## _wp0734

## Source details

**Canonical URL:** [_wp0734](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp0734.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp0734.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp0734.pdf.json)

---

### Introduction: research question and approach
- Research question: Can monetary policy be autonomous in the context of a fully pegged exchange rate and a capital account officially free?
- Context: The franc zone has had a fixed exchange rate and officially free capital movements for more than 50 years, a configuration that is inconsistent with classic “trilemma” logic (Triffin, 1963).
- Main argument:
  - Autonomy of monetary policy in the franc zone is achieved because of a contingent credit line provided by the French treasury (the compte d’operation, CO), which reduces the need to build up reserves, together with de facto widespread capital controls.
- Comparative design:
  - Systematic comparison with the Eastern Caribbean Currency Union (ECCU), which operates under a currency board.
  - Empirical tests show:
    - Monetary policy has been overall autonomous in the franc zone, although this varies by period.
    - Monetary policy clearly appears nonautonomous in the ECCU.

### Franc Zone institutional setup
- Composition and structure:
  - Arrangement links six central banks: Banque de France, two monetary institutions directly dependent on French authorities, Banque des Etats d’Afrique Centrale (BEAC), Banque des Etats d’Afrique de l’Ouest (BCEAO), and the central bank of the Comoros.
  - BCEAO issues CFA francs for WAEMU; BEAC issues CFA francs for CAEMC. Two distinct CFA francs legal tender only in their respective issuance areas.
  - Monetary unions and the Comoros governed by three separate conventions de cooperation (CC).
- Historical departures and membership:
  - Algeria, Guinea, Madagascar, Morocco, and Tunisia left soon after independence; Mauritania abandoned both the franc zone and WAEMU in 1972.
  - Comoros issues the Comorian franc; “CFA franc zone” label is misleading because there are the Comorian and French francs plus two distinct CFA francs.
- Selected institutional dates/values preserved from Table 1:
  - Banque de France (1998); Banque centrale européenne (1800) and Banque de France; Euro (since 1999, previously French francs); France (metropolitan); Monaco.
  - IEDOM (1959); IEOM (1967); BCEAO (1959); BEAC (1959); Banque centrale des Comores (1962).
  - Currency conversion notes: 1 CFPF = 0.055 FF; 1 CFA = 0.01 FF; 1 CF = 0.0133.

### The compte d’operations (CO): role, conventions, and fiscal implications
- Role and features:
  - CO is the key instrument guaranteeing access of zone members to foreign reserves.
  - Through the CO, the French treasury guarantees unlimited convertibility of CFA and Comorian francs at a given rate.
  - Guaranteed convertibility reduces reserves required by central banks to support the peg; main item distinguishing franc zone from other fixed-rate arrangements.
- Institutional mechanics and convention rules:
  - CO is a special treasury account with overdrafting facilities; credited with foreign reserves of central banks; can become negative when the treasury provides foreign reserve advances.
  - Since 1972, beneficiaries must pool 65 percent of their foreign reserves in the CO; historically BCEAO kept 89 percent and BEAC 83 percent.
  - CO deficit is unlimited.
  - Interest practices (medians, historical):
    - Since 1956, BEAC and BCEAO respectively received (paid) interest corresponding to 2.1 (0.72) and 2.9 (1.5) of their base money (medians).
    - French treasury received (paid) interest corresponding on median to 0.11 (0.36) percent of its fiscal revenues.
  - Ratissage: rapid decrease of CO credit triggers surrendering of reserves by public entities; exporters required to surrender currency receipts.
  - French treasury appoints auditors and can cancel conventions in case of misuse.
  - Additional provisions:
    - Foreign reserves should correspond to at least 20 percent of base money.
    - Historically, advances to treasuries limited to 20 percent of tax revenue of previous year; advances ruled out in WAEMU but not in CAEMC.
  - Footnote financial rates:
    - Credit interest: The European Central Bank intervention rate (or at least 2.5 percent) for more than 10 million franc.
    - Debit interest schedule: 1 percent for 0 to 5 million francs, 2 percent for 5 to 10 million francs, and the European Central Bank intervention rate (or at least 2.5 percent) for more than 10 million francs.
- Financial implications and potential fiscal effort for the French treasury:
  - Median actual effort on debit period corresponds to 0.13 percent of the fiscal revenue; largest increase in deficit represents 1.89 percent.
  - During 1962–2005, ratio of money and quasi-money in the zone to French fiscal revenue never exceeded 3.6 percent; period median 2.8 percent.
  - Catastrophic scenario (demand for foreign reserves equals liability of the financial system, M1): cost would not have corresponded to more than 3 percent of French revenue.
  - Cost in terms of French income (GDP) during 1962–2005: median potential cost 0.8 percent, maximum 1.3 percent.
  - Chronology: costs increased until 1986, decreased after tightened monetary policy and structural adjustment, 1994 devaluation exacerbated downward trend; post-devaluation recovery increased cost again.
  - Risk from enlargement (Ghana and Nigeria): potential cost would have been 7.4 percent (median for 1962–2005); maximum cost would have been 17 percent higher; standard deviation more than six times higher.

