## _cr0739

## Source details

**Canonical URL:** [_cr0739](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2007/_cr0739.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2007/_cr0739.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2007/_cr0739.pdf.json)

---

### I. DETERMINANTS OF INFLATION — Overview and key findings
- Inflation (end-of-period CPI) averaged "7.4 percent" during "2000–05" and is estimated at "7.0 percent in 2006".
- Early 1990s: triple digit inflation in "1993–94" and an average of nearly "20 percent during 1995–2000".
- Monetary expansion during "2000–05": broad money expanded by "32.4 percent" and reserve money increased by "16.8 percent".
- Identified supply and demand influences:
  - Relative inelasticity of food supply, especially meat products.
  - Import prices, especially petroleum products.
  - Changes in world commodity prices (copper and gold) affecting FX inflows and the exchange rate.
  - Loose government spending (wages, pensions, social transfers) as potential inflationary pressure.

### II. Recent CPI and monetary aggregates (selected statistics and volatility)
- Inflation volatility observations:
  - Peak "17.6 percent (y/y)" in June 2005 (meat shortage).
  - CPI "1.8 percent (y/y)" in July 2006 (lowest in four years).
  - CPI "6.8 percent (y/y)" at end-November 2006.
  - Core inflation as of November "6.7 percent (y/y)".
- Annual indicators (year-on-year / end-period as provided):
  - CPI: "2000: 8.1", "2001: 7.9", "2002: 1.7", "2003: 4.7", "2004: 10.6", "2005: 9.2", "2006: 6.8" (1/ Year on year; 2/ End-November).
  - Core Inflation: "2001: 4.3", "2002: 1.3", "2003: 5.2", "2004: 8.0", "2005: 3.5", "2006: 6.7" (1/ Year on year; 2/ End-November).
  - Broad Money: "2000: 17.5", "2001: 28.0", "2002: 41.9", "2003: 49.7", "2004: 20.3", "2005: 37.3", "2006: 38.3" (4/ End-October).
  - Togrog/U.S.$ (end of period, (+)/(-) appreciation/depreciation): "2000: -2.2", "2001: -0.5", "2002: -2.0", "2003: -3.8", "2004: -3.2", "2005: -1.0", "2006: +4.9" (2)5) notes: 5/ Since end-2005.

### III. Remonetization and financial deepening
- Key indicators and trends:
  - Total deposits relative to nominal GDP almost doubled between "2000 and 2006".
  - Money multiplier rose from "1.9 in 2000" to "4.8 in 2006".
  - Credit expansion to private sector outpaced deposit increases since 2000.
  - Credit-to-private-sector (end-2005 share of GDP): Mongolia "37.5 percent"; comparisons: Kazakhstan "36.3 percent", Ukraine "33.5 percent".
- Remonetization table (end-period / notes preserved):
  - Broad Money/GDP: "2000: 25.4", "2001: 29.6", "2002: 37.8", "2003: 47.7", "2004: 43.9", "2005: 46.7", "2006: 47.1" (1/ End-September).
  - Money multiplier: "2000: 1.9", "2001: 2.3", "2002: 2.7", "2003: 3.5", "2004: 3.6", "2005: 4.1", "2006: 4.8" (2/ Broad money/Reserve money).
  - Deposit/GDP: "2000: 15.5", "2001: 19.9", "2002: 28.1", "2003: 38.8", "2004: 36.5", "2005: 40.6", "2006: 40.9".
  - Loan/GDP: "2000: 6.8", "2001: 12.4", "2002: 19.0", "2003: 30.5", "2004: 33.5", "2005: 36.4", "2006: 37.3" (1/ End-September).

### IV. Empirical analysis — structural model (1993 Q1 - 2006 Q1) and VAR (1995 Dec. - 2006 May)
- Structural model: six endogenous variables estimated by OLS: CPI, BM, RM, ER, EX, IM; exogenous: OIL (Brent), COPP, GOLDP, LR, DUM2, DUM4.
  - Estimation Period: "1993 Q1 - 2006 Q1".
  - Frequency: Quarterly; Observations: "53".
  - Simulation period: "2001 Q1 – 2006 Q1".
  - Functional forms preserved (CPI = f(Constant, BM, ER, OIL(-1), COPP, CPI{-1}), etc.).
- VAR model:
  - Estimation Period: "1995 Dec.- 2006 May".
  - Frequency: Monthly; Observations: "126".
  - Variables: CPI, exchange rate, copper prices, oil prices, reserve money and broad money.
  - Variable ordering: copper prices, oil prices, exchange rate, reserve money, Broad money and CPI.
  - Lags: 2 (AIC).
- Main empirical conclusions (structural model & VAR confirmations):
  - Inflation strongly influenced by exogenous shocks to meat, copper, and oil import prices.
  - Monetary expansion and exchange rate changes are critical determinants of inflation.
  - Exchange rate has the largest influence on inflation; a 1 percent appreciation of the togrog causes a 0.90 percent decrease in inflation (annual average, structural simulations and VAR impulse responses).
  - Increases in copper prices have larger impacts on inflation than increases in oil prices; gold price impact negligible.
  - Broad money and reserve money growth significantly affect inflation, with lags (reserve money ~ two-quarter lag; broad money ~ three-month lag).
  - Statistical significance of most coefficients is not robust except for the exchange rate.

### V. RMSE goodness-of-fit and structural model policy simulation results
- RMSE (definition preserved): RMSE = SQRT((1/T)*∑(Ys-Ya)/Ya)2 * 100, with T: number of observations, Ys: simulated value, Ya: actual value.
- RMSEs (percent) — all endogenous variables less than 5 percent:
  - CPI: 1.4184
  - Broad Money: 0.5426
  - Reserve Money: 0.4823
  - Exchange Rate: 0.9622
  - Exports: 4.3878
  - Imports: 4.2135
- Structural model simulation: impact on CPI (Permanent Shocks; Impact on CPI in percent)
  - 10 % increase in copper price:
    - 1st year: 0.83
    - 2nd year: 1.25
    - 3rd year: 1.39
    - 4th year: 1.47
    - 5th year: 1.51
    - Annual average: 1.29
  - 10 % increase in gold price:
    - 1st year: -0.05
    - 2nd year: -0.04
    - 3rd year: -0.04
    - 4th year: -0.03
    - 5th year: -0.03
    - Annual average: -0.04
  - 10 % increase in oil price:
    - 1st year: 0.17
    - 2nd year: 0.36
    - 3rd year: 0.41
    - 4th year: 0.43
    - 5th year: 0.44
    - Annual average: 0.36
  - 5 % increase in reserve money:
    - 1st year: 0.06
    - 2nd year: 0.19
    - 3rd year: 0.30
    - 4th year: 0.38
    - 5th year: 0.44
    - Annual average: 0.27
  - 5 % increase in broad money:
    - 1st year: 0.18
    - 2nd year: 0.27
    - 3rd year: 0.30
    - 4th year: 0.32
    - 5th year: 0.33
    - Annual average: 0.28
  - 1 % appreciation of Togrog:
    - 1st year: -0.70
    - 2nd year: -0.94
    - 3rd year: -0.96
    - 4th year: -0.95
    - 5th year: -0.95
    - Annual average: -0.90

