## _wp06140

## Source details

**Canonical URL:** [_wp06140](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06140.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06140.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06140.pdf.json)

---

### Summary and purpose
- Examines whether Botswana’s actual real exchange rate deviates from its sustainable equilibrium using the behavioral equilibrium exchange rate (BEER) approach.
- Motivation: successive devaluations of the pula and the move from a fixed to a crawling peg exchange rate regime suggest potential misalignment with fundamentals.

### Macroeconomic performance and external/fiscal positions
- Real GDP growth averaged over 8 percent for two decades.
- Inflation: "has been kept broadly to single digits except during the 1992–94 recession."
- Per capita income in 2003: over US$4,800.
- Foreign reserves built to over US$5.5 billion, equivalent of about 18½ months of imports.
- Fiscal balance: surplus until 2000/01; external current account surplus sizable due to diamond exports.
- 1992–94 recession:
  - Real GDP declined by 0.1 percent.
  - Formal sector employment fell by about 6 percent.
  - Primary reason: low demand for diamonds.
- Privatization of the public pension fund reduced foreign reserves to some extent.

### Structure, concentration risks, and competitiveness
- Mining (including diamonds): 35–40 percent of GDP and 75 percent of total exports in recent years.
- Other mining: copper, nickel, soda ash, coal, and gold make small but noticeable contributions.
- Risks identified:
  - Potential overvaluation of the pula from large natural resource exports (Dutch disease).
  - Balassa-Samuelson effect: productivity differential between tradables and nontradables may cause external REER appreciation.
- Diversification efforts (tourism, livestock) have faced challenges; dependence on diamonds unchanged over several decades.

### Exchange rate history and policy actions
- February 2004 devaluation: the pula was devalued by 7.5 percent against the currency basket.
- May 2005 devaluation: the pula was devalued by 12 percent.
- Post-May 2005: crawling peg arrangement announced.
- Spread between buy and sell rates widened from ±0.125 percent to ±0.5 percent to encourage interbank market development.
- Long-term nominal movements (relative to 1980 levels, prior to February 2004 devaluation):
  - Depreciated by about 80 percent against the U.S. dollar.
  - Depreciated by 65 percent against the euro.
  - Appreciated by over 20 percent against the South African rand.
- Short-run movements:
  - 2001–2003: pula increased by about 50 percent against the U.S. dollar and doubled against the euro.
  - The two devaluations (February 2004 and May 2005) caused roughly a 20 percent decline by July 2005 in the value of the pula against the U.S. dollar, the euro, and the rand.
- Currency basket: comprises the South African rand and the SDR; precise weights not published.

### Analytical framework and econometric approach
- Approach: BEER to capture equilibrium behavior, cyclical and transitory deviations.
- Key fundamentals: Terms of trade (TOT); relative price of nontradables to tradables (ltnt); real interest differential (r-r*); net foreign assets (nfa); fiscal-risk premium (λ).
- Empirical model:
  - Reduced-form log specification: ln REER_t = f(r-r*_t, ltot_t, ltnt_t, nfa_t, λ_t).
  - Estimation method: vector error correction mechanism (VECM) with annual data for 1985–2004; lag length p set to one year.
- Data sources and construction:
  - REER: CPI-based multilateral REER (IMF Effective Exchange Rate Facility).
  - TOT: IMF WEO; ltnt from Bank of Botswana Annual Reports and Botswana Central Statistics Office.
  - Home real interest rate: CPI-deflated prime lending rate; foreign real rate: weighted average of long-term government bond yields deflated by CPI using REER weights.
  - nfa: ratio of net foreign assets to GDP (IMF IFS).
  - λ: difference between Botswana fiscal deficit and weighted average deficits of trading partners (ratio to GDP).

### Summary statistics (preserved exactly)
- lnREER: Mean 4.5; Std. Dev. 0.1; Min 4.4; Max 4.7
- r-r*: Mean -0.9; Std. Dev. 3.3; Min -6.4; Max 5.7
- ltot: Mean 4.6; Std. Dev. 0.3; Min 3.8; Max 4.8
- ltnt: Mean 0.006; Std. Dev. 0.012; Min -0.011; Max 0.032
- nfa: Mean 92.721; Std. Dev. 21.757; Min 57.6; Max 134.9
- λ: Mean -9.8; Std. Dev. 8.5; Min -24.8; Max 4.8
- Memorandum items:
  - Actual REER: Mean 92.3; Std. Dev. 7.4; Min 85.0; Max 111.0
  - Fiscal deficit for Botswana: Mean -6.4; Std. Dev. 7.5; Min 6.8; Max -19.9
  - Fiscal deficit for trading partners: Mean 3.4; Std. Dev. 1.7; Min 6.4; Max 0.9

