## _wp04167

## Source details

**Canonical URL:** [_wp04167](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04167.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04167.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04167.pdf.json)

---

### I. INTRODUCTION — core argument and scope
- Core focus: shortcomings in liberalization of Pakistan’s capital account via foreign currency deposits (FCD).
- Expected policy objective versus outcome:
  - Policymakers expected FCDs to mobilize domestic and external savings to boost investment; outcome: increase in potential savings from FCDs did not result in an investment boom.
  - Short-term foreign liabilities financed large fiscal and external current account deficits for a somewhat longer period than would otherwise have been possible, producing a debt overhang that constrained long-term growth prospects.
- Central concern: public debt sustainability amid high degree of dollarization and structural financial sector weaknesses.
- Operational detail of original FCD scheme:
  - Banks sold foreign exchange deposited with them to the State Bank of Pakistan (SBP); SBP provided foreign exchange cover to commercial banks and provided forward cover at subsidized rates for almost all years during the original scheme’s operation.
- Resulting dynamic:
  - Foreign exchange earnings that could have built international reserves were instead used to finance persistent fiscal and external imbalances, worsening the SBP’s open foreign exchange position as FCDs grew and reserves ran down.
- Private-sector behavior and model framing:
  - Underinsurance against international liquidity shortages amplified vulnerability; expectations about central bank action during crises affected extent of underinsurance.
  - Analytical approach employs the Caballero-Krishnamurthy dual-liquidity model (Appendix I) to interpret underinsurance and implications for policy.

### II. FOREIGN CURRENCY DEPOSITS, FISCAL DEFICITS, AND EXTERNAL DEBT — findings and mechanisms
- Historical context and export structure:
  - Pakistan experienced protracted balance of payments problems throughout the 1980s and 1990s reflecting expansionary financial policies and external sector weaknesses.
  - Export base narrow: over half of exports were from cotton and cotton-based products.
  - Workers’ remittances were sizable relative to exports, declined in the 1990s, but later reversed.
- Financing composition of current-account deficits:
  - Financed inter alia by foreign direct investment and portfolio capital, but more so by accumulation of short-term foreign liabilities from FCDs and foreign assistance by bilateral and multilateral donors.
- Operational constraints and commercial-bank incentives:
  - Commercial banks required to surrender foreign exchange from FCDs to the SBP, which provided rupees at prevailing exchange rate.
  - Commercial banks closed open forex positions by buying six-month forward cover from the SBP sold at subsidized rates (rate of depreciation exceeded cost of forward cover) for almost all years of the original scheme.
  - Result: FCDs were highly profitable for commercial banks, which favored mobilizing FCDs over rupee deposits and rolled forward cover contracts on expiry.
- Crisis chronology and restructuring:
  - External debt crisis peaked after 1998 nuclear tests, G-8 sanctions, suspension of economic assistance.
  - ESAF/EFF arrangement with IMF went off track; government suspended repayment of bilateral debt while maintaining payments to international financial institutions to enable IMF negotiations and later Paris Club debt relief and regularization of arrears.
  - Official external debt restructured to achieve medium-term debt sustainability; private creditor approach preserved face value while reducing coupon/interest rates below market rates.
- Crisis classification and balance-sheet fragilities:
  - Evolved from “first-generation” fiscal/current-account crisis into a “third generation” financial crisis reflecting fragilities in government, central bank, and commercial bank balance sheets.
  - First-generation: monetization of unsustainable budget deficits → reserve losses → abandonment of exchange rate peg.
  - Second-generation: self-fulfilling expectations; short-term dollar liabilities and SBP-provided forward cover without adequate reserves created chronic risks of self-fulfilling rollover crises.
  - Third-generation: liquidity and currency mismatches revealed vulnerabilities in corporate and financial sectors; capital-account crises can adjust in days, magnifying real effects and causing exchange-rate overshooting.
- Monetary and exchange-rate policy constraints:
  - SBP often injected liquidity (mostly inflationary or sterilized via reserve losses); used open market operations and interest-rate hikes intermittently; frequently allowed gradual rupee depreciation to preserve reserves.
  - Discrete exchange-rate adjustments raised domestic debt-service costs and worsened SBP balance sheet via costly provision of forward cover.
- Distribution of currency risk:
  - Currency risk largely borne by SBP via forward cover rather than corporations; depreciation still raised borrowing costs, credit rationing, and solvency pressures for public enterprises and private firms.
- Reserve-policy critique:
  - Authorities kept reserve holdings unduly low for cost reasons, increasing vulnerability to shocks and elevating country risk premia on external borrowing.
  - Higher proportion of owned reserves would have reduced vulnerability to abrupt reductions in borrowed reserves.