### VI. VAR impulse responses and variance decomposition (contributions to CPI variance)
- VAR impulse-response findings:
  - Exchange rate depreciation shock: instant and persistent impact on inflation.
  - Copper price shocks: affect inflation with about a three-month lag; responses are more lasting.
  - Oil price shocks: more rapid impact.
  - Reserve money growth: affects inflation with about a two-quarter lag; broad money with about a three-month lag.
- Variance decomposition (percentage contributions to CPI variance by variable across lags)
  - Lags 1: CPI 100.0, COPP 0.0, OIL 0.0, ER 0.0, RM 0.0, BM 0.0
  - Lags 2: CPI 98.2, COPP 0.2, OIL 0.1, ER 0.2, RM 0.3, BM 1.0
  - Lags 3: CPI 92.0, COPP 0.8, OIL 0.6, ER 2.7, RM 3.1, BM 0.9
  - Lags 4: CPI 79.8, COPP 1.0, OIL 2.4, ER 8.7, RM 7.5, BM 0.7
  - Lags 6: CPI 60.6, COPP 1.5, OIL 5.1, ER 22.1, RM 9.5, BM 1.2
  - Lags 8: CPI 51.3, COPP 6.6, OIL 5.0, ER 26.4, RM 8.8, BM 1.9
  - Lags 10: CPI 44.8, COPP 12.3, OIL 4.5, ER 24.6, RM 11.7, BM 2.2
  - Lags 12: CPI 40.8, COPP 15.6, OIL 4.6, ER 22.7, RM 14.0, BM 2.3
  - Lags 14: CPI 39.3, COPP 17.3, OIL 4.7, ER 22.0, RM 14.3, BM 2.4
  - Lags 16: CPI 38.5, COPP 18.5, OIL 4.5, ER 22.0, RM 13.8, BM 2.6
- Interpretation:
  - Exchange rate explains more than 20 percent of inflation and the impact is persistent.
  - Copper price changes contribute materially to CPI changes.
  - Reserve money growth explains more than 14 percent of inflation one year ahead.

### VII. Policy implications on inflation control
- Determinants: food supply constraints (meat) and mineral commodity prices (copper, oil) are primary drivers; money supply and exchange rate changes also significant.
- Recent low and stable inflation largely attributable to exchange rate appreciation despite commodity price hikes and sharp money supply increases.
- Policy recommendations:
  - Monetary policy should be more prudent to cushion effects of high mineral prices and the economic boom.
  - To lower inflation toward around 5 percent (2007 Monetary Policy Guidelines), monetary authorities may need to lower reserve money growth further, thereby reducing broad money.
  - Monetary policy should be pre-emptive because inflationary pressures from excessive money supply materialize with time lags and are harder to control after exchange rate depreciation, loose fiscal policy, or adverse weather.

### VIII. Mineral revenues, fiscal sustainability, and nonrenewable resource fund options
- Corporate income tax (CIT) for mining (from January 1, 2007):
  - Split rate: 10 percent on profits up to tog 3 billion and 25 percent on additional profits.
  - Loss carry-forward: up to two years, and reduce profits by no more than 50 percent.
  - Basic deduction: $2,600 per ton of copper.
- Mineral revenue elements and timing:
  - Licensing fees: small budget contribution; auctioning allowed but not used.
  - Dividends from Erdenet (51 percent state-owned) typically reach budget with a two-year delay.
  - CIT from Erdenet historically largest and most volatile share.
  - Windfall tax (introduced mid-2006) became the most important source of mineral revenue.
  - Total mineral revenues over the last decade ranged from less than 1 percent to more than 6 percent of nonmineral GDP.
- Mining operations and prospects (operational status as in source):
  - Erdenet — yes; Oyu Tolgoi — No; Tavan Tolgoi — Partially; Burenkhaan — No; Tumurtei — No; Tumurtiin Ovoo — Yes; Asgat — No; Tsagaan Suvraga — No; Bor Undur — Yes; Boroo — Yes; Gatsuurt — No; Hailaast — Yes; Baganuur — Yes.
- Mineral wealth estimation assumptions (Box II.1 highlights):
  - Total production assumptions: Copper: Erdenet: 23 747 000 tons. Ivanhoe : 41 549 000 tons. Gold: Ivanhoe: 10 336 000 oz. Others: 28 215 000 oz.
  - Prices based on WEO projections; beyond seven-year WEO forecast, copper and gold prices assumed to increase in line with world inflation (2 percent).
  - Tax regime assumed unchanged except windfall tax abolished after 2011.
  - Discount rate: Nominal rate of return (5 percent); real rate of return on financial investment (3 percent).
- NPV of Mongolia’s mineral revenues (2006–50) — in millions of U.S. dollars:
  - Copper prices return to long-run level: 2,010.3
  - Central case: 3,703.2
  - Copper prices remain at current level: 11,179.3
- NPV as percent of nonmineral GDP:
  - Copper prices return to long-run level: 193.6
  - Central case: 238.2
  - Copper prices remain at current level: 719.0
- Long-term fiscal benchmarks:
  - Constant real expenditure per year (2006–50): $113.3 million (5.9 percent of projected nonmineral GDP in 2006).
  - 2007 budget: nonmineral primary deficit of 16.3 percent of nonmineral GDP — sustainable only under most optimistic copper-price assumption; otherwise unsustainable.
  - Recommended rule of thumb: limit the nonmineral primary deficit to about 6 percent of nonmineral GDP in addition to what would be sustainable in a country without mineral wealth.
- Fund options and guidance:
  - Types: Stabilization funds, Savings funds, Financing funds.
  - Financing fund preferred for transparency and integration with consolidated government accounts; special design feature for Mongolia: the fund should not crowd out concessional loans and should invest equivalent amounts in the currency of concessional debt.
  - Funds are not substitutes for fiscal discipline; success depends on sound macroeconomic management.
  - Assessment: existing development fund weaknesses — receives only windfall tax, low saving rate (one third), arbitrary thresholds, easily circumvented provisions.