### Estimation results and equilibrium relationships
- Unit-root tests: most variables I(1); first differences stationary.
- Johansen Trace test: one cointegrating relationship identified.
- Estimated long-run (cointegrating) equation (coefficients and standard errors preserved):
  - ln REER* = -0.289(0.024) + 1.348(0.018) ltnt - 0.030(0.013) ltot + 0.0013(0.0002) (r-r*) + 0.0004796(0.0007) nfa - 0.0007(0.0002) λ
- Coefficient interpretations:
  - r-r*: higher domestic real interest relative to foreign raises equilibrium REER.
  - ltot: terms-of-trade improvement increases REER (consistent with Dutch-disease influence).
  - ltnt: increase leads to significant external REER appreciation (Balassa-Samuelson effect).
  - λ: higher fiscal-risk premium leads to depreciation; economic effect very small.
  - nfa: unexpected negative coefficient; marginal economic impact; possible explanations include pension privatization and measurement issues.
- Alpha adjustment matrix (Coef.; Std. Err.):
  - lq: -0.093; 0.193
  - r-r*: 31.820; 9.642
  - ltot: 2.813; 0.997
  - ltnt: 0.100; 0.037
  - nfa: 66.376; 86.149
  - λ: 20.001; 34.182
- Most significant adjustments to disequilibria occur in the relative real interest rate and TOT equations.

### Sustainable equilibrium, misalignment, and recent dynamics
- Permanent values for fundamentals: Holt-Winters filtering; nfa uses five-year moving average.
- Selected estimated misalignment entries (index: 1980=100; positive values represent overvaluation; values preserved exactly):
  - 1985: Actual REER 82.2; Fitted 102.2; Sustainable 89.2; Misalignment -16.9
  - 1990: Actual REER 90.6; Fitted 84.2; Sustainable 98.7; Misalignment -8.2
  - 1995: Actual REER 91.5; Fitted 90.4; Sustainable 91.6; Misalignment -0.1
  - 2000: Actual REER 93.5; Fitted 91.4; Sustainable 91.2; Misalignment 2.3
  - 2001: Actual REER 106.6; Fitted 111.3; Sustainable 97.4; Misalignment 9.5
  - 2002: Actual REER 111.0; Fitted 112.7; Sustainable 102.0; Misalignment 8.8
  - 2003: Actual REER 111.0; Fitted 112.7; Sustainable 102.0; Misalignment 8.8
  - 2004: Actual REER 106.3; Fitted 109.7; Sustainable 105.8; Misalignment 0.4
- Summary conclusions:
  - Pula undervalued in the late 1980s.
  - Misalignment marginal in the 1990s.
  - Overvalued by 5 to 10 percent in 2000–2003.
  - 2004 shows realignment, possibly due to the February 2004 devaluation.

### Policy discussion and regime tradeoffs
- Evidence of recent overvaluation provides rationale for February 2004 and May 2005 devaluations and move to a crawling peg.
- Tradeoffs:
  - Fixed peg:
    - Merits: credible nominal anchor, historically low inflation.
    - Disadvantages: high volatility under real shocks, vulnerability to speculative attacks.
  - Crawling peg:
    - Advantages: can realign with shifting fundamentals and stabilize short-term exchange rate.
    - Disadvantages: no medium-term nominal anchor; risk of "galloping" inflation if adjustments aggressive.
- Operational recommendations for crawling peg:
  - Active crawling peg: preannounce rate of crawl for a year in advance; commit to announced path to stabilize expectations and induce gradual disinflation.
  - Risks: credibility problems if market drives rates to band edges; crawling band may be seen as weaker commitment than hard peg.
- Cross-country evidence cited:
  - Rogoff et al. (2003): "limited flexibility" regimes tended to achieve lower inflation, higher per capita GDP growth, and less annual growth volatility than pegged regimes over 1970-1999.
  - Bubula and Otker-Robe (2003): crisis incidence higher for conventional fixed regimes pegged to a basket than for crawling pegs during 1990-2001.
  - Ramcharan (2005): exogenous shocks have significant adverse impacts on developing countries with fixed regimes due to limited exchange rate flexibility.