### Historical usage of the CO and reserve coverage
- Usage patterns:
  - BCEAO and BEAC effectively used the CO before the 1994 devaluation; BCEAO withdrew reserve currency relatively more frequently.
  - Apart from Comoros (and France), all zone members used CO facilities several times, resulting in significant CO debits.
  - For BCEAO: deficits appeared first in 1983; episodes in 1980s lasted several months up to 3 years; deficit periods longer and more frequent and maximum deficit in terms of base money larger than BEAC.
  - For BEAC: deficits appeared in 1987; deficit faded after some months in 1994, although foreign liabilities increased after the 1994 devaluation.
  - After the 1994 devaluation, no new CO deficits appeared.
- Reserve adequacy and coverage statistics (Table 2, 1956–2005 monthly data):
  - CO below 20% of base money:
    - BCEAO: 105 months (17.6%)
    - BEAC: 161 months (27.4%)
    - ECCU (1983–2005): 0 (0%)
  - Foreign reserves below 20% of base money:
    - BCEAO: 36 months (6%)
    - BEAC: 116 months (16%)
  - Coverage of demand liability with foreign reserves:
    - ECCU: 98 percent
    - BEAC: 50 percent (40 percent with CO deposits)
    - BCEAO: 45 percent (40 percent with CO deposits)
- Historical variation in foreign reserve/base money ratio:
  - 1956–1980: ratio between 20 to 80 percent.
  - 1980–1994: foreign reserves declined; often under 20 percent.
  - After 1994 devaluation: banks built high levels of foreign reserves; ratio often exceeded 100 percent.

### Capital mobility and exchange rate regime in practice
- Official principle: free capital mobility among zone members, but in practice controls are implemented.
- Evolution:
  - 1939: French administration implemented exchange controls; zone currencies remained freely convertible among themselves.
  - 1967: France removed most capital controls, making zone francs convertible to other currencies; conventions call for common exchange regulation and free intrazone transfers.
- De facto practice:
  - Conventions not uniformly enforced; capital regulations among zone members are not uniform.
  - Intrazone capital controls exist for many transactions: controls on foreign assets of financial institutions; deposit and repatriation of export receipts; controls on issuing/selling shares, gold, bills and coins, foreign investments and borrowing abroad.
  - Transfers and invisibles remain free.
- Exchange rate:
  - Theoretically pegged but adjustable; in practice long record of unconditional fixity.
  - Exchange rate adjustments require consultations with French government and unanimous vote of member countries; since 1945 the peg changed only once (1994).
  - CAEMC CFA and WAEMU CFA pegged to the euro (and francs before) at same rate even after 1994 devaluations, except Comoros where devaluation differed.

### Comparative note: franc zone vs. ECCU
- Regime difference:
  - Franc zone: peg supported by CO and de facto capital controls.
  - ECCU: currency board arrangement without an unlimited contingency line.
- ECCU features:
  - Adopted lower foreign reserve target (60 percent of base money in 1975) compared with orthodox currency board target (100 percent).
  - ECCU does not have an unlimited contingency line like the CO; foreign reserves adequacy is crucial.

### Testing monetary autonomy — long-run relationships (cointegration)
- Model: Bs_t = α R_t + e_t, with interpretation:
  - e_t stationary → monetary policy nonautonomous (reserves determine base money).
  - e_t with stochastic trend → autonomous monetary policy.
- Key empirical findings:
  - ECCU: cointegration present between base money and foreign reserves; domestic assets stationary → monetary policy nonautonomous.
  - CAEMC and WAEMU: cointegration rejected → evidence of autonomous monetary policy.
- Temporal changes / structural breaks considered at 1981 and 1994; subperiods 1956–1980, 1980–1994, 1994–2005.
- Seasonally adjusted OLS co-integration selected values (Table 6):
  - Zone-wide foreign assets - base money:
    - 1994–2005: coefficient 2.14, p-value 0.03.
    - 1980–1994: coefficient -1.00, p-value 0.31.
    - 1956–1980: coefficient 1.28, p-value 0.19.
  - Co-integration CO - base money:
    - 1994–2005: coefficient -3.13, p-value 0.00.
    - 1980–1994: coefficient 0.47, p-value 0.63.
    - 1956–1980: coefficient 1.64, p-value 0.10.
  - Member-specific:
    - BCEAO (1994–2005): CO–base money cointegration coefficient 2_t e = -3.99, p-value 0.00.
    - BEAC (1994–2005): foreign assets–base money cointegration coefficient 1_t e = 4.11, p-value 0.00.