### IX. Bank of Mongolia (BOM) gold operations — monetary, reserve, income, and governance implications
- Context:
  - BOM engaged in gold operations since early 1990s; until 2001 monopolistic buyer; since 2002 commercial banks allowed to purchase gold from miners.
  - Gold production increased fivefold from 1995 to 2005, boosting international reserves.
- Monetary policy and sterilization:
  - BOM gold purchases inject reserve money unless fully sterilized.
  - Reserve money growth and NIR changes:
    - Reserve money grew by 19.7 percent in 2005.
    - NIR rose by 82.3 percent from end-2004.
    - As of end-August 2006, reserve money increased by 33.3 percent from end-December 2005; NIR increased by 70.9 percent over same interval.
    - Outstanding CBBs declined from tog 125.7 billion to tog 65.3 billion during the same period.
  - Sterilization coefficient (Box III.1):
    - Definition: ratio of change in CBBs to change in NIR.
    - Monthly data Dec. 1995 - Dec. 2005, observations: 121.
    - Elasticity estimates: short-term elasticities ~59-67 percent; long-run elasticities <80 percent (0.78 and 0.70 reported).
    - Implication: gold operations are not fully sterilized, reducing monetary policy effectiveness.
- Seasonality and reserve composition:
  - Gold purchases mostly between Q2 and Q3; most gold sold in Q4 per internal rule.
  - Share of gold in total international reserves: peaked 57.4 percent (end-November 2004), declined to zero (Dec 2005), rose 44.5 percent (end-June 2006), fell to 25.3 percent (end-September).
- Use of derivatives and risk:
  - BOM used forwards and options to hedge and to take market positions.
  - Derivatives activity exposes BOM to reputational and financial risks; recommended continuous monitoring, reporting, and independent audit.
- BOM 2005 income statement (In Millions of togrogs):
  - Total Revenue: 142,385
  - Total Expenses: 133,055
  - Net profit: 9,330
  - Gold revenue: 94,820
  - Gold-related expenses: 74,194
  - Gold net profit: 20,626
    - Trade from gold & silver: Revenue 15,865; Expenses 19,172; Net -3,307
    - Revaluation: Revenue 76,366; Expenses 52,839; Net 23,527
    - Derivative dealing: Revenue 2,589; Expenses 2,183; Net 406
  - Other revenue: 47,565; Other expenses: 58,861; Other net -11,296
  - Share of gold in BOM’s 2005 income:
    - Gold revenue: tog 94.8 billion (67 percent of total income).
    - Gold expenses: tog 74.2 billion (about 56 percent of expenses).
    - Net profit from gold transactions: tog 20.6 billion.
    - BOM’s net profit in 2005: tog 11.4 billion (largely explained by gold net profit).
- Risks and fiscal implications:
  - A sharp decline in world gold prices could reverse gains, erode BOM capital, damage credibility, and strain the budget via recapitalization needs, potentially fueling inflation expectations and weakening the togrog.
- Policy recommendations for BOM and authorities:
  - Adopt forceful measures to address negative impacts of gold operations on the monetary framework.
  - Improve monitoring and control of gold-related risks: reliable information, efficient reporting, independent audit.
  - Adopt more prudent guidelines for gold risk exposures; refrain from risky speculative positions in gold and derivatives.
  - Prepare a timely exit strategy from the gold market; consider phased withdrawal to minimize adverse effects on small producers and market absorption constraints.

### X. BOM midpoint exchange rate and market effects
- Midpoint setting practice:
  - BOM sets daily midpoint exchange rate based on lagged interbank rates; usually the simple average of previous day’s buying and selling rates reported by commercial banks.
  - The midpoint rate is used for government transactions, customs valuation, accounting and taxation.
  - BOM buying/selling rates set at five togrogs below/above the midpoint (10 togrogs trading margin), higher than commercial banks’ typical margins (~4 togrogs).
- Market share and interventions:
  - BOM’s market share declined: commercial banks purchased 22 percent of FX needs from BOM in 2002; 8 percent in 2005.
  - BOM interventions atypical and mainly smoothing; limited impact on the trend of exchange rate movements.
- Exchange rate volatility (standard deviation of first difference of logarithms of daily exchange rate):
  - 2001: 0.00064
  - 2002: 0.00066
  - 2003: 0.00068
  - 2004: 0.00088
  - 2005: 0.00103
  - 2006 1/: 0.00120 (1/ Based on data through end-July.)
- Midpoint rate affecting market rate (Box IV.1 findings):
  - Granger Causality tests (daily data Jan 2001–Jul 2006, 4 lags) indicate the BOM’s midpoint rate "seems to affect the market rate."
  - Null of no Granger Causality rejected in all three tested cases, "with p-values close to zero."
  - Conceptual channels: midpoint rate provides additional liquidity and arbitrage opportunities; being essentially the lagged market rate can help reduce volatility.
  - Caveat: BOM’s discretion to refuse FX requests may limit practical impact; Granger tests indicate forecasting causality not definitive structural causation.

*Source: IMF staff report content (content unit _cr0739).*

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

### _cr0739 - References..............................................................................................................

### I. DETERMINANTS OF INFLATION IN MONGOLIA — Overview and Key Findings
- Inflation (end-of-period CPI) averaged "7.4 percent" during "2000–05" and is estimated at "7.0 percent in 2006".
- During the early 1990s Mongolia recorded triple digit inflation in "1993–94" and an average inflation rate of nearly "20 percent during 1995–2000".
- Broad money expanded by "32.4 percent" and reserve money increased by "16.8 percent" during "2000–05", rates much higher than inflation.
- Key supply and demand influences on inflation identified:
  - Relative inelasticity of food supply, especially meat products, is a key contributor.
  - Import prices, especially petroleum products, are potential sources of inflationary pressures.
  - Changes in world commodity prices (e.g., copper and gold) can indirectly affect inflation via foreign exchange inflows and exchange rate movements.
  - Loose government spending (wages, pensions, social transfers) may have potential inflationary impacts (not main focus of the chapter).

### II. Recent CPI and Monetary Aggregates Developments (selected statistics)
- Inflation volatility:
  - Peak of "17.6 percent (y/y)" in June 2005 due to a severe shortage of meat supply.
  - CPI "1.8 percent (y/y)" in July 2006 (lowest in four years).
  - CPI "6.8 percent (y/y)" at end-November 2006.
  - Core inflation as of November "6.7 percent (y/y)".
- Annual indicators table (year on year / end-period data preserved):
  - CPI: "2000: 8.1", "2001: 7.9", "2002: 1.7", "2003: 4.7", "2004: 10.6", "2005: 9.2", "2006: 6.8" (1/ Year on year; 2/ End-November).
  - Core Inflation: "2001: 4.3", "2002: 1.3", "2003: 5.2", "2004: 8.0", "2005: 3.5", "2006: 6.7" (1/ Year on year; 2/ End-November).
  - Broad Money: "2000: 17.5", "2001: 28.0", "2002: 41.9", "2003: 49.7", "2004: 20.3", "2005: 37.3", "2006: 38.3" (4/ End-October).
  - Togrog/U.S.$ (end of period, (+)/(-) indicates an appreciation/depreciation): "2000: -2.2", "2001: -0.5", "2002: -2.0", "2003: -3.8", "2004: -3.2", "2005: -1.0", "2006: +4.9" (2)5) notes: 5/ Since end-2005.
- Meat and oil price shocks showed large movements and were associated with exchange rate dynamics (chart annotations preserved).