### Policy implications, constraints, and recommendations
- Main empirical finding summary:
  - BEER (1985–2004): undervalued late 1980s; marginal misalignment in 1990s; overvalued by 5 to 10 percent in 2000–2003.
- Implementation requirements for crawling peg:
  - Comprehensive monetary and fiscal adjustments.
  - Acceleration of structural reforms to stimulate domestic private business and address high interest and unemployment rates.
- Constraints and risks:
  - Fiscal position expected to tighten due to HIV/AIDS, poverty, and high unemployment (>20 percent), limiting room for fiscal and interest-rate responses.
  - Need for credible operational framework and coordination across policy areas.
- Appendix managerial lessons and operational guidance:
  - Preconditions: mature interbank market and central bank responsibility for exchange rate policy implementation.
  - Recommended band width (Williamson, 1996): between ±7 and ±10 percent; Botswana’s arrangement at time of writing allowed ±0.5 percent (up from ±0.125 percent).
  - Suggested defensive tools: readiness to change interest rates and reserve requirements; allow movements outside the band under extreme pressure with commitment to restore band later; set depreciating crawl consistent with inflation target minus expected foreign inflation and productivity differentials.

### Appendix comparative indicators (preserved exactly for Botswana and comparator countries)
- Botswana (May 2005) — annual series 2002 2003 2004 2005:
  - GDP growth: 7.8 5.7 4.0 3.5
  - CPI: 8.0 9.3 6.9 8.6
  - Change in bilateral exchange rate against US$ (depreciation -): 27.7 23.1 3.8 -22.3
  - Change in REER (depreciation -): 4.7 3.1 -5.9 -9.5
  - Current account balance: 3.6 6.0 3.2 13.5
  - Overall budget balance (including grants): -4.1 -0.4 1.0 2.4
  - Total reserves (in months of imports): 30.3 23.0 18.4 19.4
  - Unemployment: 23.8 n.a. n.a. n.a.

- Chile (September 1982), Colombia (November 1991), Israel (December 1991), and Mexico (January 1989): full annual series and averaged outcomes after adoption are reported in the Appendix, with summary policy lessons including:
  - Stabilizing the real exchange rate under a crawling peg often involves short- to medium-term inflationary pressures that can delay achievement of low inflation for "five to ten years."
  - Restoring credibility may require coordinated tightening of monetary and fiscal policy and potentially discrete devaluations to shape expectations.
  - Structural reforms are critical where fiscal space is limited and unemployment is high.

*Source: _wp06140 (IMF working paper content provided).*

### 1. Summary Statistics ..................................................................................................

### _wp06140 - 1. Summary Statistics

### Introduction and purpose
- The paper examines whether Botswana’s actual real exchange rate deviates from its sustainable equilibrium level using the behavioral equilibrium exchange rate (BEER) approach.
- Motivation: successive devaluations of the pula and the move from a fixed to a crawling peg exchange rate regime suggest potential misalignment with fundamentals.

### Macroeconomic performance and fiscal/external positions
- Real GDP growth has averaged over 8 percent for two decades.
- Inflation has been kept broadly to single digits except during the 1992–94 recession.
- Per capita income in 2003 was estimated at over US$4,800.
- Foreign reserves were built to over US$5.5 billion, the equivalent of about 18½ months of imports.
- The fiscal balance continued to be in surplus until 2000/01; the external current account surplus has remained sizable due to diamond exports.
- Note on 1992–94 recession:
  - Real GDP declined by 0.1 percent.
  - Formal sector employment fell by about 6 percent.
  - Primary reason: low demand for diamonds during recession in industrial countries.
- Recent privatization of the public pension fund decreased foreign reserves to some extent.

### Structure of the economy and competitiveness concerns
- Mining (including diamonds) contributed 35–40 percent of GDP and 75 percent of total exports in recent years.
- Other mining: copper, nickel, soda ash, coal, and gold make small but noticeable contributions to GDP.
- Risks identified:
  - Potential overvaluation of the pula from large natural resource exports (Dutch disease).
  - Balassa-Samuelson effect: rapid growth in tradables relative to nontradables may lead to larger appreciation of the external real exchange rate.
- Diversification efforts (tourism, livestock) have faced challenges; dependence on diamonds has not changed much over several decades.

### Exchange rate policy actions and nominal exchange rate history
- February 2004 devaluation: the pula was devalued by 7.5 percent against the currency basket.
- May 2005 devaluation: the pula was devalued by 12 percent.
- Following May 2005, the exchange rate was announced to be adjusted in a crawling peg arrangement rather than in discrete steps.
- The spread between the buy and sell rates of the pula was widened from ±0.125 percent to ±0.5 percent to encourage interbank market development.
- Long-term nominal movements (relative to 1980 levels, prior to February 2004 devaluation):
  - Pula had depreciated by about 80 percent against the U.S. dollar.
  - Pula had depreciated by 65 percent against the euro.
  - Pula had appreciated by over 20 percent against the South African rand.
- Short-run movements:
  - An unusually sharp appreciation of the rand against the U.S. dollar starting in 2001 and continuing for three years led to a parallel hike of the pula against major currencies.
  - In 2002–03 the value of the pula increased by about 50 percent against the U.S. dollar and doubled against the euro.
  - The two devaluations (February 2004 and May 2005) caused roughly a 20 percent decline by July 2005 in the value of the pula against the U.S. dollar, the euro, and the rand.
- The currency basket comprises the South African rand and the SDR; the precise weights are not published.