### Short-run dynamics, error correction, and offsetting behavior
- Short-run model: first-difference equations; where cointegration holds (ECCU) includes an error correction term (ECT).
- Seasonally Adjusted OLS short-run key values (Table 7, March 1956–August 2005; p-values in parentheses):
  - Autoregressive term (lagged change in Bs):
    - Zone: 0.13 (0.00)
    - BEAC: -0.31 (0.00)
    - BCEAO: 0.21 (0.00)
    - ECCU: -0.04 (0.39)
  - Change in foreign reserves (1956–2005):
    - Zone: 0.12 (0.00)
    - BEAC: 0.19 (0.09)
    - BCEAO: 0.11 (0.00)
    - ECCU: 0.17 (0.00)
  - Change in CO balance (1956–2005):
    - BEAC: 0.016 (0.09)
    - BCEAO: -0.07 (0.54)
    - Zone: 0.015 (0.11)
  - Error correction model (ECCU only):
    - Error correction term CT (domestic assets) = -0.11 (0.0) indicating more than 10 percent of deviation corrected each month.
  - Observations:
    - Zone: 1176
    - BEAC: 589
    - BCEAO: 579
    - ECCU: 319
- Interpretation:
  - ECCU: significant error correction; base money responsive to foreign reserves; each month >10 percent of deviation corrected (three-quarters corrected in a year).
  - BCEAO and BEAC: no valid error correction model for entire period; monthly share of deviations corrected close to zero (best case 3.8 percent after a year).
  - Effect of 1994 devaluation: responsiveness of base money to changes in foreign reserves increased after 1994, reflecting tighter targeting post-devaluation.

### Offsetting coefficients (degree of sterilization) — Table 8 results
- Definition: proportion of reserve changes that are offset by monetary authorities; coefficient close to -1 indicates full offsetting (active monetary policy); close to 0 indicates little or no offsetting.
- Change in foreign reserves (1956–2005):
  - Zone: -0.88
  - BEAC: -1.0
  - BCEAO: -0.89
  - ECCU (1956–2005 1/): -0.14
- Subperiods for foreign reserves:
  - 1956–80:
    - Zone: -1.1
    - BEAC: -1.6
    - BCEAO: -1.08
  - 1980–94:
    - Zone: -0.9
    - BEAC: -0.93
    - BCEAO: -0.86
  - 1994–2005:
    - Zone: -0.7
    - BEAC: -0.55
    - BCEAO: -0.7
- Change in CO balance (1956–2005):
  - Zone: -0.98
  - BEAC: -1.07
  - BCEAO: -0.98
  - 1994–2005:
    - Zone: -0.68
    - BEAC: -0.54
    - BCEAO: -0.68
- Interpretation:
  - Before 1994, BCEAO and BEAC largely offset changes in foreign reserves (coefficients close to -1).
  - After 1994, offsetting decreased (coefficients closer to zero), indicating reduced activism.
  - ECCU: offset coefficient close to zero throughout, suggesting monetary policy does not offset reserve changes.

### Foreign reserve thresholds (Table 9) — sensitivity regimes
- Procedure: thresholded regressions for reserve/base-money ratios between 5 and 100 percent; central banks supposed to maintain ratio above 20 percent.
- Main findings:
  - Zone-wide CO-based threshold:
    - Change of CO balance: 0 to 35% coefficient 0.12 (p-value 0.00); 35 to 100% coefficient 0.037 (p-value 0.00).
    - Monetary policy becomes more responsive to CO position under CO/base money threshold of 35 percent.
    - No change in monetary policy related to the ratio foreign reserve/base money found for the zone overall.
  - BCEAO:
    - Change of CO balance: 0 to 35% coefficient 0.11 (p-value 0.00); 35 to 100% coefficient 0.026 (p-value 0.00).
    - No foreign reserve target found using foreign reserves; CO-based threshold estimated at 35 percent.
  - BEAC:
    - Change in foreign reserves: 0 to 30% coefficient 0.16 (p-value 0.00); 30 to 100% coefficient 0.10 (p-value 0.00).
    - Change of CO balance: 0 to 30% coefficient 0.18 (p-value 0.00); 30 to 100% coefficient 0.12 (p-value 0.00).
    - Threshold corresponds to 30 percent (foreign assets or CO/base money) entailing moderate tightening.