### III. Remonetization and Financial Deepening Indicators
- Rapid financial deepening described:
  - Total deposit in the banking system relative to nominal GDP almost doubled during "2000 and 2006".
  - Money multiplier increased from "1.9 in 2000" to "4.8 in 2006".
  - Since 2000, credit expansion to private sector outpaced deposit increases.
- Comparative credit-to-private-sector (end-2005 shares of GDP):
  - Mongolia reached "37.5 percent" (end-2005); comparison values listed: Kazakhstan "36.3 percent", Ukraine "33.5 percent".
- Remonetization indicators table (preserved end-period / specific notes):
  - Broad Money/GDP: "2000: 25.4", "2001: 29.6", "2002: 37.8", "2003: 47.7", "2004: 43.9", "2005: 46.7", "2006: 47.1" (1/ End-September).
  - Money multiplier: "2000: 1.9", "2001: 2.3", "2002: 2.7", "2003: 3.5", "2004: 3.6", "2005: 4.1", "2006: 4.8" (2/ Broad money/Reserve money).
  - Deposit/GDP: "2000: 15.5", "2001: 19.9", "2002: 28.1", "2003: 38.8", "2004: 36.5", "2005: 40.6", "2006: 40.9".
  - Loan/GDP: "2000: 6.8", "2001: 12.4", "2002: 19.0", "2003: 30.5", "2004: 33.5", "2005: 36.4", "2006: 37.3" (1/ End-September).

### IV. Empirical Analysis: Structural Model and VAR Approach
- Objective: assess whether inflation is predominantly affected by commodity prices or money supply developments; use simulation and impulse response analyses.
- Main conclusion from empirical work:
  - Inflation is strongly influenced by exogenous shocks to meat, copper, and oil import prices.
  - Monetary expansion and changes in exchange rate are also critical determinants of inflation.
- Structural model estimation details (Box I.1 — preserved specifications and data):
  - Estimation Period: "1993 Q1 - 2006 Q1".
  - Frequency: Quarterly; Number of observations: "53".
  - All variables are log-transformed; seasonal dummy variables included.
  - Six endogenous variables estimated by OLS: CPI, broad money (BM), reserve money (RM), nominal exchange rate (ER), export (EX), import (IM).
  - Model equations (functional forms preserved):
    - CPI = f (Constant, BM, ER, OIL(-1), COPP, CPI{-1})
    - BM = f (Constant, RM, ER, DUM2, BM{-1})
    - RM = f (Constant LR, D2, RM{-1})
    - ER = f (Constant, EX, IM, ER{-1})
    - EX = f (Constant, COPP, GOLDP, ER, DUM4, EX{-1})
    - IM = f (Constant, OIL, EX, CPI, IM{-1})
  - Endogenous / Exogenous variable mapping preserved:
    - Endogenous: CPI, BM, RM, ER, EX, IM
    - Exogenous: OIL (Brent), COPP (Copper price), GOLDP (Gold price), LR (Lending Rates), DUM2 (2nd quarter dummy), DUM4 (4th quarter dummy)
  - Simulation period: "2001 Q1 – 2006 Q1".
  - RMSE test applied to verify stability; policy simulation applied generating shocks to observe dynamic impacts on inflation.
- Empirical results summary:
  - Estimated coefficients generally have expected signs in most cases.
  - Inflation is positively related to increases in broad money, increases in copper and oil prices, and depreciation of the togrog.
  - Statistical significance is not robust except for the exchange rate.
  - Broad money is positively affected by an increase in reserve money and the second quarter seasonal dummy, and negatively affected by depreciation of the exchange rate.
  - Reserve money supply is positively related to increases in lending interest rates (reflecting high money demand).
  - Exchange rate influenced by exports and imports; exports positively affected by copper price increases and exchange rate depreciation.
  - Imports are insignificantly influenced by world oil prices in the estimations.
- Methodological notes preserved:
  - Central Bank Bill rates could be an alternative explanatory variable for reserve money, but lending rates used due to lack of interest rate signaling channel of CBBs.
  - Foreign remittances not included because of inaccessibility of quarterly data.

### V. Boxes and Statistical Appendices (content inventory highlights)
- Box titles present in the chapter:
  - I.1. Empirical Method Using the Structural Model
  - I.2. Empirical Method Using VAR Model
  - II.1. Assumptions Underlying Mineral Wealth Calculations
  - II.2. The Financing Fund of Timor-Leste
  - III.1. Estimation of Sterilization Coefficient
  - IV.1. Does BOM’s Midpoint Rate Affect the Market Rate?
- Statistical appendix tables cover series from "Basic Data, 2001–06" through "Nominal Exchange Rates, 2001–06" including fiscal, monetary, sectoral, and external accounts (table list preserved in the source).

*Source: _cr0739 - References..............................................................................................................*

### 9.      The simulation results of this structural model sufficiently fit the statistical

### 9.      The simulation results of this structural model sufficiently fit the statistical 

### RMSE goodness-of-fit
- RMSE definition: RMSE = SQRT((1/T)*∑(Ys-Ya)/Ya)
2
 * 100,  where T: number of observations in simulation period, Ys: simulated value,  Ya: actual value.
- RMSEs in each endogenous variable are less than 5 percent.
- Results of RMSE (percent):
  - CPI: 1.4184
  - Broad Money: 0.5426
  - Reserve Money: 0.4823
  - Exchange Rate: 0.9622
  - Exports: 4.3878
  - Imports: 4.2135

### Empirical results of policy simulations on inflation (structural model)
- General findings:
  - Inflation in Mongolia is positively affected by increases in copper and oil prices, money growth, and the depreciation of exchange rate.
  - Increases in copper prices have a larger impact on inflation than increases in oil prices.
  - Impact of gold prices on inflation seems negligible.
  - Money growth has a significant impact on inflation.

- Specific simulation impacts on CPI (Permanent Shocks; Impact on CPI in percent):
  - 10 % increase in copper price:
    - 1st year: 0.83
    - 2nd year: 1.25
    - 3rd year: 1.39
    - 4th year: 1.47
    - 5th year: 1.51
    - Annual average: 1.29
  - 10 % increase in gold price:
    - 1st year: -0.05
    - 2nd year: -0.04
    - 3rd year: -0.04
    - 4th year: -0.03
    - 5th year: -0.03
    - Annual average: -0.04
  - 10 % increase in oil price:
    - 1st year: 0.17
    - 2nd year: 0.36
    - 3rd year: 0.41
    - 4th year: 0.43
    - 5th year: 0.44
    - Annual average: 0.36
  - 5 % increase in reserve money:
    - 1st year: 0.06
    - 2nd year: 0.19
    - 3rd year: 0.30
    - 4th year: 0.38
    - 5th year: 0.44
    - Annual average: 0.27
  - 5 % increase in broad money:
    - 1st year: 0.18
    - 2nd year: 0.27
    - 3rd year: 0.30
    - 4th year: 0.32
    - 5th year: 0.33
    - Annual average: 0.28
  - 1 % appreciation of Togrog:
    - 1st year: -0.70
    - 2nd year: -0.94
    - 3rd year: -0.96
    - 4th year: -0.95
    - 5th year: -0.95
    - Annual average: -0.90