### III. FOREIGN CURRENCY DEPOSITS AND CONSTRAINTS ON FINANCIAL POLICIES — institutional effects and market segmentation
- FCDs as premature capital-account liberalization:
  - Residents allowed to hold FCDs in 1991, opening the capital account and accelerating dollarization.
- Market segmentation and multiple exchange rates:
  - FCDs contributed to dollarization and segmentation by diverting home remittances from official channels to money-changers, creating special-interest groups that profited from multiple exchange-rate practices and opposed exchange reform.
  - Use of dollars from FCDs to buy Foreign Exchange Bearer Certificates (FEBC) traded on stock market produced implicit exchange rates diverging from official rates.
- Policy-design trade-offs:
  - Original FCD scheme design (SBP forward cover at subsidized rates, compulsory surrender of foreign exchange) magnified macroeconomic and financial-policy constraints.
  - New FCD regime (post-May 1998 convertibility freeze and subsequent regulations) removed some problematic features (e.g., SBP forward cover) but rapid deposit growth and other issues persist.
- Prudential features under the new scheme:
  - Banks face limits on net foreign-exchange position (10 percent of capital).
  - FCDs must not exceed 20 percent of deposits.
  - Domestic foreign-exchange loans strictly limited to export finance.
- Trade-off from prudential caps:
  - The 20 percent cap on FCDs imposes an arbitrary ceiling on mobilizing savings into the domestic banking system; excess demand for foreign-currency-denominated savings may exit the banking system or be held as foreign cash or deposited abroad.
  - Once export-finance lending limits are exhausted, banks must keep FCD proceeds with SBP or hold them abroad, limiting domestic financial intermediation and private investment.

### IV. ANALYTICAL FRAMING (Caballero-Krishnamurthy) AND POLICY IMPLICATIONS
- Dual-liquidity model insights (from Appendix I):
  - Private sector underinsures against external crises by over-borrowing in foreign currency when international liquidity is scarce (vertical/“vertical region” case).
  - When international liquidity is binding, monetary policy cannot relax investment constraints; international-rate constraints determine investment and domestic interest rates can exceed international rates.
  - Credible SBP international-liquidity backing strengthens private incentives to self-insure.
- Policy recommendations (preserved exactly as stated in text):
  - Fiscal consolidation to restore debt sustainability.
  - Reform of FCD scheme to limit quasi-fiscal exposures and remove incentives that encourage short-term foreign liabilities without adequate reserve backing.
  - Increase owned international reserves to reduce vulnerability to sudden stops and volatile external financing costs.
  - Implement ex-ante measures to encourage private sector provision of international liquidity (i.e., reduce underinsurance), rather than encouraging FCD build-up via subsidized policies.
  - Strengthen financial sector balance sheets, improve regulatory oversight to address dollarization and maturity/currency mismatches.
  - Reorient public expenditure toward social and development spending expeditiously while managing debt dynamics.
  - Link capital account liberalization to appropriate sequencing of domestic financial reforms (including bank supervision).
  - Set international reserve targets with attention not only to foreign trade volumes but also to growth in short-term liabilities.
  - Implement measures to enhance international liquidity and to make domestic currency–denominated assets more attractive (reduce inflation; ensure adequate rates of return on domestic currency deposits).
  - Maintain vigilance on dollarization risks; SBP should monitor dollar loans and strategies to prevent NPL problems in case of further dollarization.
- Suggested debt-instrument reforms:
  - Consider indexed financial instruments: inflation-indexed debt and GDP-indexed bonds as instruments to self-insure against growth slowdowns and reduce probability of explosive debt/GDP paths.
  - Caveats: indexation can create structural rigidities and may weaken incentives to fight inflation.

### V. BOX 1 — Evolution of FCDs (selected empirical facts and dynamics)
- Historical growth and dollarization:
  - Foreign currency accounts introduced in 1973.
  - Residents allowed to hold FCDs in 1991.
  - Foreign currency deposits grew from US$26 million in 1975/76 to a peak of US$10 billion in 1998 (excluding institutional deposits).
  - Per annum growth examples: 83 percent in 1985/86; 43 percent in 1991/92.
  - FCDs represented 558 percent of foreign exchange reserves in 1997/1998.
  - FCDs amounted to 94 percent of imports in 1997/98.
  - Dollarization ratio (resident foreign currency deposits to broad money inclusive of foreign currency deposits) reached 23 percent in 1997/98.
  - Footnote: estimates of dollarization are biased downward because data on foreign exchange deposited abroad and foreign cash held by residents are not available; domestic foreign currency deposits provide a lower bound.
- Forward cover, SBP losses, and quasi-fiscal effects:
  - From January 1973 to June 1992, the State Bank provided forward cover to financial institutions without charging a fee.
  - Forward cover fee introduced effective July 1, 1992 at 3 percent per annum on U.S. dollar deposits; only minor upward revisions followed.
  - SBP incurred substantial losses from forward cover operations; net losses in excess of Rs. 13 billion in 1993/94 and 1995/96.
  - SBP targeted a profit margin for banks on foreign currency deposits of about 2 percent when setting forward cover fees.
  - Institutional attractiveness of FCDs: subsidized rates, taxation exemption (FCDs remained tax free since inception), protection from official inquiries about deposit sources, unrestricted withdrawal facilities.
- Impact on monetary, fiscal, and exchange-rate policies:
  - Dollarization lowered demand for domestic cash balances and made reserve-money demand sensitive to foreign capital movements.
  - Provision of forward cover lowered SBP profits and nontax revenues.
  - Average interest on FCDs was more than twice that on Pakistan’s medium- and long-term debt towards the end of the original scheme; debt servicing on foreign currency deposits comparable in magnitude to medium- and long-term foreign debt.
- FCDs, international reserves, and underinsurance:
  - FCDs provided foreign exchange inflows but were not net resources available to defend the rupee; build-up of FCDs hid costs in SBP’s open foreign exchange position.
  - Pakistan’s international reserves in 2001–04 increased from less than a billion dollars to more than US$12 billion.
  - Borrowed reserves incur net carrying costs equal to difference between borrowing cost (LIBOR plus spreads) and yield on reserve assets; for Pakistan these carrying costs were substantial and often in double digit levels for borrowing spreads.
- Prudential constraints under new FCD scheme:
  - Banks’ net foreign-exchange position limit = 10 percent of capital.
  - FCDs cap = 20 percent of deposits.
  - Domestic foreign-exchange lending restricted to export finance.

