chile will reactivate foreign exchange swap tenders as from tomorrow, when financial conditions will be announced prior to the opening of the market.”9 The second case corresponds to an offer of US dollar swaps (up to $500 million in each tender). The central bank board announced this new program “to facilitate the normal operation of the monetary market in exchange currency […] The program will be transitory, so it will be valid for a period of four weeks.”10 Nevertheless, in the end, these swaps were not used.

These two episodes did not involve changes in the net asset position in foreign currency of the central bank. In that sense, there were more liquidity measures than foreign exchange interventions, so this chapter does not study them (see Claro and Soto (2013) for a more in-depth discussion).

Intervention Impact on Exchange Rates

Several well-known identification issues are present when trying to estimate the effects of intervention on the exchange rate (such as the two-way causality mentioned in Daude, Levy-Yeyati, and Nagengast 2016), which preclude the proper identification of those effects.

One way of addressing these issues is to use high-frequency data to run event studies. This chapter follows that strategy to estimate the effects of announcements and actual interventions on the exchange rate, trends, and volatility during the four intervention episodes.

Other studies have analyzed the effect of the interventions in Chile on the exchange rate. For example, Tapia and Tokman (2003) conducted an event study with daily data for the interventions of 2001 and 2002. BIS (2014) ran an event study with intraday data to see the effectiveness of the interventions in 2008 and 2011, and Pincheira (2013), among other goals, analyzed the effect of the announcement and actual interventions of 2008 and 2011 on the level of exchange rate (it first looks at the intraday exchange rate around each announcement and then estimates the effect of announcements and actual interventions on daily exchange rate). Also, Contreras, Pistelli, and Saez (2013) also conducted an event study for several countries, including Chile in 2011. This chapter uses an identification strategy similar to those works.

We estimate the following regression:

Yt=β0+β1Postj+ϵt,(4.1)

where Y corresponds to the logarithm, the variation, or the standard deviation of the nominal exchange rate, respectively.11 The events correspond to announcements or actual interventions (sales or purchases), j corresponds to the period before and after the event, and the variable Post corresponds to a dummy that takes the value of 1 if the observation is after each episode, but inside the window of size j, and zero otherwise. So, β1 denotes the effect of each event on level, trend, or volatility of the nominal exchange rate.

Announcement Effects: Event Study with Daily Data

This section presents the results of the event study around each announcement date using daily data for different window sizes j = {1, 2, 3, 5, 10, and 15} days before and after the event. It also estimates the effects of the early ending of the third program, in September 2008.

This analysis starts by looking at the movements of the exchange rate around the announcement dates, presented in Figures 8.3 and 8.4. It then runs the event study regressions, with results in Table 8.1.

Figure 8.3.
Figure 8.3.

Interventions: Sale Announcements, 2001 and 2002

(Chilean pesos per US dollar, observed Central Bank of Chile)

Source: Central Bank of Chile.
Figure 8.4.
Figure 8.4.

Interventions: Purchase Announcements, 2008 and 2011

(Chilean pesos per US dollar, observed Central Bank of Chile)

Source: Central Bank of Chile.
Table 8.1.

Event Study of Announcement: Daily Exchange Rate

article image
article image
Source: Authors’ estimation based on Central Bank data. Note: The “Window Size” column reflects the number of days for each side of the announcement date. *p < 0.1; **p < 0.05; ***p < 0.01.

Figure 8.3 presents the analysis of the first two announcements of intervention (sales). Panel 1 shows that, after the first announcement, the exchange rate did appreciate, showing a change in its tendency. Apparently, the effect did not last long: the exchange rate appreciated by 3.9 percent during August 15–30, with the maximum daily appreciation of 1.8 percent on August 20. This effect was then offset by a depreciation in September (as a result of the 9/11 events) and October, followed by appreciations again in November and December. The total effect from the date of the announcement to the end of the program was an appreciation of 3.42 percent.

Panel 2 of Figure 8.3 shows that the tendency of the daily exchange rate also changed around the date of the second announcement. After that, the maximum level of appreciation was reached two months later (December 13, 6.97 percent), but that appreciation was almost completely undone in the following months. At the end of the program, the exchange rate had appreciated by only 0.5 percent.

Figure 8.4 shows the path of the exchange rate around the purchase announcements. Panel 1 shows that the tendency of the exchange rate around the third announcement of intervention did change, from appreciation to a persistent depreciation, so when the program ended, the exchange rate had reached 23.5 percent depreciation since its onset. Note that this program ended sooner than originally planned because of the beginning of the financial crisis (with the associated depreciation of the currency).