### VAR model findings and impulse responses
- Key qualitative confirmations of the structural model:
  - The togrog exchange rate has the largest influence on inflation.
  - A 1 percent appreciation of the togrog causes a 0.90 percent decrease in inflation (annual average).
  - Exchange rate shocks impact inflation with a short time lag; other shocks have more gradual impacts.
- Impulse response evidence:
  - A one percent standard deviation shock to copper prices, oil prices, money supply and the exchange rate affects inflation significantly.
  - Increases in copper prices affect inflation with about a three-month lag; oil price increases have a more rapid impact.
  - The length of the response is more lasting for copper price changes than for oil price changes.
  - Nominal exchange rate depreciation shock has an instant and persistent impact on inflation.
  - Reserve money growth affects inflation with about a two-quarter lag; broad money increase affects inflation with about a three-month lag.
- VAR model empirical details (Box I.2):
  - Estimation Period: 1995 Dec.- 2006 May
  - Frequency: Monthly
  - Number of observations: 126
  - All variables are log-transformed.
  - Model consists of 6 variables: CPI, exchange rate, copper prices, oil prices, reserve money and broad money.
  - Variable ordering: copper prices, oil prices, exchange rate, reserve money, Broad money and CPI (determined by exogenity).
  - Lags of the VAR model set as 2 by the AIC.

### Variance decomposition (explanatory contributions to CPI variance)
- Variance decomposition table (Lags; percentage contributions to CPI variance by variable):
  - Lags 1: CPI 100.0, COPP 0.0, OIL 0.0, ER 0.0, RM 0.0, BM 0.0
  - Lags 2: CPI 98.2, COPP 0.2, OIL 0.1, ER 0.2, RM 0.3, BM 1.0
  - Lags 3: CPI 92.0, COPP 0.8, OIL 0.6, ER 2.7, RM 3.1, BM 0.9
  - Lags 4: CPI 79.8, COPP 1.0, OIL 2.4, ER 8.7, RM 7.5, BM 0.7
  - Lags 6: CPI 60.6, COPP 1.5, OIL 5.1, ER 22.1, RM 9.5, BM 1.2
  - Lags 8: CPI 51.3, COPP 6.6, OIL 5.0, ER 26.4, RM 8.8, BM 1.9
  - Lags 10: CPI 44.8, COPP 12.3, OIL 4.5, ER 24.6, RM 11.7, BM 2.2
  - Lags 12: CPI 40.8, COPP 15.6, OIL 4.6, ER 22.7, RM 14.0, BM 2.3
  - Lags 14: CPI 39.3, COPP 17.3, OIL 4.7, ER 22.0, RM 14.3, BM 2.4
  - Lags 16: CPI 38.5, COPP 18.5, OIL 4.5, ER 22.0, RM 13.8, BM 2.6
- Interpretation:
  - Exchange rate explains more than 20 percent of inflation and the impact is persistent.
  - Copper price changes contribute materially to CPI changes.
  - Reserve money growth explains more than 14 percent of inflation one year ahead.

### Concluding remarks and policy implications
- Determinants of inflation in Mongolia:
  - Largely affected by food supply constraints (especially meat products) and mineral commodity prices (mostly copper and oil import prices).
  - Increases in money supply and changes in the exchange rate also have very significant impacts on inflation.
- Recent low and stable inflation:
  - Mostly attributable to the exchange rate appreciation, despite commodity price hikes and sharp increases in money supply.
- Policy recommendations:
  - Monetary policy needs to be more prudent to cushion effects of high mineral prices and the consequent economic boom.
  - To pursue lowering inflation in the medium-term to around 5 percent (as called for under 2007 Monetary Policy Guidelines), monetary authorities may need to lower reserve money growth further, thereby that of broad money.
  - Monetary policy should be pre-emptive to prevent renewed inflation, particularly because inflationary pressures from excessive money supply materialize with time lags and become more difficult to control when the exchange rate depreciates or if loose fiscal policy or adverse weather conditions occur.

### Annex: Estimation results of the structural model (selected equations and statistics)
- Log(CPI) = -1.591 + 0.011 log(BM) + 0.017 log(OIL(-1)) + 0.051 log(COPP) + 0.437 log(ER) + 0.561 log(CPI(-1))
  - t-values (in parenthesis): (-2.27) (0.26) (0.31) (0.93) (3.59) (6.37)
  - adjusted R2 = 0.983, D.W. = 1.98
- Log(BM) = 0.902 + 0.175 log(RM) - 0.232 log(ER) + 0.895 log(BM(-1)) + 0.065 DUM2
  - t-values: (3.83) (1.92) (-2.30) (15.14) (2.39)
  - adjusted R2 = 0.993, D.W. = 1.20
- Log(RM) = 0.004 + 0.001 log(1+LR/100) + 0.998 log(RM(-1)) + 0.151 DUM2
  - t-values: (0.01) (1.85) (30.62) (3.86)
  - adjusted R2 = 0.984, D.W. = 2.27
- Log(ER) = 1.049 - 0.101 log(EX) + 0.143 log(IM) + 0.814 log(ER(-1))
  - t-values: (7.31) (-2.92) (4.06) (32.76)
  - adjusted R2 = 0.977, D.W. = 1.34
- Log(EX) = -9.840 + 0.383 log(COPP) + 0.447 log(GOLDP) + 0.644 log(ER) -0.243 log(EX(-1)) + 0.301 DUM4
  - t-values: (-5.26) (1.74) (3.43) (5.68) (-1.75) (3.81)
  - adjusted R2 = 0.674, D.W. = 1.60
- Log(IM) = -0.142 - 0.003 log(OIL) + 0.658 log(EX) + 0.378 log(CPI) + 0.059 log(IM(-1))
  - t-values: (-0.30) (-0.02) (5.01) (3.54) (0.53)
  - adjusted R2 = 0.732, D.W. = 1.56

*Source: _cr0739 - 9.      The simulation results of this structural model sufficiently fit the statistical*

### 25.      The corporate income tax (CIT) regime for mining sector is the same as in other

### _cr0739 - 25.      The corporate income tax (CIT) regime for mining sector is the same as in other

### Corporate income tax (CIT) regime and sectoral implications
- Tax structure:
  - Split rate system from January 1, 2007: 10 percent on profits up to tog 3 billion and 25 percent on any additional profits.
  - Losses can be carried forward for up to two years and reduce profits by no more than 50 percent.
  - Current basic deduction fixed at $2,600 per ton of copper.
- Sectoral impacts and concerns:
  - Long delay between first investment expenditure and first profit makes the limited loss carry-forward potentially a greater concern for mining than other firms.
  - Economies of scale imply mining firms tend to be larger and are more likely to fall under the higher of the two tax rates.
  - Exploration costs may be amortized over five years (partially mitigates limited loss carry-forward).