### Analytical framing and policy relevance
- Although nominal exchange rate adjustments do not easily solve long-term external competitiveness issues, realignment can be justifiable because it influences competitiveness in the short to medium term.
- The Bank of Botswana (BoB) states that its objective of low inflation depends on domestic producers maintaining international competitiveness and contributing to broader national growth.
- The BEER approach is used to assess actual versus equilibrium real exchange rates and to inform whether the observed exchange rate requires realignment.

*Source: _wp06140 - 1. Summary Statistics (IMF PDF content provided).*

### Section II reviews earlier studies on exchange rate misalignment. Section III describes recent

### _wp06140 - Section II reviews earlier studies on exchange rate misalignment. Section III describes recent

### II. THE LITERATURE
- Three principal approaches to determining equilibrium exchange rates:
  - Purchasing Power Parity (PPP): relates nominal exchange rate to price differentials; examples cited include Johansen and Juselius (1992), MacDonald and Nagayasu (1998).
  - Fundamental Equilibrium Exchange Rate (FEER): determined by current account balance targets, underlying sustainable international asset positions, and national income at full employment; Williamson (1994) found the actual U.S. dollar was 14 percent overvalued and the Japanese yen 27 percent undervalued in late 1989.
  - Behavioral Equilibrium Exchange Rate (BEER): focuses on dynamic behavior including short-run deviations and broader macroeconomic fundamentals (Clark and MacDonald, 1998).
- This paper employs the BEER approach to capture equilibrium exchange rate behavior, including cyclical and transitory deviations, with key Botswana fundamentals identified as:
  - Terms of trade (TOT).
  - Dutch disease considerations (resource-export effects on nontradable/tradable prices).
  - Balassa-Samuelson effect (productivity differential between tradable and nontradable sectors).
- Emerging fiscal and fiscal-risk considerations for Botswana:
  - Growing fiscal pressures from large HIV/AIDS-related expenditures and the possibility that diamond revenues will plateau.
  - Reference: Moody’s Investment Services has rated Botswana’s sovereign bonds A1.
  - Botswana AIDS Impact Survey (2004) prevalence estimates: overall 17.3 percent for the general population aged 18 months and older; 33 percent for age group 25-29 years; 40.7 percent for 30-34; 36.9 percent for 35-39.
  - UNAIDS estimate at end-2003: 37.3 percent of the adult population (15-49).
  - Share of health expenditure in total government spending increased from 4 percent to over 10 percent over the past decade.
- Econometric choice:
  - Reduced-form single-equation model estimated using the vector error correction mechanism (VECM), following precedents such as MacDonald and Nagayasu (1998).

### III. RECENT DEVELOPMENTS IN EXCHANGE RATES IN BOTSWANA
A. External Real Effective Exchange Rate (REER)
- Definitions and measurement notes:
  - REER is CPI-based; NEER and REER definitions and formulae provided (footnote 12/13).
  - Exchange rates defined in foreign-currency terms: an increase in the exchange rate is an appreciation.
- Historical REER movements (1978–80 base; index: 1980=100 in figures):
  - 1980s: actual REER depreciated by about 15 percent.
  - Late 1990s through 2004 (before devaluations): the pula appreciated by approximately 25 percent.
  - February 2004 devaluation: nominal devaluation of 7.5 percent against the basket; estimated real effective impact 5.5 percent.
  - May 2005 devaluation: decreased the REER by 10.5 percent in a month.
- Inflation and REER link:
  - Botswana inflation "has remained relatively high at 6-9½ percent, though mostly it has stayed in single digits."
  - During 1998-2003: a 15 percent increase in Botswana’s relative prices led to REER appreciation by 25 percent and NEER increase by 10 percent.
- Figures referenced:
  - Figure 4: Botswana: Real Effective Exchange Rates (Monthly), 1980–2005 (Index: 1980=100).
  - Figure 5: Botswana: REER, NEER, and Relative Consumer Prices, 1980–2004 (Index: 1980=100).