Panel 2 of Figure 8.4 shows that the movement around the last announcement was different. The exchange rate jumped, and then continued appreciating. Two days after the announcement, the exchange rate had depreciated by 4 percent, and the maximum percentage of depreciation was reached eight days later (6.5 percent).

Table 8.1 presents the results of the event study regressions using daily data. Jointly with the estimation of the coefficients (columns 3, 6, and 9), it presents the average of preannouncement levels of each dependent variable (columns 1, 4, and 7) and postannouncement levels (columns 2, 5, and 8).12

The sale announcements (programs I and II) in column 3 show that the change in the level of the exchange rate is positive for small windows, and it becomes negative for bigger ones. Thus, the exchange rate continued depreciating until a few days after each announcement of intervention, although the effects are not statistically significant.

Consistently, in column 6, which presents the difference in the average daily variation of the exchange rate, the exchange rate was depreciating before each announcement, and then started to appreciate, or at least smoothed its depreciating tendency (see window size 1 in program I). Also, in window sizes j = {5, 10} of program II, the exchange rate increased its tendency to appreciate after the announcement. These effects are significant in bigger window sizes.

The last definition of the dependent variable corresponds to the change in the volatility of the nominal exchange rate presented in column 9.13 An increase can be seen in the volatility in both sale announcements, but the difference becomes smaller as the window size increases.

Regarding purchase announcements (programs III and IV), column 3 shows that for windows greater than one, the difference in the level of the exchange rate is positive and generally increasing (except for the biggest window size). The same phenomena in column 6 refer to the difference in the average of daily variation, reflecting a reduction in the rate of appreciation or a shift to depreciation. Also, a positive but decreasing difference in the volatility of the nominal exchange rate can be seen in column 9.

For the announcement of the end of the third intervention program, the effect on the exchange rate level is positive (column 3), as it is the difference in daily variations (column 6), which indicates that the tendency of depreciation increased after the end of the program. Also, the volatility increased sharply after this announcement—however, those results are very likely related to the turmoil of the financial crisis.

This evidence suggests that intervention announcements have the expected directional impact on the level and trend of the nominal exchange rate, and the effects are bigger for the two purchase announcements. The evidence on volatility shows that the announcement increased, rather than reduced, the volatility of the currency, suggesting that these announcements could have introduced some uncertainty, which tends to decrease as window sizes increase (see Claro and Soto (2013) for further discussion).

Effects of the Last Two Announcements: Event Study with Intraday Data

For these announcements, intraday data on exchange rates allow better identification of the effect of an announcement by narrowing the window of analysis around an event. This analysis closely follows the one presented in BIS (2014)14 and in Pincheira (2013).

The announcements were published on the Chilean central bank’s website, followed by a simultaneous notification to commercial banks about the recent upload. Press releases from the central bank board also explained the program, usually on the day following the announcement. The timing of these purchase announcements was the following:

  • On April 10, 2008, the communiqué was uploaded at 7:15 p.m. The next day, the central bank governor published a press release at 11:00 a.m. announcing that the program of US dollar purchases would start on April 14.

  • On September 29, 2008, the communiqué to terminate the program was published at 5:30 p.m., announcing the immediate cessation of the intervention. US dollar purchases were discontinued then and there.

  • On January 3, 2011, the communiqué was published at 6 p.m. and the governor gave a press release on January 4 at 10:30 a.m. As explained, a program of US dollar purchases was announced, which was to start on January 5, 2011.

Figure 8.5 presents the results on the exchange rate level in a window size of j = {1} for the day before and after each event. The preannouncement exchange rate was taken 90 minutes before the announcement, and for the postannouncement rate, two different definitions were used to compute results: (1) the first observation after the announcement (between 10 and 20 minutes after, depending on the data available for each date), and (2) the observation at 8:15 a.m. of the following day.15

Figure 8.5.
Figure 8.5.

Intraday Effect, 2008 and 2011

(Chilean pesos per US dollar)

Source: Bolsa Electrónica de Chile.Note: The “1” represents the first observation after the announcement, so the effects are reflected on the orange line.

Figure 8.5 shows that at the moment of the announcement (in both cases, 2008 and 2011), the nominal exchange rate depreciated considerably. When the ending of the third program was announced, the exchange rate appreciated, as expected.

Between 90 minutes before the announcement and the first observation after the announcement, the exchange rate depreciated by 0.69 percent in April 2008, and by 4 percent in January 2011, and it appreciated by 2.58 percent in September 2008.

Also, because the announcements were made in the late afternoon, the same observation before the announcement (90 minutes before) is compared with the observation at 8:15 a.m. of the following day: The exchange rate depreciated by 3.43 percent in April 2008, and 3.68 percent in January 2011, and it appreciated by 2.39 percent in September 2008.