### Conclusions and policy implications
- Comparative autonomy:
  - ECCU: monetary policy nonautonomous, consistent with currency-board rule.
  - Franc zone: able to combine autonomy with fixity mainly because of small size of economies and limited financial development, aided by the CO and de facto capital controls.
- Time variation and structural change:
  - Structural shift in 1994 reduced autonomy; central banks target specific reserve levels so responsiveness changes with reserve level.
  - Instruments to limit CO deficit (e.g., progressively increasing interest rate) did not have significant influence on monetary policy.
- Policy trade-offs and future risks:
  - Near-term: zone size likely to stay the same; cooperation may be unchanged.
  - Risks from financial development and growth: increased pressures on fixed exchange rate and monetary cooperation; maintaining the zone requires stricter compliance with reserve-preserving rules.
  - Trade-offs:
    - Strictly targeting foreign reserves would preserve the peg but abolish the historic objective of allowing monetary policy autonomy under fixity; would require building reserve stocks consistent with peg and limit authorities’ ability to react to shocks.
    - If zone members become emerging economies, they may dismiss the peg or terminate cooperation or preserve it with more flexible exchange rate.
    - Deepening monetary union integration would ease softening the peg and reforming arrangement by reducing exposure to external shocks and allowing monetary autonomy.

*Italic: Source PDF filename: _wp0734 - References..............................................................................................................*

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

### _wp0734 - References..............................................................................................................

### Introduction: research question and approach
- Research question: Can monetary policy be autonomous in the context of a fully pegged exchange rate and a capital account officially free?
- Context: The franc zone has had a fixed exchange rate and officially free capital movements for more than 50 years, a configuration that is inconsistent with classic “trilemma” logic (Triffin, 1963).
- Main argument: Autonomy of monetary policy in the franc zone is achieved because of a contingent credit line provided by the French treasury (the compte d’operation), which reduces the need to build up reserves, together with de facto widespread capital controls.
- Comparative design: The paper systematically compares the franc zone with the Eastern Caribbean Currency Union (ECCU), which operates under a currency board; empirical tests show:
  - Monetary policy has been overall autonomous in the franc zone, although this varies by period.
  - Monetary policy clearly appears nonautonomous in the ECCU.

### Franc Zone institutional setup
- Composition and structure:
  - The franc zone is the result of historic monetary cooperation between France and 15 countries in Africa.
  - The arrangement links six central banks (Table 1): the Banque de France, two monetary institutions directly dependent on French authorities, the Banque des Etats d’Afrique Centrale (BEAC), the Banque des Etats d’Afrique de l’Ouest (BCEAO), and the central bank of the Comoros.
  - The BCEAO issues CFA francs for the West Africa Monetary Union (WAEMU); the BEAC issues CFA francs for the Central Africa Economic Monetary Community (CAEMC). Despite sharing the acronym CFA, the two CFA francs are distinct and are legal tender only in their respective issuance areas.
  - The monetary unions and the Comoros are governed by three separate conventions de cooperation (CC).
- Historical departures and membership details:
  - Some former French colonies left the zone: Algeria, Guinea, Madagascar, Morocco, and Tunisia left soon after independence; Mauritania abandoned both the franc zone and WAEMU in 1972.
  - The central bank of the Comoros issues the Comorian franc; the “CFA franc zone” label is thus misleading because there are the Comorian and French francs plus two distinct CFA francs.

### The compte d’operations (CO) and reserve support
- Role and features of the CO:
  - The compte d’operations (CO) is the key instrument guaranteeing access of zone members to foreign reserves.
  - Through the CO, the French treasury guarantees unlimited convertibility of CFA and Comorian francs at a given rate.
  - The guaranteed convertibility reduces the level of reserves required by central banks to support the peg.
  - The guaranteed convertibility is the main item distinguishing the franc zone from other fixed rate arrangements.
- Institutional trade-offs and disciplining mechanisms:
  - In compensation for the guaranty, zone members agreed on rules to avoid excessive deficits (Box 1).
  - Pooled reserves are less vulnerable to external shocks because the unions’ economies are more diversified than individual members.
  - The need to define a common monetary policy reduces individual governments’ influence and lowers the probability of policy reversals, thereby protecting foreign reserves.