B. Internal Real Exchange Rate (IRER) and Terms of Trade (TOT)
- Definitions:
  - IRER = P_T / P_NT (tradable price to nontradable price ratio); REER decomposed via IRER and trading-partner analogs (equations (2) and (3)).
  - TOT = P_EX / P_IM (exports relative to imports).
- Botswana-specific observations:
  - Relative price index of nontradables to tradables (IRER) has been very stable over two decades.
  - TOT index increased nearly fourfold between 1980 and 2000, driven by long-term declining international oil prices and increases in world commodity prices, including diamonds, copper, and nickel.
  - Recent hikes in fuel prices caused TOT deterioration, mitigating appreciation pressures on the external REER.
  - Commodity-specific note: diamond prices increased by 40 percent in the past two decades (BoB Annual Report 2004).
- Figure referenced:
  - Figure 6: Botswana: External and Internal Real Exchange Rates, and Terms of Trade, 1980–2004 (Index: 1985=100).
- Structural context:
  - In the consumption commodity basket, the share of nontradables (mainly services) is about 30 percent; domestic and imported tradables amount to 70 percent.
  - South Africa contributes about 75 percent of total imports for Botswana.

### IV. ANALYTICAL FRAMEWORK
A. Specifying an Empirical Model
- BEER specification for 1985-2004 period:
  - Reduced-form equation: ln REER_t = f(r-r*_t, ltot_t, ltnt_t, nfa_t, λ_t)  (Equation (4) notation).
  - Variables:
    - r-r*: real interest differential.
    - ltnt: relative price of nontradables to tradables (log).
    - nfa: net capital inflows proxy—rate of net foreign assets to GDP.
    - λ: risk premium factor proxied by government income and expenditure (difference between Botswana’s fiscal deficit and weighted average deficits of trading partners, expressed as ratio to GDP).
- Theoretical foundations:
  - Risk-adjusted interest parity and relative Fisher conditions lead to decomposition (Equation (5)), assuming expected future exchange rate converges to long-run equilibrium dependent on fundamentals.
- Empirical strategy:
  - VEC model with annual data; vector z' = (ln REER*, r-r*, ltot, ltnt, nfa, λ, constant).
  - Lag length p set to one year to conserve degrees of freedom.

B. Data
- Variable sources and measurement:
  - REER: multilateral CPI-based exchange rate of the pula (IMF Effective Exchange Rate Facility).
  - TOT: relative price of exports to imports (IMF WEO).
  - ltnt: computed from Bank of Botswana Annual Reports and Botswana Central Statistics Office.
  - Home real interest rate: CPI-deflated prime lending rate (long series available).
  - Foreign real interest rate: weighted average of long-term government bond yields deflated by CPI inflation, using same REER weights.
  - nfa: ratio of net foreign assets to GDP (IMF International Financial Statistics).
  - λ: difference between Botswana fiscal deficit and weighted average deficits of trading partners (ratio to GDP).
- Summary statistics (Table 1):
  - lnREER: Mean 4.5; Std. Dev. 0.1; Min 4.4; Max 4.7
  - r-r*: Mean -0.9; Std. Dev. 3.3; Min -6.4; Max 5.7
  - ltot: Mean 4.6; Std. Dev. 0.3; Min 3.8; Max 4.8
  - ltnt: Mean 0.006; Std. Dev. 0.012; Min -0.011; Max 0.032
  - nfa: Mean 92.721; Std. Dev. 21.757; Min 57.6; Max 134.9
  - λ: Mean -9.8; Std. Dev. 8.5; Min -24.8; Max 4.8
- Memorandum items:
  - Actual REER: Mean 92.3; Std. Dev. 7.4; Min 85.0; Max 111.0
  - Fiscal deficit for Botswana: Mean -6.4; Std. Dev. 7.5; Min 6.8; Max -19.9
  - Fiscal deficit for trading partners: Mean 3.4; Std. Dev. 1.7; Min 6.4; Max 0.9