Event Study of US Dollar Sales in Spot and BCD Markets (2001 and 2002–03)

This section assesses the effect of actual interventions by running a daily event study (defined in regression 4.1) around the days of sale of foreign currency and BCDs. The number of observations is small, so the significance has to be taken with care.

The analysis is not done with daily data for purchases, because they took place every day, which reduces the capacity to identify the effects of a particular operation. As explained below, in those cases, the analysis runs event studies with intraday data. The analysis is done for each type of sale (spot and BCD), for window size j = {1, 3} days and for the dependent variables: level, trend, and volatility of the nominal exchange rate (Table 8.2).16

Table 8.2.

Event Study of Spot and (Central Bank Dollar-Denominated Bond) Sales: One-Day Window Size

article image
Source: Authors’ estimation is based on central bank data. Note: The “Window Size” column reflects the number of days for each side of the sale operations date. *p<.1; ***p<.01.

To improve the identification capacity of the exercise, the sample is restricted to those interventions with at least two days between them, in the case of a one-day window, and four days for a three-day window.17 In that sample, 54 dates of sales are in the one-day window size, and 12 are in the three-day window size.

Table 8.2 shows that the level and the daily variation effects were small and not statistically significant; they also varied across different window sizes. For the volatility, it is found that, if anything, it increased after the selling operations of BCDs in 2001.

The small impact on the level and the trend of the exchange rate is not surprising, given the intervention strategy of these programs, in which the central bank gave information about the interventions days before starting them. These findings are consistent with Tapia and Tokman (2003), who determined that the effects of the interventions came from the announcements and not from the actual interventions.

Event Study of US Dollar Purchases (2008 and 2011)

This section assesses the effectiveness of the actual purchases of US dollars on the programs of 2008 and 2011. It runs the event study regression (4.1) with the same three specifications of the dependent variable as before. As these programs consist of daily purchases, it relies on the intraday data of the nominal exchange rate to estimate the event study, and it runs the previous regression separately for each day of intervention.

Tables 8.3–8.5 report the average of the estimated coefficients, by year and by quarter.

Table 8.3.

Event Study of Purchases, 2008 and 2011: Level

article image
Source: Authors’ estimation based on central bank data. Note: Values shown are β1 (standard errors). The “Window Size” column reflects the amount of time before 9 a.m. and after 1 p.m. Intraminute variation was too small, so now variation corresponds to the variation between the first and last observation per window. *p < .1; **p < .05.
Table 8.4.

Event Study of Purchases, 2008 and 2011: Variation

article image
Source: Authors’ estimation based on central bank data. Note: Values shown are β1 (standard errors). The “Window Size” column reflects the amount of time before 9 a.m. and after 1 p.m. Intraminute variation was too small, so now variation corresponds to the variation between the first and last observation per window. *p < .1; **p < .05.
Table 8.5.

Event Study of Purchases 2008 and 2011: Volatility

article image
Source: Authors’ estimation based on central bank data. Note: Values shown are β1 (standard errors). The “Window Size” column reflects the amount of time before 9 a.m. and after 1 p.m. The intraminute variation was too small, so now variation corresponds to the variation between the first and last observation per window. *p < .1; **p < .05; ***p < .01.

The purchases were made between 9 a.m. and 1 p.m., so observations before 9 a.m. are compared with the ones after 1 p.m., using different window sizes j = {10 minutes, 1 hour, 2 hours, and all the observations until the end of the day}.18

Table 8.3 shows the percentage change in the nominal exchange rate. The effects are small and statistically significant only for some cases in 2008, where the coefficients have the expected sign. Table 8.4 shows that the effects on the variation of the exchange rate are again not significantly different from zero.

For exchange rate volatility, Table 8.5 shows negative and significant effects for 2008, but results are less clear for 2011, since the signs vary across different window sizes. This is consistent with BIS (2014), which found no significant effects on the trend of the exchange rate when considering a narrow time period around the actual interventions.

Communication

Central bank communication of the interventions to the market varied across regimes. When the exchange rate band was operating, the central bank did not communicate the spot interventions to the market. However, since the central bank adopted inflation targeting, it communicates the interventions, although their content varied across the four intervention episodes.

The first two interventions established the maximum reserves to be sold, but only the second one announced the duration of the program. In neither of them did the central bank communicate the daily spot or BCD interventions, but it published the foreign reserves two weeks after the interventions occurred. The two subsequent episodes, however, explicitly established the amount to be purchased, the starting date and the duration, along with the daily schedule.