### Mineral-related revenue elements and timing
- Revenue components:
  - Licensing fees: relatively small contribution to the budget despite increases under the Minerals Law adopted in mid-2006; play an important role in limiting speculation on licenses; auctioning of mining sites is allowed but has not occurred.
  - Dividends from Erdenet: 51 percent owned by the Mongolian state, remainder by the Russian government; Mongolian share of dividend goes into the budget usually with a delay of two years after the accounting year in which they were earned.
  - CIT revenues from Erdenet have historically accounted for the largest share and most volatility.
  - Windfall tax: introduced in mid-2006 and since then has become the most important source of mineral revenue.
- Volatility:
  - Total mineral revenues over the last decade have ranged from less than 1 percent to more than 6 percent of nonmineral GDP.

### Current and prospective mining operations — production, firms, and tax exposure
- Major firms and timelines:
  - Erdenet: exploiting a major copper mine since the 1970s; expects to continue extraction at the current rate for a further 50 years; historically pays out two-thirds of profits as dividends with significant lags (typically reaching the budget two years after the tax year).
  - Ivanhoe: expected to start extracting copper and gold from 2009 and to surpass Erdenet as the largest copper producer from around 2011; uncertainty about government equity participation and whether Ivanhoe will be liable for the windfall tax (on-site smelter would apparently exempt company from the windfall tax as it applies only to raw copper ores).
- Gold sector:
  - Over 100 companies of different sizes.
  - Gold companies to date have not paid CIT due to tax holidays or exemptions.
  - Boroogold: exempt from the windfall tax; agreed to pay an additional US$10 million this year and to begin paying CIT at the full income tax rate from 2007; negotiating a new investment contract for another deposit — likely to pay taxes on profits from the new deposit from the beginning given the imminent abolition of tax holidays.

### Mineral reserves and major deposits
- Mapping coverage and uncertainty:
  - Only 15 percent of the total area has been geographically mapped (USGS 2004); estimates of proven and potential resources are very approximate.
- Notable reserve figures:
  - Baganuur estimated reserves include 6 billion tons of coal and 2.4 billion tons of coke.
- Select operating and prospective mines (operational status preserved exactly as in source): Erdenet (Copper, molybdenum) — yes; Oyu Tolgoi (Copper, gold) — No; Tavan Tolgoi (Coking coal) — Partially; Burenkhaan (Phosphate) — No; Tumurtei (Iron) — No; Tumurtiin Ovoo (Zinc) — Yes; Asgat (Silver) — No; Tsagaan Suvraga (Copper) — No; Bor Undur (Fluorspar) — Yes; Boroo (Gold) — Yes; Gatsuurt (Gold) — No; Hailaast (Gold) — Yes; Baganuur (Coal) — Yes.

### Mineral wealth estimation — assumptions and scenario results
- Key assumptions (Box II.1):
  - Total Production:
    - Copper: Erdenet: 23 747 000 tons. Ivanhoe : 41 549 000 tons.
    - Gold: Ivanhoe: 10 336 000 oz. Others: 28 215 000 oz.
  - Prices: Copper and gold price estimates based on World Economic Outlook (WEO) projections; beyond the seven-year WEO forecast period, copper and gold prices assumed to increase in line with world inflation (i.e., 2 percent).
  - Tax regime: assumed unchanged as described above, except the windfall tax assumed to be abolished after 2011.
  - Discount rate: Nominal rate of return (5 percent); real rate of return on financial investment (3 percent).
- Sensitivity to price scenarios:
  - Three scenarios considered for copper prices: central (WEO-based), high (prices remain at 2006 level), low (prices fall back immediately to long-run average of $2,000 per ton).
- Net Present Value (NPV) of Mongolia’s mineral revenues (2006–50) — in millions of U.S. dollars:
  - Copper prices return to long-run level: 2,010.3
  - Central case: 3,703.2
  - Copper prices remain at current level: 11,179.3
- NPV as percent of nonmineral GDP:
  - Copper prices return to long-run level: 193.6
  - Central case: 238.2
  - Copper prices remain at current level: 719.0
- Notes on price assumptions:
  - Low price case assumes copper prices of $2,000 per ton until 2011 (average price of the last ten years).
  - High price case assumes copper prices of $6,300 until 2011 (WEO assumption for 2006).
  - Beyond 2011 copper prices grow by 2 percent per year in all scenarios.

### Long-term fiscal sustainability and constant real expenditure benchmark
- Rationale:
  - Use of the nonmineral primary balance to remove commodity-price cyclicality and reveal sustainable fiscal position.
  - Sustainable nonmineral primary deficit ceiling defines permissible long-term deficit that does not exhaust assets.
- Constant real expenditure approach:
  - Discounted government net worth multiplied by assumed real rate of return yields a fixed annual constant dollar amount that can be spent indefinitely.
  - Estimated constant real expenditure per year during 2006–50: $113.3 million (5.9 percent of projected nonmineral GDP in 2006).
  - Trajectory: fixed annual dollar amount remains constant in real terms but declines relative to nonmineral GDP over time (by 2050 it declines to 0.6 percent of nonmineral GDP).
- Assessment of the 2007 budget:
  - 2007 budget proposal budgeted to achieve a nonmineral primary deficit as a share of nonmineral GDP of 16.3 percent.
  - Sustainability implication:
    - This would be sustainable under the most optimistic copper price assumption only.
    - If copper prices drop back as expected under the WEO, in 2007 more than twice as much of the mineral wealth will be spent as would be possible in the long run.
    - If prices return to their long-term average, three times the sustainable expenditure would be spent.

### Institutional arrangements and nonrenewable resource fund options
- Rationale for funds:
  - Governments should prudently save some mineral revenues in years of exceptionally high prices/production and may spend more than current revenues in years of low prices or declining production.
  - Political difficulties in justifying budget surpluses motivate some countries to set up special commodity funds.
- Main types of nonrenewable resource funds:
  - Stabilization funds:
    - Aim to smooth volatility of commodity prices.
    - Payments/withdrawals can be triggered by market price relative to a reference price; reference price must be adapted over time.
  - Savings funds:
    - Aim to save a share of exhaustible resource revenues for future generations.
    - Withdrawals typically restricted by tight rules (e.g., limited to the real return on assets).
    - Can combine elements of stabilization and savings if proportions vary.
  - Financing funds:
    - Receive all revenues from mineral resources; withdrawals finance budget deficits and remaining funds are invested.
    - Achieve transparency without rigid fiscal restrictions; prevent simultaneous saving inside the fund while issuing debt.
- Fiscal rules:
  - Alternative or complement to funds; could specify overall balance not exceed the real return on mineral wealth, or restrict nonmineral primary balance to a fixed share of nonmineral GDP.
  - Both funds and fiscal rules can be circumvented; success depends on fiscal discipline and sound macroeconomic management more than institutional form.
- Design guidance:
  - A well-designed fund could help achieve sound fiscal objectives, but funds are not guarantees; prudent fiscal policy and macroeconomic management remain essential.