### Historical usage of the CO and episodes of CO debits
- Usage patterns:
  - Both the BCEAO and BEAC effectively used the CO facility before the 1994 devaluation; the BCEAO withdrew reserve currency relatively more frequently than the BEAC.
  - Apart from the Comoros (and France), all zone members used CO facilities several times, resulting in significant CO debits.
- Timing and magnitude of deficits:
  - For the BCEAO:
    - Deficits appeared first in 1983.
    - Episodes in the 1980s lasted several months up to 3 years.
    - Deficit periods were longer and more frequent for the BCEAO than for the BEAC.
    - The maximum deficit in terms of base money was larger for the BCEAO.
  - For the BEAC:
    - Deficits appeared in 1987.
    - The deficit faded away after some months in 1994, although foreign liabilities increased after the 1994 devaluation.
  - After the 1994 devaluation, no new CO deficits appeared.

### Comparative note: franc zone vs. ECCU
- Regime difference:
  - The franc zone maintains a peg supported by the CO and de facto capital controls.
  - The ECCU is based on a currency board arrangement.
- Empirical outcome:
  - Tests corroborate that the franc zone broadly achieved monetary policy autonomy (with period variation).
  - Tests show the ECCU’s monetary policy clearly appears nonautonomous.

### Key institutional and numeric details preserved from Table 1 and text
- Institutions and selected dates/values as presented:
  - Banque de France (1998); Banque centrale européenne (1800) and Banque de France; Euro (since 1999, previously French francs); France (metropolitan); Monaco.
  - IEDOM (1959) Took over from the Caisse centrale de l’outre-mer; Euro (since 1999, previously French francs); Guadeloupe, French Guyana, Martinique, Reunion, Mayotte, St-Pierre et Miquelon.
  - IEOM (1967) Took over from la banque d’Indochine; French franc; French CFP (1945) (Communauté Financière du Pacifique); 1 CFPF = 0.055 FF; New-Caledonia, Wallis-et-Futuna, French Polynesia.
  - BCEAO (1959) Took over la Caisse centrale de l’outre-mer, which took over from la banque d’Afrique de l’Ouest; Franc CFA (Communauté Financière Africaine) (1945); 1 CFA = 0.01 FF; Benin, Burkina Faso, Côte d’Ivoire, Guinea-Bissau (1997), Mali, Niger, Senegal, Togo.
  - BEAC (1959) Took over from la Caisse centrale de l’outre-mer, which took over from la banque d’Afrique de l’Ouest; Franc CFA (Coopération Financière en Afrique centrale) (1945); 1 CFA = 0.01 FF; Cameroon, Central African Republic, Congo (Republic of), Gabon, Equatorial Guinea, Chad.
  - Banque centrale des Comores (1962) Took over from la Banque centrale de Madagascar et des Comores; (1973) Madagascar abandoned the zone; (1979) The Comoros joined the franc zone as an independent state; Franc Comorian; 1 CF = 0.0133; Comoros.
- Data sources and coverage note:
  - For BCEAO and BEAC, the database is built on their respective monthly central banks bulletins, issued since 1956; bulletins report the balance sheet of the monetary authorities, including base money and its counterparts.
  - Foreign assets (FA) consist of: reserves deposited on the CO, reserves position at the IMF, and other FA; other FA incorporates reserves invested in convertible currencies.

*Italic: Source PDF filename: _wp0734 - References..............................................................................................................*

### BOX 1. THE COMPTE D’OPERATIONS

### BOX 1. THE COMPTE D’OPERATIONS

### Origins
- The compte d’operations (CO) was created by the French treasury to handle balance-of-payment issues between France and Algeria; first version appeared in Algeria in 1878.
- Mechanism: French treasury deposited reserves in Algerian francs into the Bank of Algeria; transfers between Algeria and France were converted at a given rate but did not entail currency exchange; French treasury netted transactions and held a net position.
- Result: French treasury deposited its balance in the Bank of Algeria and conditioned monetary policy of the beneficiary country via advances and progressively increasing interest on advances.
- 1925: an account opened in the French treasury for the Bank of Morocco; subsequently accounts opened for other central banks in the Zone.