In addition, for the last two episodes, the central bank made it clear that it would reserve the right to alter the program whenever it deemed necessary. As we already explained, that was the case in September 2008. The communiqué explicitly stated that “[the program] could be modified according to market conditions” and the one in 2011 stated that “[stages of the program] could revise them according to market conditions.”

Also, the central bank board communicated to the market the main reason(s) for each intervention, which, in the first two cases, were related to an overreaction of the level and volatility of the currency because of financial stress in other economies in the region. In the last two cases, the interventions were related mainly to an accumulation of reserves for precautionary reasons.

In addition, the four communiqués highlight the commitment of the central bank to its inflation target. In the first two announcements, the central bank said that the board reiterates its confidence in the Chilean economy and in the effectiveness of the macroeconomic policy framework. In the last two announcements, it was even more explicit in its commitment to the inflation-targeting regime, stating in 2008 that “this exceptional measure is coherent with the overall conduct of monetary policy, which aims to keep inflation at 3% most of the time, within a tolerance range of ± 1%.” In 2011 it ended by saying: “The Board of the Central Bank reiterates its commitment to the conduct of monetary policy, based on an inflationary goal and exchange rate flotation […] our flotation scheme allows interventions in exceptional situations.”

Lessons from Chile

Implementation of the free-floating regime made the inflation target more credible and the economy more resilient to external shocks. Interventions are a risk for an inflation-targeting regime, because if there are too many of them, central banks can lose credibility on their commitment to the target. One of the main lessons of the Chilean experience is that interventions need to be rare and occur only in extraordinary circumstances.

The method of communicating interventions is also relevant. Transparency and clear public explanation is crucial. If interventions are not transparent, the temptation to intervene more frequently increases. This transparency, without conditioning it to the evolution of the exchange rate, reflects the commitment of the Central Bank of Chile to exchange rate flexibility.

The central bank on its website states that “such interventions are transparent, sound, and exceptional measures. When they occur, the timings and amounts involved are explicitly defined, and the reasons are clearly stated.”19

Conclusion

After decades of experimenting with fixed, multiple exchange rates and crawling pegs, in 1999 the Central Bank of Chile abandoned exchange rates as a policy goal, which left the inflation rate as the only target.

Chile has a full inflation-targeting regime that rests on a central bank commitment to exchange rate flexibility. Since this adoption, the central bank has intervened in the foreign exchange market four times. The first two (2001 and 2002–03) responded to an overreaction of the exchange rate and its volatility because of difficult situations in neighboring countries, and the last two (2008 and 2011) responded to a desire to build up the stock of foreign reserves for precautionary reasons.

In each case, interventions were preannounced, along with the maximum amount of sales/purchases in spot markets and BCD, the beginning and ending dates of the intervention period (except for the first episode of intervention), and a daily schedule for the program in the last two cases.

The evidence indicates that the announcements have a short-run effect on the nominal exchange rate and trend, and those go in the expected direction, with bigger effects for the last two announcements (purchases). Also, no significant effect of actual interventions was found on the level and trend of the nominal exchange rate. For volatility, the analysis found that, if anything, the four announcements and the actual selling interventions tended to increase the volatility of the nominal exchange rate.

References

  • BIS (Bank for International Settlements). 2014. “The Effects of Intraday Foreign Exchange Market Operations in Latin America: Results for Chile, Colombia, Mexico, and Peru.” BIS Working Paper 462, Bank for International Settlements, Basel.

    • Search Google Scholar
    • Export Citation

  • Claro, Sebastian, and Claudio Soto. 2013. “Exchange Rate Policy and Exchange Rate Interventions: The Chilean Experience.” BIS Working Paper 73, Bank for International Settlements, Basel.

    • Search Google Scholar
    • Export Citation

  • Contreras, Gabriela M., Alfredo M. Pistelli, and Camila Sáez. 2013. “Efecto de las intervenciones cambiarias recientes en economías emergentes.” Notas de Investigación Journal Economía Chilena 16 (1): 122–37.

    • Search Google Scholar
    • Export Citation

  • Daude, Christian, Eduardo Levy-Yeyati, and Arne J. Nagengast. 2016. “On the Effectiveness of Exchange Rate Interventions in Emerging Markets.” Journal of International Money and Finance 64: 239–61.

    • Search Google Scholar
    • Export Citation

  • Morandé, Felipe G., and Matías Tapia. 2002. “Exchange Rate Policy in Chile: From the Band to Floating and Beyond.” Working Paper 130, Stanford Center on Global Poverty and Development, Stanford, CA.