### VI. VULNERABILITIES, DATA, AND SUPERVISION
- Main vulnerabilities and mechanisms:
  - Large volumes of short-term FCDs and their institutional features made the economy highly vulnerable to exogenous shocks.
  - Use of FCDs to accumulate international reserves resulted in borrowed, rather than owned, reserves; required periodic refinancing and exposed reserve accumulation to volatility in financing costs.
  - Reliance on short-term capital was driven mainly by large fiscal deficits.
- Bank supervision and data needs:
  - Financial institutions should provide comprehensive quarterly reports that reflect financial positions accurately, including complete income statements showing accruals, amortization, and updated reserves for nonperforming assets.
  - SBP should develop an early-warning system for deterioration in loan portfolios based on an accurate and comprehensive database; rely initially on off-site prudential supervision and timely financial analysis of quarterly returns.

### VII. SELECTED KEY STATISTICS (preserved exactly)
- Table 2 (Foreign Currency Deposits and Official Reserves, fiscal years):
  - 1990/1991: TFCD = 2,592; International Reserves (a) = 1,390; Difference (b-a) = -1,202.
  - 1996/1997: TFCD = 9,843; International Reserves = 1,977; Difference = -7,866.
  - 2000/2001: TFCD = 2,253; International Reserves = 2,666; Difference = 413.
  - 2001/2002: TFCD = 1,194; International Reserves = 5,439; Difference = 4,245.
  - Period averages:
    - 1990/91–1999/00: TFCD = 5,888; International Reserves = 2,307; Difference = -3,581.
    - 2000/01–2002/03: TFCD = 1,247; International Reserves = 6,268; Difference = 5,022.
- Table 3 (Resident Foreign Currency Deposits and M2, fiscal years):
  - 1990/91: M2 = 368,783; RFCDs = 9,957; Ratio of RFCD to M2 (In percent) = 2.70.
  - 1996/97: M2 = 1,053,234; RFCDs = 222,882; Ratio = 21.16.
  - 1997/98: M2 = 1,206,320; RFCDs = 278,556; Ratio = 23.09.
  - 2003/04 2/: M2 = 2,312,307; RFCDs = 148,282; Ratio = 6.41.
  - Period averages:
    - 1990/91–1999/00: M2 = 873,181; RFCDs = 120,563; Ratio = 12.94.
    - 2000/01–2002/03: M2 = 1,788,728; RFCDs = 145,916; Ratio = 8.37.
- Table 1 (selected indicators):
  - 2003: GNI (percent) = 14.70; growth = 2.91; Inflation (percentage change) = 9.27; Official Exchange Rate (Rupees per dollar) = 57.22.
  - Period averages:
    - 2000–03: GNI (percent) = 9.61; growth = 3.43; Inflation = 5.85; Official Exchange Rate = 58.66.
- Table 4 (External Debt, totals and percent of GDP):
  - Total external debt projected 2003/04 = 35,452 (In millions of U.S. dollars).
  - Total external debt (In percent of GDP) period averages:
    - 1991/92–1999/00 average (labeled) = 53.2.
    - 2000/01–2003/04 average (labeled) = 54.2.

### VIII. APPENDIX I — Caballero-Krishnamurthy model (concise mechanics and implications)
- Model structure:
  - Three periods: date 0 (investment/financing), date 1 (near crisis), date 2 (repayment); two states at date 1 with probabilities π and 1-π.
  - Firms face productivity shocks; distressed plants can invest kθ goods to offset declines.
- Key analytical distinctions (selected equations and inequalities preserved in spirit):
  - Firms first borrow against net international collateral (equation (3): fww n −≡ ).
  - Horizontal-region: international finance not binding; domestic interest rate equals international interest rate (equation (5): *1 i i d = ); aggregate investment determined by firms’ constraints (equation (6)).
  - Vertical-region: international liquidity scarce; investment determined by stock of international liquidity (equation (9): *1 1 2 i w k n V + = θ); domestic interest rate exceeds international rate (equation (10): *1 1 i i d > ).
  - Banking-system amplification: declines in asset prices weaken banks’ ability to aggregate collateral, producing further asset-price declines and reduced activity.
- Policy implications highlighted by model:
  - In the vertical region, expansionary monetary policy cannot loosen the international liquidity constraint; stabilization of exchange rate by SBP without adequate reserves distorts private insurance decisions and encourages underinsurance.
  - Credible SBP international-reserve backing would strengthen private incentives to insure against international liquidity shortages.

*Source: _wp04167 (IMF PDF excerpts provided)*

### 1.     Basic Economic Indicators, 1973–2003 ..........................................................................28

### Basic Economic Indicators, 1973–2003

### Major sections (as listed)
- 1.     Basic Economic Indicators, 1973–2003 ..........................................................................28
- 2.     Foreign Currency Deposits and Official Reserves, 1990/91–2003/04 ............................29
- 3.     Resident Foreign Currency Deposits (RFCD) and Money Supply, 1990/91–2003/04 ...30
- 4.     External Debt, 1991/92–2003/04 .....................................................................................31

### Boxes and appendices
- Box 1. Evolution of FCDs .......................................................................................................11 
- Appendix  
  - I.      Caballero-Krishnamurthy      Model .....................................................................................32
- Appendix Table 
  - A1.  Caballero-Krishnamurthy Model .....................................................................................33

*https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04167.pdf*

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

### _wp04167 - References..............................................................................................................