### Institutional setup and conventions
- Legally, the CO is a special treasury account distinct from the general French budget; functions as a current account with overdrafting facilities.
- Account mechanics:
  - Account credited with foreign reserves of central banks; can become negative when members have unfavorable balance of payments and the treasury provides foreign reserve advances.
  - Based on international conventions (convention de compte d’operation) that stipulate conditions to avoid excessive deficits.
- Key convention rules and historical practices:
  - Since 1972, CO beneficiaries must pool 65 percent of their foreign reserves in the CO. Historically, banks kept larger amounts than required (89 percent for the BCEAO and 83 percent for the BEAC).
  - The deficit of the CO is unlimited.
  - The central bank receives interest for CO credits, and pays a progressively increasing interest rate on CO debits.
    - Since 1956, the BEAC and BCEAO respectively received (paid) interest corresponding to 2.1 (0.72) and 2.9 (1.5) of their base money (medians).
    - The French treasury received (paid) interest corresponding on median to 0.11 (0.36) percent of its fiscal revenues.
    - Financial conditions fixed by the conventions and are no longer used as an instrument to control account position.
  - Ratissage: in case of rapid decrease of CO credit, public entities are supposed to surrender their reserves to central banks; exporters required to surrender currency receipts.
  - French treasury appoints auditors to check compliance with conventions; can cancel conventions in case of misuse.
- Additional convention de cooperation provisions:
  - Foreign reserves should correspond to at least 20 percent of base money.
  - Historically, advances to treasuries were limited to 20 percent of the tax revenue of the previous year; advances ruled out in WAEMU, but not in CAEMC.
  - French treasury still appoints members to central bank boards.
- Footnote financial rates:
  - Credit interest: The European Central Bank intervention rate (or at least 2.5 percent) for more than 10 million franc.
  - Debit interest schedule: 1 percent for 0 to 5 million francs, 2 percent for 5 to 10 million francs, and the European Central Bank intervention rate (or at least 2.5 percent) for more than 10 million francs.

### Financial implications and potential fiscal effort for the French treasury
- Historical assessment of cost to French treasury:
  - Median of the actual effort demanded on debit period corresponds to 0.13 percent of the fiscal revenue; the largest increase in deficit represents 1.89 percent.
  - During 1962–2005, ratio of money and quasi-money in the zone to French fiscal revenue never exceeded 3.6 percent; period median was 2.8 percent.
  - In a catastrophic scenario where demand for foreign reserves equals liability of the financial system (M1), cost for the treasury would not have corresponded to more than 3 percent of its revenue.
  - Cost in terms of French income (GDP) during 1962–2005: median potential cost 0.8 percent, maximum 1.3 percent.
- Chronology and trends:
  - Costs increased until 1986 (sustained growth and financial development), then decreased sharply after tightened monetary policy during structural adjustment.
  - The 1994 devaluation exacerbated the downward trend in cost; post-devaluation recovery increased cost again.
- Risks with financial development and zone enlargement:
  - Sustained increase in credit increases financial system liabilities and potential bailout cost; could reach levels French taxpayers unwilling to pay.
  - Inclusion of Ghana and Nigeria: potential cost would have been 7.4 percent (median for 1962–2005); maximum cost would have been 17 percent higher; standard deviation more than six times higher.

### Reserve targets, coverage, and CO usage
- Reserve adequacy rules and empirical coverage:
  - Central banks supposed to maintain foreign reserves at least 20 percent of base money.
  - Table 2 (1956–2005, monthly data): number of months CO balance below 20 percent of base money and number of months foreign reserves below 20 percent of base money:
    - BCEAO: CO below 20% for 105 months (17.6%); foreign reserves below 20% for 36 months (6%).
    - BEAC: CO below 20% for 161 months (27.4%); foreign reserves below 20% for 116 months (16%).
    - ECCU (1983–2005): CO below 20% = 0 (0%); reference for ECCU is 60 percent of base money.
  - Coverage of demand liability with foreign reserves:
    - ECCU covered 98 percent of demand liability with foreign reserves.
    - BEAC covered 50 percent (40 percent with CO deposits).
    - BCEAO covered 45 percent (40 percent with CO deposits).
- Historical variation in foreign reserve/base money ratio:
  - 1956–1980: ratio between 20 to 80 percent.
  - 1980–1994: foreign reserves declined; often under 20 percent despite protective measures.
  - After 1994 devaluation: banks built high levels of foreign reserves; ratio often exceeded 100 percent.

### Capital mobility and exchange rate regime
- Capital mobility:
  - Official principle: free capital mobility among zone members, but in practice controls are implemented.
  - 1939: French administration implemented exchange controls; zone currencies remained freely convertible among themselves.
  - 1967: France removed most capital controls, making zone francs convertible to other currencies; conventions call for common exchange regulation and free intrazone transfers.
  - In practice, conventions not uniformly enforced; capital regulations among zone members are not uniform and intrazone capital controls now exist for many transactions (controls on foreign assets of financial institutions; deposit and repatriation of export receipts; controls on issuing/selling shares, gold, bills and coins, foreign investments and borrowing abroad). Transfers and invisibles remain free.
- Exchange rate:
  - Theoretically pegged but adjustable; in practice, long record of unconditional fixity.
  - Exchange rate adjustments require consultations with French government and unanimous vote of member countries; since 1945 the peg changed only once (1994).
  - CAEMC CFA and WAEMU CFA pegged to the euro (and francs before) at same rate even after 1994 devaluations, except Comoros where devaluation differed.