### V. ESTIMATION RESULTS
- Stationarity and cointegration:
  - Unit-root tests (ADF and PP) indicate most variables are I(1); first differences are stationary (Table 2).
  - Johansen Trace test finds one cointegrating relationship (Table 3); null of no cointegration rejected, one cointegrating vector not rejected at 1 percent.
- Estimated long-run equilibrium (cointegrating) equation (Equation (7)):
  - ln REER* = -0.289(0.024) + 1.348(0.018) ltnt - 0.030(0.013) ltot + 0.0013(0.0002) (r-r*) + 0.0004796(0.0007) nfa - 0.0007(0.0002) λ
    - Note: coefficients and standard errors presented in the source (parentheses).
- Interpretation of coefficients:
  - Real interest differential: higher domestic real interest rate relative to foreign increases equilibrium REER.
  - Terms of trade (ltot): improvement increases REER (consistent with possible Dutch-disease influence).
  - Relative price of nontradables to tradables (ltnt): increase leads to significant external REER appreciation (Balassa-Samuelson effect).
  - Risk premium (λ): higher fiscal-risk premium leads to depreciation, but economic effect is very small.
  - Net foreign assets (nfa): unexpected negative coefficient; economic impact marginal. Possible explanations include privatization of public pension system reducing nfa ratio 2001–2004, and measurement issues.
- Alpha adjustment matrix (Table 4):
  - Coefficients (Coef.; Std. Err.):
    - lq: -0.093; 0.193
    - r-r*: 31.820; 9.642
    - ltot: 2.813; 0.997
    - ltnt: 0.100; 0.037
    - nfa: 66.376; 86.149
    - λ: 20.001; 34.182
  - Most significant adjustments to disequilibria occur in the relative real interest rate and TOT equations.
- Sustainable equilibrium computation:
  - Permanent values for fundamentals computed with Holt-Winters filtering; nfa uses five-year moving average.
- Estimated misalignment (Table 5; Actual REER, Fitted REER, Sustainable level, Misalignment percent):
  - Selected entries (index: 1980=100; Positive values represent overvaluation):
    - 1985: Actual REER 82.2; Fitted 102.2; Sustainable 89.2; Misalignment -16.9
    - 1990: Actual REER 90.6; Fitted 84.2; Sustainable 98.7; Misalignment -8.2
    - 1995: Actual REER 91.5; Fitted 90.4; Sustainable 91.6; Misalignment -0.1
    - 2000: Actual REER 93.5; Fitted 91.4; Sustainable 91.2; Misalignment 2.3
    - 2001: Actual REER 106.6; Fitted 111.3; Sustainable 97.4; Misalignment 9.5
    - 2002: Actual REER 111.0; Fitted 112.7; Sustainable 102.0; Misalignment 8.8
    - 2003: Actual REER 111.0; Fitted 112.7; Sustainable 102.0; Misalignment 8.8
    - 2004: Actual REER 106.3; Fitted 109.7; Sustainable 105.8; Misalignment 0.4
  - Summary conclusion from estimates:
    - Pula undervalued in the late 1980s.
    - Misalignment marginal in the 1990s.
    - Overvalued by 5 to 10 percent from 2000 to 2003.
    - 2004 shows realignment, possibly due to February 2004 devaluation.
- Figures referenced:
  - Figure 7: Botswana: Actual and Equilibrium Real Exchange Rates, 1980–2004 (Index: 1980=100).

### VI. DISCUSSION
- Policy regime implications:
  - Evidence of recent overvaluation provides a rationale for devaluations and for the May 2005 move from a fixed peg to a crawling peg.
  - Under a crawling peg the currency value is adjusted continuously rather than in discrete steps; credibility of the nominal anchor requires coordination of fiscal, monetary, and structural policies.
- Tradeoffs of regimes:
  - Fixed peg merits: credible nominal anchor, lower inflation (historically single-digit inflation for Botswana across two decades except 1992-94).
  - Fixed peg disadvantages: high volatility under real shocks and nominal rigidities; vulnerability to speculative attacks (though Botswana has sufficient reserves and a sound banking system making such attacks less likely).
  - Crawling peg advantages: can realign in line with shifts in fundamentals and stabilize the exchange rate in the short term.
  - Crawling peg disadvantages: provides no medium-term nominal anchor; risks of “galloping” inflation if adjustment is aggressive.
- Operational recommendations for a crawling peg:
  - Active crawling peg: preannounce rate of crawl for a year in advance; commit strongly to announced path to stabilize expectations and induce gradual disinflation.
  - Risks: if market rates are driven to the edge of the band, similar credibility problems as with hard pegs may arise; crawling band may be regarded as a weaker commitment than a hard peg.
- Cross-country evidence:
  - Rogoff et al. (2003): countries with “limited flexibility” tended to achieve lower inflation, higher per capita GDP growth, and less annual growth volatility than pegged regimes over 1970-1999.
  - Bubula and Otker-Robe (2003): crisis incidence higher for countries with conventional fixed regimes pegged to a basket than for those using crawling pegs during 1990-2001.
  - Ramcharan (2005): exogenous shocks (e.g., windstorms) have significant adverse impacts on developing countries with fixed regimes due to limited exchange rate flexibility.