*Italic: Source — IMF staff report content (content unit _cr0739).*

### 47.      The fund should be coherently integrated into the budget process. This is best

### 47.      The fund should be coherently integrated into the budget process. This is best

### Integration with the budget and institutional design
- The fund should operate only as a government account rather than a separate institution.
- Budget formulation and reporting should focus on the consolidated presentation and expenditure should be executed by the Treasury.
- The fund should ideally be a “financing” fund, where the fund’s balance reflects government saving of its mineral wealth and is presented in the context of all the government’s financial assets and liabilities.
- Because of Mongolia’s access to highly concessional debt, the fund should not finance the part of the overall deficit that can be financed by concessional debt. Instead the fund should invest equivalent amounts in the currency of the debt to reduce risks.

### Asset management and investment strategy
- Fund assets should be prudently managed, coordinated with other government financing operations and largely invested offshore.
- Rationale: Funds can accumulate large amounts of resources and excessive risk-taking would not be appropriate.
- Holding the assets offshore reduces the impact on the domestic economy, such as through the exchange rate.
- Funds should not lend or otherwise encumber their assets.

### Transparency, accountability and oversight
- The rules and operations of the fund should be transparent with stringent mechanisms to ensure accountability and prevent misuse.
- Requirements:
  - Regular and frequent disclosure and reporting on the principles governing the fund, its inflows and outflows, and the allocation and return on assets.
  - The fund’s activities should be audited by an independent agency.
  - Investment performance should be periodically evaluated.

### Assessment of the existing development fund
- The current development fund is unlikely to play a useful role in shaping a sustainable fiscal policy.
- Reasons:
  - The fund receives just the windfall tax rather than all mineral revenues.
  - If copper prices fall as expected, it will receive only negligible funds from 2008, when the copper price is expected to fall below the threshold for windfall tax liability, and only the smaller windfall tax revenue from gold would continue.
  - Given the arbitrary thresholds for the windfall tax, it is unlikely that these revenues represent the entire windfall from temporarily high copper prices.
  - The low savings rate of one third would not achieve much stabilization or intergenerational equity.
  - The provisions of the fund can easily be circumvented, by using its components to replace other spending or saving, so that its main impact may be on administrative costs.

### Fiscal policy implications and recommendations (Conclusion)
- Mongolia will receive substantial revenue from its mineral resources in the coming years.
- Authorities should undertake regular estimates of the country’s mineral wealth; the estimates in this paper provide a first approximation.
- Fiscal policy should take the mineral wealth into account when setting targets for the fiscal balances.
- A rule limiting the nonmineral primary balance to the real return on the mineral wealth would ensure both stabilization and intergenerational equity.
- Even if the government does not fully follow that approach, the calculation can still serve as an important input into fiscal policy considerations.
- For practical reasons, the government could initially work with a shorter term horizon and target a nonmineral primary balance that would yield stable debt ratios over the medium term.
- The paper finds that the authorities’ current fiscal strategy envisaged in the 2007 budget would be sustainable only under very optimistic copper price assumptions. Any fall in copper prices would put the government on an unsustainable path, with enormous adjustments necessary, if copper prices follow instead the WEO assumptions.
- Recommended rule of thumb: limit the nonmineral primary deficit to about 6 percent of nonmineral GDP in addition to what would have been sustainable in a country without any mineral wealth.
- Any additional funds received from temporary mineral price swings or concessional loans should be saved.
- While not indispensable, a special fund could help achieve a sustainable fiscal policy. Based on international experience, a financing fund would be ideal, though with the particular feature for Mongolia that the fund should not crowd out concessional loans.
- The detailed design of the fund is important for its success; some features of the existing development fund raise doubts that it can play a useful role.

### Transition to Bank of Mongolia (BOM) gold operations — key contextual points
- The document next shifts to discuss the Bank of Mongolia’s gold operations (beginning Section III), noting:
  - The BOM has been engaged in gold operations since the early 1990s to provide liquidity and economies of scale.
  - Until 2001 the BOM acted as a monopolistic buyer; since 2002 designated commercial banks have been allowed to purchase gold from local miners.
  - Gold production increased fivefold from 1995 to 2005, contributing to a large increase in Mongolia’s international reserves.
- Main concerns outlined for BOM gold operations include potential interference with monetary policy (without adequate sterilization), governance issues (quality and transparency of risk management), and the risk that adverse gold price movements could erode the BOM’s capital base.

*Source: IMF country report text (excerpts as provided).*

### 66.      Despite these measures, the BOM’s gold trading activities have given rise to

### _cr0739 - 66.      Despite these measures, the BOM’s gold trading activities have given rise to

### Impact on the Conduct of Monetary Policy
- BOM’s gold purchases, without full sterilization, lead to reserve money injection into the economy.
- Reserve money growth figures:
  - Reserve money grew by 19.7 percent in 2005.
  - Net foreign assets (NIR) rose by 82.3 percent from end-2004.
  - As of end-August 2006, reserve money increased by 33.3 percent from end-December 2005.
  - NIR increased by 70.9 percent over the same interval.
- Outstanding CBBs issued declined from tog 125.7 billion to tog 65.3 billion during the same period.
- Conclusion: Sterilization through CBB issuance was not sufficient to mop up excess liquidity generated by BOM’s gold purchases, potentially dampening monetary policy effectiveness and preventing the BOM from meeting its inflation objective.

### Estimation of Sterilization Coefficient (Box III.1)
- Sterilization coefficient defined as the ratio of the change in CBBs to the change in NIR.
- Regression framework:
  - Frequency: monthly data
  - Sample period: Dec. 1995 - Dec. 2005
  - Number of observation: 121
  - Estimation method: OLS
- Main empirical results:
  - Elasticity of changes in CBBs with respect to changes in NIR about 59-67 percent.
  - Long-run elasticities less than 80 percent (0.78 = 0.324/(1-0.584) in equation 2 and 0.70 = 0.294/(1-0.579) in equation 4).
  - Short-term elasticities estimated around 60-70 percent; long-run elasticities less than 80 percent.
- Implication: BOM’s gold operations are not fully sterilized, reducing monetary policy effectiveness.

### Seasonality and Money Supply Implications
- BOM’s gold operations display seasonal pattern since 2003:
  - Most gold purchases made between the second and the third quarter.
  - Most gold sold in the fourth quarter (as required by internal rule).
- Seasonality partly reflects mining and agriculture cycles.
- Money supply may expand further in response to BOM’s increasing gold purchases given weak coordination with monetary policy operations.

### Composition of International Reserves and Reserve Management
- Share of gold in Mongolia’s total international reserves:
  - Peaked at 57.4 percent at end-November 2004.
  - Declined to zero in December 2005.
  - Rose to 44.5 percent at end-June 2006.
  - Fell to 25.3 percent at end-September (year not explicitly restated in this excerpt).
- Comparative note: South Africa’s share of gold in total international reserves stood at around 10 percent at end-September 2006.
- Use of derivatives:
  - BOM relied on derivative instruments (forwards and options) to hedge gold price risk.
  - BOM also used derivatives to take market positions to maximize returns, exposing it to significant risks to reputation and credibility.
- Recommendation: Monitor BOM’s risk exposure continuously through reliable information, efficient reporting system, and an independent audit function.