### Comparison with the Eastern Caribbean Currency Union (ECCU)
- ECCU features:
  - Historical example of a currency board in developing countries; regime maintained a pegged exchange rate for more than 55 years.
  - ECCU adopted lower foreign reserve target (60 percent of base money in 1975) compared with orthodox currency board target (100 percent).
  - ECCU does not have an unlimited contingency line like the CO; foreign reserves adequacy is crucial.
- Table 3 summary of evolution and rules:
  - British Caribbean Currency Board (1950–1965): foreign reserves/base money = 100 percent.
  - Eastern Caribbean Currency Authorities (1965–1983): foreign reserves/base money = 70 percent.
  - Eastern Caribbean Currency Union (1983–2004): foreign reserves/base money = 60 percent (1975); exchange rate evolved from pound sterling peg to U.S. dollar peg in 1976.
- Table 4 highlights differences:
  - Franc Zone: fixed by international treaties since 1948 (one adjustment in 1994); capital account: free movement with anchor currency since 1939 but de facto controls; monetary policy: pool of foreign reserves plus French Treasury guaranty through the CO.
  - ECCU: fixed by law since 1955; open capital account; monetary policy: pool of foreign reserves and high foreign reserves cover of base money.

### Testing monetary autonomy in the franc zone — Long-run relationships
- Model and interpretation:
  - Cointegration model: Bs_t = α R_t + e_t, where R_t = stock of reserves, Bs_t = base money, α = responsiveness of base money to reserves, e_t = domestic counterparts.
  - If e_t stationary → monetary policy nonautonomous (reserves determine base money); if e_t has a stochastic trend → autonomous monetary policy.
  - Phillips-Perron unit root tests applied to domestic assets (base money minus net foreign reserves).
- Key empirical findings (Table 5 and narrative):
  - ECCU: long-run relationship between base money and foreign reserves holds (cointegration present); domestic assets stationary → monetary policy nonautonomous.
  - CAEMC and WAEMU: cointegration rejected → evidence of autonomous monetary policy.
- Temporal changes and structural breaks:
  - Potential structural breaks considered at 1981 (first negative CO) and 1994 (devaluation).
  - Allowing for time breaks (subperiods 1956–1980, 1980–1994, 1994–2005) shows cointegration mainly valid during last period (1994–2005).
  - Table 6 seasonally adjusted OLS results (selected values):
    - Zone-wide co-integration foreign assets - base money:
      - 1_t e (1994–2005): coefficient 2.14, p-value 0.03.
      - 1_t e (1980–1994): coefficient -1.00, p-value 0.31.
      - 1_t e (1956–1980): coefficient 1.28, p-value 0.19.
    - Co-integration CO - base money:
      - 2_t e (1994–2005): coefficient -3.13, p-value 0.00.
      - 2_t e (1980–1994): coefficient 0.47, p-value 0.63.
      - 2_t e (1956–1980): coefficient 1.64, p-value 0.10.
    - Member-specific: BCEAO shows cointegration between base money and CO in 1994–2005 (2_t e = -3.99, p-value 0.00); BEAC shows cointegration between foreign assets and base money in 1994–2005 (1_t e = 4.11, p-value 0.00).

### Short-run dynamics and offsetting behavior
- Short-run model: first-difference form capturing how changes in foreign reserves influence changes in base money; where cointegration holds (ECCU) includes error correction term (ECT).
- Table 7 key values (Seasonally Adjusted OLS, March 1956–August 2005; p-values in parentheses):
  - Dependent variable: log(Bs_t) - log(Bs_{t-1})
  - Autoregressive term (lagged change in Bs):
    - Zone: 0.13 (0.00)
    - BEAC: -0.31 (0.00)
    - BCEAO: 0.21 (0.00)
    - ECCU: -0.04 (0.39)
  - Change in foreign reserves (1956–2005):
    - Zone: 0.12 (0.00)
    - BEAC: 0.19 (0.09)
    - BCEAO: 0.11 (0.00)
    - ECCU: 0.17 (0.00)
  - Change in CO balance (1956–2005) for BEAC/BCEAO:
    - BEAC: 0.016 (0.09)
    - BCEAO: -0.07 (0.54)
    - Zone: 0.015 (0.11)
  - Error correction model (ECCU only): error correction term CT (domestic assets) = -0.11 (0.0) indicating more than 10 percent of deviation corrected each month.
  - Observations:
    - Zone: 1176
    - BEAC: 589
    - BCEAO: 579
    - ECCU: 319
- Interpretation:
  - ECCU: significant error correction; base money responsive to foreign reserves; each month >10 percent of deviation corrected (three-quarters corrected in a year).
  - BCEAO and BEAC: no valid error correction model for entire period; monthly share of deviations corrected close to zero (best case 3.8 percent after a year).
- Effect of 1994 devaluation:
  - Responsiveness of base money to changes in foreign reserves increased after 1994, reflecting tighter targeting of foreign reserves post-devaluation.