### VII. CONCLUSION
- Main empirical findings:
  - Applying the BEER approach for 1985–2004, the Botswana pula appears:
    - Undervalued in the late 1980s.
    - Marginal misalignment in the 1990s.
    - Overvalued by 5 to 10 percent in recent years (2000–2003).
  - These findings provide a rationale for the February 2004 and May 2005 devaluations and the move to a crawling peg.
- Policy implications and required actions:
  - Successful implementation of crawling peg requires:
    - Comprehensive monetary and fiscal adjustments.
    - Acceleration of structural reforms to stimulate domestic private business and address high interest and unemployment rates.
  - Botswana faces constraints:
    - Fiscal position expected to tighten due to HIV/AIDS and poverty problems and high unemployment, limiting room for fiscal and interest-rate responses.
    - Need for credible operational framework and coordination across policy areas.
- Appendix (experiences from other countries adopting crawling pegs; managerial lessons):
  - Precondition: mature interbank market and central bank responsibility for exchange rate policy implementation.
  - Recommended band width: between ±7 and ±10 percent (Williamson, 1996); Botswana’s arrangement at time of writing allowed ±0.5 percent (up from ±0.125 percent).
  - Suggested policies to defend the band: readiness to change interest rates and reserve requirements; allow movements outside the band under extreme pressure with commitment to restore band later; set depreciating crawl consistent with inflation target minus expected foreign inflation and productivity differentials.

*Italic: Source — _wp06140 (IMF working paper content provided).*

### Appendix Table 1 shows macroeconomic indicators for four countries that have introduced

### _wp06140 - Appendix Table 1 shows macroeconomic indicators for four countries that have introduced

### Overview and purpose
- Appendix Table 1 compares macroeconomic indicators for four countries that introduced crawling peg regimes: Chile, Colombia, Israel, and Mexico.
- The analytical focus is on medium to long-term economic performance, ignoring short-run effects.
- The country selection follows Williamson (1996) and the IMF (1998).

### Shared preconditions and broad outcomes
- Common preconditions: weak external competitiveness and high inflation.
- General post-adoption tendencies:
  - National currencies tended to depreciate against the U.S. dollar.
  - Real exchange rates more or less stabilized, as intended.
  - It took a considerable amount of time, "five to ten years," for inflation to fall to single digits.
  - Putting high priority on external competitiveness created inflationary pressures and complicated lowering inflation expectations.
  - As inflation is contained, current account balances appear to have improved to sustainable levels.

### Botswana: advantages, disadvantages, and policy implications
- Relative advantages:
  - Botswana’s inflation is much lower than the four comparator countries, reducing the need for sterilized interventions when implementing a crawling peg.
  - Strong external position: a large current account surplus and an abundance of foreign reserves.
- Recent challenge:
  - A 12 percent devaluation of the pula that coincided with the move to a crawling peg pushed inflation outside the Bank of Botswana’s target range.
- Monetary/fiscal implications:
  - To restore credibility to the inflation objective, both monetary and fiscal policies may need to be tightened in the short to medium term.
  - To strengthen devaluation expectations, further discrete devaluations of the parity as the new regime may be needed (similar to Israel’s experience).
  - To maintain policy credibility, the government may need to defend the band under appreciation pressures.
- Structural constraints:
  - Emerging fiscal difficulty.
  - HIV/AIDS and poverty issues may complicate medium-term fiscal positions.
  - Unemployment is currently over 20 percent, limiting room to rely on fiscal policies and interest rates if intensive sterilization interventions are required.
  - Full implementation of a new exchange rate regime likely requires monetary and fiscal adjustments plus vigorous structural reforms to tackle high interest and unemployment rates.

### Comparative country experiences and specific notes
- Mexico:
  - The 1994 peso crisis elevated inflation above 30 percent and undermined the crawling peg announced in 1989.
- Botswana:
  - Inflation increased to a range of 17–18 percent in January–March 2006, mainly due to the reintroduction of school fees as well as the devaluation.
  - To account for the impact of devaluation, the central bank changed its inflation objective from 3-6 percent to 4-7 percent in August 2005.
- Israel:
  - The Israeli band policy was modified after introduction in 1991: the rate of crawl was adjusted downward and three discrete devaluations of parity were implemented between 1992 and 1995.
- Intervention approaches:
  - A crawling band makes sense only if it represents an attempt to keep the exchange rate reasonable when the market pushes it elsewhere.
  - Colombia used intramarginal intervention solely to reduce volatility.
  - Israel attempted to defend an inner band.

### Key statistics from Appendix Table 1 (preserved exactly as in source)

- Botswana (May 2005) — annual series 2002 2003 2004 2005:
  - GDP growth: 7.8 5.7 4.0 3.5
  - CPI: 8.0 9.3 6.9 8.6
  - Change in bilateral exchange rate against US$ (depreciation -): 27.7 23.1 3.8 -22.3
  - Change in REER (depreciation -): 4.7 3.1 -5.9 -9.5
  - Current account balance: 3.6 6.0 3.2 13.5
  - Overall budget balance (including grants): -4.1 -0.4 1.0 2.4
  - Total reserves (in months of imports): 30.3 23.0 18.4 19.4
  - Unemployment: 23.8 n.a. n.a. n.a.