### I. INTRODUCTION — core argument and scope
- Focus: shortcomings in liberalization of Pakistan’s capital account via foreign currency deposits (FCD).
- Policymakers expected FCDs to mobilize domestic and external savings to boost investment; outcome:
  - Increase in potential savings from FCDs did not result in an investment boom.
  - Short-term foreign liabilities financed large fiscal and external current account deficits for a somewhat longer period than would otherwise have been possible, producing a debt overhang that constrained long-term growth prospects.
- Central concern: public debt sustainability amid high degree of dollarization and structural financial sector weaknesses.
- Analytical approach: pinpoint structural weaknesses in the financial system, with particular focus on how short-term foreign liabilities affected financial policy conduct.
- Operational detail (original FCD scheme): banks sold foreign exchange deposited with them to the State Bank of Pakistan (SBP); SBP provided foreign exchange cover to commercial banks and provided forward cover at subsidized rates for almost all years during the original scheme’s operation.
- Resulting dynamic: foreign exchange earnings that could have built international reserves were instead used to finance persistent fiscal and external imbalances, worsening the SBP’s open foreign exchange position as FCDs grew and reserves ran down.
- Private sector behavior: underinsurance against international liquidity shortages (where external supply of funds is a vertical curve) amplified vulnerability; expectations about central bank action during crises affected extent of underinsurance.
- Policy error highlighted: authorities encouraged FCD build-up through preferential treatment (taxes, SBP-provided subsidized forward cover) despite underinsurance and high crisis risk.
- Organization of paper: Section II—effects of fiscal deficits on savings and external indebtedness; Section III—evolution and implications of FCDs and dollarization; Section IV—government policy responses and additional recommended steps; Section V—conclusions and priorities for fiscal consolidation and FCD reform. Appendix: Caballero-Krishnamurthy model analytics.

### II. FOREIGN CURRENCY DEPOSITS, FISCAL DEFICITS, AND EXTERNAL DEBT — findings and mechanisms
- Historical context:
  - Pakistan experienced protracted balance of payments problems throughout the 1980s and 1990s reflecting expansionary financial policies and external sector weaknesses (Table 1).
  - Export base narrow: over half of exports were from cotton and cotton-based products.
  - Workers’ remittances were a sizable proportion of total exports, declined in the 1990s, but later reversed.
- Financing of persistent current account deficits:
  - Financed inter alia by foreign direct investment and portfolio capital, but more so by accumulation of short-term foreign liabilities from FCDs and foreign assistance by bilateral and multilateral donors (Tables 2 and 3).
- Operational constraints of the FCD scheme:
  - Commercial banks required to surrender foreign exchange from FCDs to the SBP, which provided rupees at prevailing exchange rate.
  - Commercial banks closed open forex positions by buying six-month forward cover from the SBP sold at subsidized rates (rate of depreciation exceeded cost of forward cover) for almost all years of the original scheme.
  - Result: FCDs were highly profitable for commercial banks, which favored mobilizing FCDs over rupee deposits and rolled forward cover contracts on expiry.
- Implications for international liquidity and reserve accumulation:
  - Capital account opening made international liquidity broader than official reserves; authorities could use borrowed assets from private residents or international capital markets.
  - Structural change (SBP providing forward cover) complicated derivation of simple rules of thumb or rigorous quantitative projections for prudent reserve accumulation.
  - Authors avoid constructing quantitative reserve metrics; focus instead on whether larger owned reserves would have improved performance.
- Use of external borrowing:
  - External borrowing can finance investment beyond domestic saving; but in Pakistan’s case increased indebtedness reflected increased consumption rather than investment, contributing to debt-servicing problems and debt reschedulings.
- Crisis events and restructuring:
  - External debt crisis peaked after 1998 nuclear tests, G-8 sanctions, suspension of economic assistance (Table 4).
  - ESAF/EFF arrangement with IMF went off track; government suspended repayment of bilateral debt while maintaining payments to international financial institutions to enable IMF negotiations and later Paris Club debt relief and regularization of arrears.
  - Official external debt restructured to achieve medium-term debt sustainability; private creditor approach preserved face value while reducing coupon/interest rates below market rates.
- Crisis classification and balance-sheet fragilities:
  - Process described as evolving from a “first-generation” fiscal/current account crisis into a “third generation” financial crisis reflecting fragilities in government, central bank, and commercial bank balance sheets.
  - First-generation mechanism: monetization of unsustainable budget deficits → reserve losses → abandonment of exchange rate peg.
  - Second-generation aspects: self-fulfilling expectations; short-term dollar liabilities and SBP-provided forward cover without adequate reserves created chronic risks of self-fulfilling rollover crises.
  - Third-generation emphasis: liquidity and currency mismatches revealed vulnerabilities in corporate and financial sectors; capital account–driven crises can adjust in days, magnifying real economy effects and causing overshooting in exchange markets.
- Monetary and exchange rate policy constraints:
  - In presence of large government borrowing, SBP injected liquidity (mostly inflationary or sterilized via reserve losses).
  - SBP occasionally used open market operations and interest rate hikes to control liquidity; frequently allowed gradual rupee depreciation to preserve reserves.
  - Exchange market pressures sometimes resisted, leading to discrete exchange rate adjustments that raised domestic debt-service costs and worsened SBP balance sheet through costly provision of forward cover.
- Distinctive distribution of currency risk:
  - In Pakistan, currency risk was largely borne by the central bank (via forward cover) rather than corporations; but depreciation still raised borrowing costs, credit rationing, and solvency pressures for both public enterprises (rigid output prices) and private firms (short-term refinancing needs).
  - The FCD scheme contributed to firms’ reliance on periodic refinancing and heightened vulnerability to volatile external financing costs and credit rationing.
- Reserve policy critique:
  - Authorities kept reserve holdings unduly low for cost reasons, increasing vulnerability to shocks and elevating country risk premia on external borrowing.
  - Higher proportion of owned reserves would have reduced vulnerability to abrupt reductions in borrowed reserves.