### Offsetting coefficients (degree of sterilization)
- Offsetting coefficients measure proportion of reserve changes that are offset by monetary authorities; coefficient close to -1 indicates full offsetting (active monetary policy); close to 0 indicates little or no offsetting.
- Table 8 selected results (March 1956–August 2005):
  - Change in foreign reserves (1956–2005):
    - Zone: -0.88
    - BEAC: -1.0
    - BCEAO: -0.89
    - ECCU (1956–2005 1/): -0.14
  - Subperiods for foreign reserves:
    - 1956–80: Zone -1.1; BEAC -1.6; BCEAO -1.08
    - 1980–94: Zone -0.9; BEAC -0.93; BCEAO -0.86
    - 1994–2005: Zone -0.7; BEAC -0.55; BCEAO -0.7
  - Change in CO balance (1956–2005):
    - Zone -0.98; BEAC -1.07; BCEAO -0.98
    - 1994–2005: Zone -0.68; BEAC -0.54; BCEAO -0.68
  - Footnote: Two lags considered for ECCU estimate.
- Interpretation:
  - Before 1994, BCEAO and BEAC largely offset changes in foreign reserves (coefficients close to -1).
  - After 1994, offsetting decreased (coefficients closer to zero), indicating reduced activism.
  - ECCU: offset coefficient close to zero throughout, suggesting monetary policy does not offset reserve changes (low activism).

### Foreign reserve thresholds
- Definition: ratio foreign reserve/base money under which monetary policy becomes more sensitive to foreign reserves.
- Estimation procedure: run thresholded regressions for ratios between 5 and 100 percent; central banks supposed to maintain ratio above 20 percent.
- Main findings (Table 9):
  - Zone-wide:
    - Change of CO balance: threshold 0 to 35% with coefficient 0.12 (p-value 0.00); 35 to 100% with coefficient 0.037 (p-value 0.00).
    - Monetary policy becomes more responsive to CO position under CO/base money threshold of 35 percent.
    - No change in monetary policy related to the ratio foreign reserve/base money found for the zone overall.
  - BCEAO:
    - Change of CO balance: 0 to 35% coefficient 0.11 (p-value 0.00); 35 to 100% coefficient 0.026 (p-value 0.00).
    - No foreign reserve target found using foreign reserves; CO-based threshold estimated at 35 percent.
  - BEAC:
    - Change in foreign reserves: 0 to 30% coefficient 0.16 (p-value 0.00); 30 to 100% coefficient 0.10 (p-value 0.00).
    - Change of CO balance: 0 to 30% coefficient 0.18 (p-value 0.00); 30 to 100% coefficient 0.12 (p-value 0.00).
    - Threshold corresponds to 30 percent (foreign assets or CO/base money) entailing moderate tightening.

### Conclusions and policy implications
- Degree of monetary autonomy differs significantly between ECCU and franc zone despite same exchange rate constraint:
  - ECCU: monetary policy nonautonomous, consistent with currency-board rule.
  - Franc zone: able to combine autonomy with fixity mainly because of small size of economies and limited financial development.
- Autonomy is time-varying:
  - Structural shift in 1994 reduced autonomy; central banks also target specific reserve levels so responsiveness changes with reserve level.
  - Instruments to limit CO deficit (e.g., progressively increasing interest rate) did not have significant influence on monetary policy.
- Policy implications concerning the future of the zone:
  - Near-term: zone size likely to stay the same; cooperation may be unchanged.
  - Risks from financial development and growth: increased pressures on fixed exchange rate and monetary cooperation; maintaining the zone requires stricter compliance with reserve-preserving rules.
  - Trade-offs:
    - Strictly targeting foreign reserves would preserve the peg but abolish historic objective of allowing monetary policy autonomy under fixity; would require building reserve stocks consistent with peg and limit authorities’ ability to react to shocks.
    - If zone members become emerging economies, they may dismiss the peg or terminate cooperation or preserve it with more flexible exchange rate.
    - Deepening monetary union integration would ease softening the peg and reforming arrangement by reducing exposure to external shocks and allowing monetary autonomy.

*Source: BOX 1. THE COMPTE D’OPERATIONS, _wp0734 - BOX 1. THE COMPTE D’OPERATIONS*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp0734.pdf_