- Chile (September 1982) — annual series 1979 1980 1981 1982; averages after adoption:
  - GDP growth: 8.7 8.1 4.7 -10.3 4.7 8.4 7.6
  - CPI: 33.4 35.1 19.7 9.9 23.4 19.0 9.2
  - Change in bilateral exchange rate against US$ (depreciation -): -15.0 -4.5 0.0 -23.4 -24.6 -9.5 -2.7
  - Change in REER (depreciation -): n.a. n.a. 18.5 -9.8 -12.4 0.1 4.8
  - Current account balance: -5.7 -7.1 -14.5 -9.5 -7.1 -1.5 -4.1
  - Overall budget balance (including grants): 4.8 5.4 2.6 -1.0 -1.4 1.4 2.1
  - Total reserves (in months of imports): 5.4 5.9 4.5 4.1 5.3 6.6 8.4
  - Unemployment: n.a. 10.4 11.3 19.6 11.5 5.4 5.2

- Colombia (November 1991) — annual series 1988 1989 1990 1991; averages after adoption:
  - GDP growth: 4.1 3.4 6.0 2.3 4.1 0.8
  - CPI: 28.1 25.8 29.1 30.4 22.9 13.0
  - Change in bilateral exchange rate against US$ (depreciation -): -18.9 -21.8 -23.8 -20.7 -9.2 -14.6
  - Change in REER (depreciation -): -2.8 -4.2 -12.1 2.9 7.2 -2.1
  - Current account balance: -0.6 -0.5 1.3 5.7 -3.2 -2.1
  - Overall budget balance (including grants): -1.3 -1.9 3.9 2.6 -2.0 -5.3
  - Total reserves (in months of imports): 5.4 5.3 6.1 9.3 6.8 5.9
  - Unemployment: 10.1 8.9 10.2 9.8 9.1 16.5

- Israel (December 1991) — annual series 1988 1989 1990 1991; averages after adoption:
  - GDP growth: 2.0 0.9 6.8 7.7 6.1 3.1
  - CPI: 16.3 20.2 17.2 19.0 11.3 4.4
  - Change in bilateral exchange rate against US$ (depreciation -): -0.3 -16.6 -4.9 -11.5 -6.4 -5.3
  - Change in REER (depreciation -): 9.7 1.0 -2.3 1.9 0.8 1.6
  - Current account balance: -1.9 0.5 0.3 -2.2 -4.0 -1.6
  - Overall budget balance (including grants): -8.4 -4.3 -5.3 -6.8 -3.7 -1.2
  - Total reserves (in months of imports): 2.6 3.3 3.3 3.0 2.6 5.6
  - Unemployment: 6.4 8.9 9.6 10.6 8.5 8.6

- Mexico (January 1989) — annual series 1986 1987 1988 1989; averages after adoption:
  - GDP growth: -3.8 1.9 1.2 4.2 3.9 2.9
  - CPI: 86.2 131.8 114.2 20.0 16.3 24.5
  - Change in bilateral exchange rate against US$ (depreciation -): -58.0 -55.6 -39.4 -7.7 -6.0 -16.9
  - Change in REER (depreciation -): -30.2 -7.9 24.0 7.3 5.0 1.7
  - Current account balance: -1.1 3.0 -1.3 -2.6 -5.4 -2.0
  - Overall budget balance (including grants): -13.0 -14.2 -8.9 -4.6 1.0 -1.0
  - Total reserves (in months of imports): 2.6 5.0 1.7 1.5 2.3 2.2
  - Unemployment: n.a. n.a. 2.5 n.a. 3.4 3.7

### Synthesis of policy lessons
- Stabilizing the real exchange rate under a crawling peg often involves short- to medium-term inflationary pressures that can delay achievement of low inflation targets for "five to ten years."
- Restoring and maintaining credibility of inflation objectives may require coordinated tightening of monetary and fiscal policy, and potentially discrete devaluations to shape expectations.
- Defending the band under appreciation pressures can be necessary to maintain credibility; intervention methods differ by country.
- Structural reforms are critical where fiscal space is limited and unemployment is high to avoid overreliance on sterilized intervention and to support a successful transition to a crawling exchange rate regime.

*Source: Appendix Table 1 and accompanying text from the supplied IMF document.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06140.pdf_