### III. FOREIGN CURRENCY DEPOSITS AND THE CONSTRAINTS ON FINANCIAL POLICIES — institutional effects and market segmentation
- FCDs as premature capital account liberalization:
  - Allowed nonresidents (e.g., Pakistanis abroad) and residents to hold foreign currency accounts and use funds for current and capital account transactions previously restricted.
  - Deregulation removed many bureaucratic hurdles for obtaining foreign exchange for travel, education, medical purposes, and enabled direct external payments.
- Beneficiaries and behaviors:
  - Residents obtained a hedge against rupee depreciation and easier mechanisms for remittances, travel and international payments.
  - Residents could buy foreign exchange from authorized dealers or kerb market and deposit into FCDs for capital account transactions.
- Market segmentation and multiple exchange rates:
  - FCDs contributed to dollarization and segmentation of foreign exchange market by diverting home remittances from official channels to money-changers.
  - Created well-connected special interest groups that profited from multiple exchange rate practices and opposed exchange reform proposals.
  - Example: use of dollars from FCDs to buy Foreign Exchange Bearer Certificates (FEBC) traded on stock market, with implicit exchange rates diverging from official rates, reinforcing multiple-currency practices.
- Policy trade-offs and regulatory design:
  - Original FCD scheme design (SBP forward cover at subsidized rates, compulsory surrender of foreign exchange) magnified macroeconomic and financial policy constraints.
  - While the “new” FCD regime (post-May 1998 convertibility freeze and subsequent new regulations) has rapid deposit growth, some problems such as SBP-provided forward cover are not present in the new scheme; much of the paper’s discussion, however, focuses on problems from the original scheme.

### IV. Analytical framing and policy implications
- The Caballero-Krishnamurthy dual liquidity model is used to interpret private sector underinsurance and dependence on central bank actions during crises; Appendix provides model analytics and implications for Pakistan’s FCD scheme.
- Policy recommendations and priorities emphasized throughout:
  - Fiscal consolidation to restore debt sustainability.
  - Reform of FCD scheme to limit quasi-fiscal exposures and remove incentives that encourage short-term foreign liabilities without adequate reserve backing.
  - Increase owned international reserves to reduce vulnerability to sudden stops and volatile external financing costs.
  - Implement ex-ante measures to encourage private sector provision of international liquidity (i.e., reduce underinsurance), rather than encouraging FCD build-up via subsidized policies.
  - Strengthen financial sector balance sheets, improve regulatory oversight to address dollarization and maturity/currency mismatches.
  - Reorient public expenditure toward social and development spending expeditiously while managing debt dynamics.
- Historical note on market access:
  - Pakistan returned to international capital markets with issuance of a US$500 million five-year sovereign bond in 2004, which was considerably over-subscribed and had broad-based market presence, including European and Far Eastern investors.

*Source: _wp04167 - References (excerpts provided).*

### Box 1. Evolution of FCDs

### Box 1. Evolution of FCDs

### Historical growth and dollarization
- Foreign currency accounts introduced in 1973 to attract savings from Pakistanis working abroad and to increase foreign saving for government deficits.
- Residents allowed to hold FCDs in 1991, opening the capital account and accelerating dollarization.
- Foreign currency deposits grew from US$26 million in 1975/76 to a peak of US$10 billion in 1998 (excluding institutional deposits).
- Per annum growth examples: 83 percent in 1985/86; 43 percent in 1991/92.
- FCDs represented 558 percent of foreign exchange reserves in 1997/1998.
- FCDs amounted to 94 percent of imports in 1997/98.
- Dollarization ratio (resident foreign currency deposits to broad money inclusive of foreign currency deposits) reached 23 percent in 1997/98.
- Footnote: estimates of dollarization are biased downward because data on foreign exchange deposited abroad and foreign cash held by residents are not available; domestic foreign currency deposits provide a lower bound.

### Forward cover, SBP losses, and quasi-fiscal effects
- From January 1973 to June 1992, the State Bank provided forward cover to financial institutions without charging a fee.
- Forward cover fee introduced effective July 1, 1992 at 3 percent per annum on U.S. dollar deposits; only minor upward revisions followed, insufficient to cover costs.
- SBP incurred substantial losses from forward cover operations as rupee depreciation accelerated in the late 1980s and 1990s.
- Examples of net losses: in 1993/94 and 1995/96 net losses from forward cover operations were in excess of Rs. 13 billion.
- Exception: 1994/95 showed a small net profit on forward cover operations.
- SBP targeted a profit margin for banks on foreign currency deposits of about 2 percent when setting forward cover fees; fluctuations in domestic and international interest rates caused deviations from this target.
- Institutional features that increased FCD attractiveness: subsidized rates, taxation exemption (FCDs remained tax free since inception), protection from official inquiries about deposit sources, unrestricted withdrawal facilities.
- Resulting quasi-fiscal motive: SBP absorbed exchange risk while banks passed on part of subsidies to depositors, creating a wedge between banks’ asset and liability returns and reducing central bank profits and nontax revenues.