    • Search Google Scholar
    • Export Citation

  • Pincheira, P. 2013. “Interventions and Inflation Expectation in an Inflation Targeting Economy.” BIS Working Paper 427, Bank for International Settlements, Basel.

    • Search Google Scholar
    • Export Citation

  • Tapia, Matías, and Andrea Tokman. 2003. “Effects of Foreign Exchange Intervention under Public Information: The Chilean Case.” Estudios de Economía 30 (1): 21–53.

    • Search Google Scholar
    • Export Citation
1

Board announcements can be found at http://www.bcentral.cl/en/web/guest/notas-de-prensa.

2

Central bank communiqué, August 16, 2001.

3

This is the total amount of BCD of the period, which includes issues from former programs.

4

Central bank communiqué, October 10, 2002.

5

The central bank had already announced that it would begin issuing BCDs on August 26, 2002; this listing takes into account the issues from Oct. 10 until February 10.

6

Central bank communiqué, April 10, 2008.

7

Central bank communiqué, September 29, 2008.

8

Central bank communiqué, January 3, 2011.

9

Central bank communiqué, September 29, 2008.

10

Central bank communiqué, December 15, 2017.

11

For the logarithm, the dependent variable is rescaled, so the coefficients show percentage changes.

12

The β1 coefficient is the difference of the average of the dependent variables before and after the event.

13

The coefficients in this last estimation are the difference in volatility between the post- and preannouncement, so we cannot report the significance.

14

The database of the intraday nominal exchange rate of La Bolsa Electrónica Chile was used, but keeping the frst observation per minute in the cases that had more than one observation in a minute.

15

This is because of different data availability for the different days of announcements.

16

The exception is for the one-day window size, where the effect on volatility is not estimated, because it is only one day before and after each intervention.

17

For the three-day window size, only sales of BCDs in 2001 and 2002–03 are presented, because all spot sales took place in less than four days between them. The regressions were run separately for BCDs and spot dates.

18

The windows are truncated from the left because they start when the market opens.

Within Same Series

Other IMF Content

Other Publishers Content

Contributor Notes

The opinions expressed in this chapter are the sole responsibility of the authors and are not necessarily those of the Central Bank of Chile.
  • Collapse
  • Expand
Cover Foreign Exchange Intervention in Inflation Targeters in Latin America
Foreign Exchange Intervention in Inflation Targeters in Latin America
Author:
Mr. Marcos d Chamon
,
Mr. David J Hofman
,
Mr. Nicolas E Magud
, and
Alejandro M. Werner
ISBN:
9781484375686
Publication Date:
28 Feb 2019
Publisher:
International Monetary Fund
DOI:
https://doi.org/10.5089/9781484375686.071
Pages:
320
  • <sup><xref ref-type="fig" rid="Rch08fig01">Figure 8.1</xref>.</sup>
    View in gallery
    Figure 8.1.

    Exchange Rate Band, 1990–99

    (Chilean pesos per US dollar, observed Central Bank of Chile)

  • <sup><xref ref-type="fig" rid="Rch08fig02">Figure 8.2</xref>.</sup>
    View in gallery
    Figure 8.2.

    Monthly Reserves over Yearly GDP, 1990–2011

    (Reserves/GDP)

  • <sup><xref ref-type="fig" rid="Rch08fig03">Figure 8.3</xref>.</sup>
    View in gallery
    Figure 8.3.

    Interventions: Sale Announcements, 2001 and 2002

    (Chilean pesos per US dollar, observed Central Bank of Chile)

  • <sup><xref ref-type="fig" rid="Rch08fig04">Figure 8.4</xref>.</sup>
    View in gallery
    Figure 8.4.

    Interventions: Purchase Announcements, 2008 and 2011

    (Chilean pesos per US dollar, observed Central Bank of Chile)

  • <sup><xref ref-type="fig" rid="Rch08fig05">Figure 8.5</xref>.</sup>
    View in gallery
    Figure 8.5.

    Intraday Effect, 2008 and 2011

    (Chilean pesos per US dollar)

Figure 8.1.

Exchange Rate Band, 1990–99

(Chilean pesos per US dollar, observed Central Bank of Chile)
Figure 8.2.

Monthly Reserves over Yearly GDP, 1990–2011

(Reserves/GDP)
Figure 8.3.

Interventions: Sale Announcements, 2001 and 2002

(Chilean pesos per US dollar, observed Central Bank of Chile)
Figure 8.4.

Interventions: Purchase Announcements, 2008 and 2011

(Chilean pesos per US dollar, observed Central Bank of Chile)
Figure 8.5.

Intraday Effect, 2008 and 2011

(Chilean pesos per US dollar)