### Impact on monetary, fiscal, and exchange rate policies
- Monetary policy:
  - Large and changing share of FCDs altered relations between reserve money (operational target), broad money (intermediate target), and inflation control.
  - Dollarization lowered demand for domestic cash balances and made reserve money demand sensitive to foreign capital movements.
  - Defending the rupee by raising returns on domestic currency assets increased banks’ funding costs and could trigger contraction in bank lending and banking sector stress (rollover loan vulnerabilities).
  - Allowing exchange rate to float risked currency mismatches because of public sector foreign-currency-denominated debt and losses on SBP balance sheet from depreciation.
- Fiscal policy:
  - Provision of forward cover lowered SBP profits and nontax revenues.
  - FCDs eroded tax base because they were exempt from all taxes, leading to lower tax revenues and reduced seigniorage revenue for a given inflation target.
  - Average interest on FCDs was more than twice that on Pakistan’s medium- and long-term debt towards the end of the original scheme; debt servicing on foreign currency deposits comparable in magnitude to medium- and long-term foreign debt.
- Exchange rate policy:
  - Currency depreciation need not attract foreign capital inflows when residents hold substantial foreign-currency-denominated assets domestically; depreciation may instead cause substitution from FCDs into domestic assets and block a channel for reserve accumulation.
  - FCDs facilitated international transactions and transfers, increasing the risk of capital flight, substitution into foreign cash, and accelerated reserve depletion in crises.

### FCDs, international reserves, and underinsurance
- FCDs provided foreign exchange inflows but were not net resources available to defend the rupee; build-up of FCDs hid costs in SBP’s open foreign exchange position.
- Increased FCDs could raise perceived sovereign repayment priority for depositors, potentially increasing risk premia on private sovereign lending and offsetting apparent savings from FCD mobilization.
- Pakistan’s international reserves in 2001–04 increased from less than a billion dollars to more than US$12 billion, a historical high relative to imports and short-term foreign liabilities.
- Caballero-Krishnamurthy dual-liquidity framework:
  - Private sector tends to underinsure against external crises by over-borrowing in foreign currency.
  - Central bank should carry additional international liquidity to mitigate private underinsurance.
  - Pakistani authorities did not proactively manage reserves largely because of persistent fiscal deficits; the FCD scheme’s preferential features encouraged excessive growth of short-term foreign liabilities.
- Borrowed reserves versus owned reserves:
  - Borrowed reserves incur net carrying costs equal to difference between borrowing cost (LIBOR plus spreads) and yield on reserve assets; for Pakistan these carrying costs were substantial and often in double digit levels for borrowing spreads.
  - Industrial countries’ net interest cost of holding reserves averaged less than 1 percent per annum in recent years—contrast with Pakistan’s higher costs.

### Financial sector constraints under the new FCD scheme and prudential measures
- New FCD scheme removed some problematic features (e.g., SBP forward cover) present in the original scheme, but key issues remain:
  - Banks face limits on net foreign exchange position (10 percent of capital).
  - FCDs must not exceed 20 percent of deposits.
  - Domestic foreign exchange loans strictly limited to export finance.
- Consequences:
  - The 20 percent cap on FCDs imposes an arbitrary ceiling on mobilizing savings into the domestic banking system; excess demand for foreign-currency-denominated savings may exit the banking system or be held as foreign cash or deposited abroad.
  - Once export-finance lending limits are exhausted, banks must keep FCD proceeds with SBP or hold them abroad, limiting domestic financial intermediation and private investment.

### Policy recommendations and reforms discussed
- Reform public debt management and National Savings Scheme (NSS):
  - Modify NSS to provide a level playing field for commercial banks; adjust profit rates on NSS instruments to align with risk and maturity characteristics.
  - Auction Pakistan Investment Bonds (PIB) as benchmarks for NSS rates (a step taken in 2001) helped reduce public debt service.
  - Medium-term reform: eliminate on-tap instruments and auction instruments such as treasury bills and PIBs through primary dealers.
  - Reduce distortive tax treatment: income tax rates on banks were reduced from 58 percent to 50 percent in 2002, with further reductions to 44 percent for tax year 2003/04 and planned reduction to corporate tax rate of 35 percent by tax year 2006/07.
- Consider indexed financial instruments to reduce currency-mismatch risks:
  - Inflation-indexed debt and GDP-indexed bonds suggested as instruments to self-insure against growth slowdowns and to reduce probability of explosive debt/GDP paths (Borensztein and Mauro, 2002).
  - Denominating external borrowing in terms of a nontraded good can mitigate self-fulfilling liquidity crises by sharing losses across creditors and reducing incentives for runs (Chamon, 2001).
  - Caveats: indexation can create structural rigidities, inflation inertia, and may weaken incentives to fight inflation; indexing bonds reduces government gains from unanticipated inflation and does not necessarily encourage higher inflation if it reduces gains from inflationary financing.
- Broader policy stance:
  - Reduce reliance on FCDs by pursuing fiscal and monetary policies aimed at low inflation and exchange rate stability to make domestic-currency instruments more attractive.
  - Build international reserves prudently to insure against capital-account crises, considering costs of borrowed reserves and benefits of higher owned reserves.
  - Recognize trade-offs in reserve accumulation (costs via interest on borrowed reserves or import compression) versus the vulnerability reduction from higher reserves.

*Source: _wp04167 - Box 1. Evolution of FCDs*

### introduction of the innovative financial instruments faces insurmountable obstacles or has to

### _wp04167 - introduction of the innovative financial instruments faces insurmountable obstacles or has to

### Vulnerabilities and main findings
- Large volumes of short-term foreign currency deposits (FCDs) and the institutional characteristics of those deposits made the Pakistan economy highly vulnerable to exogenous shocks.
- The FCD scheme created an additional channel for government borrowing, changing the nature of fiscal sustainability when the economy is more “open” versus “closed.”
- Reliance on short-term capital was driven mainly by large fiscal deficits; strengthening financial policies remains necessary to promote internal and external stability.
- Using FCDs to accumulate international reserves resulted in borrowed, rather than owned, reserves; this required periodic refinancing and exposed reserve accumulation to marked volatility in financing costs.
- Integration with international financial markets can finance investment and provide risk diversification, but it increases the need for fiscal vigilance. Quoted: “It is clear from experience that large unsustainable fiscal deficits and open capital markets make uncomfortable bedfellows. Large fiscal deficits pose increasing risks for the financial sector as capital flows are freed up. And here I want to emphasize that deficits financed in large part by domestic creditors can be especially dangerous.” (Krueger, 2004)

### Trade-offs and policy framing
- Current controls (tight limits on banks’ net foreign exchange position and restricting domestic foreign exchange loans to export finance) control foreign exchange risks in the banking system but divert savings into low-return foreign assets.
- The trade-off: (a) lower domestic banking-system risk but savings flowing abroad into low-return foreign financial assets versus (b) higher domestic investment supporting growth but increased bank exposure to foreign exchange and credit risks.
- Emphasis on maintaining adequate international reserve holdings or using savings for domestic investment “as long as there are adequate safeguards.”

### Bank supervision and data needs
- Continued development of SBP’s ability to be well-informed in a timely way about banking trends and undesirable developments is essential.
- Financial institutions should provide comprehensive quarterly reports that reflect financial positions accurately, including complete income statements showing accruals, amortization, and updated reserves for nonperforming assets.
- SBP should intensify development of an early warning system for deterioration in loan portfolios based on a sufficiently accurate and comprehensive database, relying initially on off-site prudential supervision and timely financial analysis of quarterly returns.

### Recommendations
- Link capital account liberalization to appropriate sequencing of domestic financial reforms (including bank supervision).
- Strengthen fiscal policies to reduce reliance on short-term foreign capital and enhance financial-sector resilience.
- Set international reserve targets with attention not only to foreign trade volumes but also to growth in short-term liabilities.
- Implement measures to enhance international liquidity and to make domestic currency–denominated assets more attractive (reduce inflation; ensure adequate rates of return on domestic currency deposits).
- Maintain vigilance on dollarization risks; SBP should monitor dollar loans and strategies to prevent NPL problems in case of further dollarization.

### Key statistics and illustrative data points (preserved exactly as in source)
- Table 2 (Foreign Currency Deposits and Official Reserves, fiscal years):
  - 1990/1991: TFCD = 2,592; International Reserves (a) = 1,390; Difference (b-a) = -1,202.
  - 1996/1997: TFCD = 9,843; International Reserves = 1,977; Difference = -7,866.
  - 2000/2001: TFCD = 2,253; International Reserves = 2,666; Difference = 413.
  - 2001/2002: TFCD = 1,194; International Reserves = 5,439; Difference = 4,245.
  - Period averages:
    - 1990/91–1999/00: TFCD = 5,888; International Reserves = 2,307; Difference = -3,581.
    - 2000/01–2002/03: TFCD = 1,247; International Reserves = 6,268; Difference = 5,022.
- Table 3 (Resident Foreign Currency Deposits and M2, fiscal years):
  - 1990/91: M2 = 368,783; RFCDs = 9,957; Ratio of RFCD to M2 (In percent) = 2.70.
  - 1996/97: M2 = 1,053,234; RFCDs = 222,882; Ratio = 21.16.
  - 1997/98: M2 = 1,206,320; RFCDs = 278,556; Ratio = 23.09.
  - 2003/04 2/: M2 = 2,312,307; RFCDs = 148,282; Ratio = 6.41.
  - Period averages:
    - 1990/91–1999/00: M2 = 873,181; RFCDs = 120,563; Ratio = 12.94.
    - 2000/01–2002/03: M2 = 1,788,728; RFCDs = 145,916; Ratio = 8.37.
- Table 1 (selected indicators, end-of-period or annual as labeled):
  - 2003: GNI (percent) = 14.70; growth = 2.91; Inflation (percentage change) = 9.27; Official Exchange Rate (Rupees per dollar) = 57.22.
  - Period averages:
    - 2000–03: GNI (percent) = 9.61; growth = 3.43; Inflation = 5.85; Official Exchange Rate = 58.66.
- Table 4 (External Debt, totals and percent of GDP):
  - Total external debt projected 2003/04 = 35,452 (In millions of U.S. dollars).
  - Total external debt (In percent of GDP) period averages:
    - 1991/92–1999/00 average (labeled) = 53.2.
    - 2000/01–2003/04 average (labeled) = 54.2.
- Observations on reserve-financing dynamics:
  - Reliance on FCDs for reserve accumulation meant periods when reserves were small relative to TFCDs (e.g., large negative differences in the 1990s) and later periods where reserves exceeded TFCDs (e.g., large positive differences in early 2000s).

*Source: _wp04167 - introduction of the innovative financial instruments faces insurmountable obstacles or has to (IMF PDF content provided).*

### APPENDIX I

### APPENDIX I

### Model structure and setup
- Three periods: date 0 (investment and financing decisions), date 1 (near financial crisis), date 2 (firms repay debt).
- Two states at date 1: b and g, with probabilities π and 1- π.
- One-half of firms receive a productivity shock with probability 0.5 at date 1 that lowers output from A to a.
- Distressed plants can offset the productivity decline by investing kθ goods, generating date 2 output as in equation (1).
- Key model equations are (1)–(11) in Table A1; the exposition in the appendix highlights features used to analyze international and domestic liquidity constraints in the Pakistan economy.

### Dual liquidity intuition (domestic vs international collateral)
- Firms first borrow against net international collateral from foreigners (equation (3): fww n −≡ ).
- After exhausting international collateral, firms finance remaining needs from intact domestic firms using domestic collateral (ak).
- Horizontal view:
  - Domestic interest rate equals the international interest rate (equation (5): *1 i i d = ).
  - International finance is not binding for intact firms; distressed firms can borrow as long as their domestic financial constraint is not binding.
  - Aggregate investment H is determined by individual firms’ financial constraints (equation (6)): *1 *1 1 2 1 i w i akw k nn H + < + + = θ, where 1< H θ.
- Vertical view:
  - International liquidity is scarce; inequality in equation (4) reverses to equation (7).
  - Investment is determined by the stock of international liquidity and domestic liquidity does not appear in the investment equation (equation (9)): *1 1 2 i w k n V + = θ.
  - Interest rate on loans against domestic collateral exceeds the international rate (equation (10): *1 1 i i d > ).
  - The supply of international loans becomes vertical; credit market segmented with d1 i > *1 i but not exceeding ∆-1 (equation (11): ()*1 *1 1 1 1 i iw ak i n d >− + = ).

### Banking system variant and amplification
- Caballero and Krishnamurthy sketch a variant where collateral aggregation occurs in banks (relevant for Pakistan’s FCD problem).
- Intact firms deposit with banks; distressed firms borrow from banks. Banks intermediate foreign claims backed partially by international collateral.
- Declines in asset prices weaken banks’ ability to aggregate collateral, leading to further asset-price declines and reduced economic activity.
- When banks are intermediaries, crises manifest through banking system distress rather than direct domestic debt-market failure.

### Dollar-denominated debt and underinsurance
- Dollar debt problem framed as microeconomic underinsurance against country-wide aggregate shocks.
- Denominating external debt in domestic currency is equivalent to holding dollar debt plus insurance against exchange-rate-depreciation shocks.
- With country-level international financial constraints, purchasing such insurance is optimal even for risk-neutral agents.
- Domestic financial constraints cause private agents to undervalue the social benefits of insurance against aggregate shocks.
- Foreign lenders exacerbate underinsurance by reducing participation in domestic markets.

### Policy implications for Pakistan and the FCD scheme
- In the vertical region (international liquidity binding):
  - Expansionary monetary policy cannot loosen the international liquidity constraint; instead, the relative price of domestic liquidity falls versus international liquidity.
  - Monetary policy has distributional effects by altering allocation of returns between lenders and borrowers.
  - The SBP cannot conduct effective countercyclical monetary policy to stabilize output because distressed-firm investment is internationally liquidity-constrained.
  - Attempts to expand monetary policy would depreciate the exchange rate; SBP concerns about its open foreign exchange position and the inflation target make it favor policies supporting the exchange rate.
- Stabilizing the exchange rate in the vertical region distorts private insurance decisions:
  - If domestic agents expect limited exchange rate flexibility and smaller declines in the value of domestic liquidity relative to international liquidity, they will underinsure against international liquidity shortages.
  - If the SBP held sufficiently large international reserves to credibly demonstrate effective countercyclical monetary policy, private incentives to insure against international liquidity shortages would be stronger.

### Key model distinctions and inequalities (as presented)
- Productivity/investment relation: () AkkakA ≤∆+=θθ ~, where aA−≡∆ (equation (1)).
- Return on investment relative to international rate: 1−∆> * 1 i (equation (2)).
- Net international collateral: fww n −≡ (equation (3)).
- Horizontal-region international constraint: akw n 2 1 2 1 > (equation (4)).
- Horizontal interest parity: * 1 1 ii d = (equation (5)).
- Horizontal aggregate-investment condition: * 1 * 1 1 2 1 i w akw k nn H + < + + = θ, where 1< H θ (equation (6)).
- Vertical-region reversal: akw n 2 1 2 1 < (equation (7)).
- International liquidity scarcity condition: * 1 12 i wk n + > (equation (8)).
- Vertical aggregate-investment: * 1 1 2 i w k n V + = θ (equation (9)).
- Vertical-region domestic interest: * 1 1 1 ii d > (equation (10)).
- Bounds on segmented market domestic rate: ()* 1 * 1 1 1 i iw ak i n d >− + = (equation (11)).

*Source: APPENDIX I, _wp04167 - APPENDIX I*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04167.pdf_