### Costs and Benefits of Holding International Reserves
- Costs proxied by:
  - Domestic interest rates or GDP growth rates (because BOM pays interest on CBBs used to sterilize reserves).
  - Interest rate differential regarded as a net cost of holding international reserves.
- Benefits proxied by:
  - Adjustment cost avoidance from external imbalances.
  - Propensity to import (higher propensity implies greater benefit from holding reserves).
- Empirical observations:
  - Spread between domestic and international interest rates remains very high (though not as large as in the past).
  - Propensity to import has been decreasing rapidly, implying diminishing need for BOM to build excessive international reserves at high cost.

### BOM’s Income Statement Effects from Gold Operations
- Three channels by which gold operations affect income statement:
  - (i) Profit or loss from spot trading of monetary and nonmonetary gold.
  - (ii) Revaluation.
  - (iii) Profit or loss from derivative transactions.
- Components of spot trading income:
  - Fees charged on domestic gold sellers for gold insurance, shipping, and refining cost.
  - Gains or losses from OTC buy/sell transactions.
- Revaluation:
  - Bulk of income/expenses from gold operations comes from revaluation (gold price changes and exchange rate movements).
  - Gold revaluation was main contributor to BOM’s income surplus in 2005.
- Derivatives:
  - BOM usually sells forwards and calls options in OTC market to hedge gold position.
  - In 2005, BOM realized a profit of about US$2 million from option premiums.
  - Unrealized derivatives marked to market; realized upon expiry.
  - BOM conducts derivatives on behalf of gold mining companies and levies a fee: 5 percent of the premium for a call option and forward gain, respectively.
  - Eligibility: Only gold mines with an annual gold production of more than 500 Kg within the last 2 years are eligible for these transactions.
  - These operations recorded as receivables or payables from mining companies and could impair BOM’s reputation and potentially breach the Central Bank Law.
- Income statement figures for 2005 (In Millions of togrogs):
  - Total Revenue: 142,385
  - Total Expenses: 133,055
  - Net profit: 9,330
  - Gold revenue: 94,820
  - Gold-related expenses: 74,194
  - Gold net profit: 20,626
    - Trade from gold & silver: Revenue 15,865; Expenses 19,172; Net -3,307
    - Revaluation: Revenue 76,366; Expenses 52,839; Net 23,527
    - Derivative dealing: Revenue 2,589; Expenses 2,183; Net 406
  - Other revenue: 47,565; Other expenses: 58,861; Other net -11,296
- Share of gold in BOM’s 2005 income:
  - Total BOM revenue from gold operations: tog 94.8 billion (67 percent of BOM’s total income).
  - Gold-related expenses: tog 74.2 billion (about 56 percent of BOM’s expense).
  - Net profit from gold transactions: tog 20.6 billion.
  - BOM’s net profit in 2005: tog 11.4 billion (largely explained by gold net profit).

### Risks and Fiscal Implications
- Heavy reliance on gold operations for BOM income poses risks:
  - A sharp decline in world gold prices could reverse gains, erode BOM’s capital base, damage credibility, and threaten financial system stability.
  - BOM recapitalization would strain the budget, potentially fuel inflation expectations, weaken confidence in the togrog, and worsen debt sustainability.

### Conclusions and Policy Recommendations
- Key concerns:
  - Gold operations inject high-powered money; sterilization via CBBs insufficient to mop up liquidity.
  - This undermines monetary policy effectiveness and jeopardizes inflation-target / price stability objectives.
  - Reserve management faces uncertainty and risks from substantial and volatile gold holdings and derivative use.
  - High domestic interest rates increase the cost of holding international reserves while propensity to import declines.
  - Heavy reliance of BOM income on gold trading raises central bank governance concerns.
- Recommended actions for BOM and authorities:
  - Adopt more forceful measures to address negative impacts of gold operations on monetary framework.
  - Improve monitoring and control of gold-related risk exposures through reliable information, efficient reporting, and independent audit.
  - Adopt more prudent guidelines for gold risk exposures and refrain from entering risky speculative positions in gold and other speculative financial derivatives.
  - Begin preparing a timely exit strategy from the gold market while minimizing adverse effects on the gold mining sector, especially small producers lacking export capacity.
  - Consider phased withdrawal given limited capacity of domestic banks and market participants to absorb immediate BOM exit.

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2007/_cr0739.pdf*

### 83.      The BOM sets its daily

### _cr0739 - 83.      The BOM sets its daily

### Midpoint exchange rate: setting rule and practice
- The BOM sets its daily midpoint exchange rate based on lagged exchange rates at the interbank market.
- The midpoint exchange rate of the togrog against the U.S. dollar is set and announced daily in late afternoon, usually calculated as the simple average of the buying and selling rates from the previous day’s transactions reported by commercial banks.
- This rate is then used as the next day’s BOM policy rate.
- Sometimes the midpoint rate is set differently from this rule, especially whenever the BOM decides to keep the BOM rate unchanged, e.g., when the BOM concludes that market fluctuations are temporary.
- Today’s exchange rates from the largest foreign exchange bureaus are also collected (dubbed “curb rates”) as a reference for setting the midpoint rate.
- The BOM rate broadly follows the interbank rate with a lag.

### Uses and scope of the midpoint rate
- The midpoint rate is only used in government transactions and customs valuation, for accounting and taxation purposes.
- All other transactions, including sales of retained foreign exchange by public sector enterprises, take place through the interbank market.
- Exchange rates for other convertible currencies are calculated on the basis of the cross rates against the dollar in international markets.

### Interventions, gold operations, and reserve effects
- Interventions are officially limited to smoothing high-frequency fluctuations in the foreign exchange market, using the U.S. dollar as the principal intervention currency.
- Because the BOM can accumulate foreign reserve through its gold operations, there is usually no need for the BOM to purchase foreign exchange from the market for that purpose.
- Indeed, the BOM has been a net seller of foreign exchange in the interbank market.
- Numerical lines as reported in source:
  - 2002200320042005
  - BOM gold purchase118158137210
  - Increase in net international reserves66-9737131
  - Source: Data provided by the Bank of Mongolia.
  - BOM's Gold Purchases and Reserve Accumulation (In millions USS$)
- The BOM has deliberately avoided intervening in the interbank market in recent years. Although there were episodes of deliberate intervention in the past, in recent years the BOM’s interventions, which are not sterilized, have been limited to fulfilling requests from commercial banks.
- The BOM also claims that no interventions for smoothing purposes were made, probably reflecting broad stability of the exchange rate, with the togrog only depreciating by 6 percent against the U.S. dollar during 2001–end-July 2006 and the real effective exchange rate appreciating by 7 percent over the same period.

### Pricing margins, market access, and market share
- The BOM’s buying and selling rates are set at five togrogs below and above the midpoint rate.
- This trading margin of ten togrogs is high compared with those of the commercial banks, which are usually four togrogs or less for transactions with business clients.
- The BOM has deliberately set a large margin to reduce commercial banks’ incentives to trade foreign exchange with the BOM. Occasionally, a different margin (e.g., eight togrogs) was applied.
- The BOM’s market share in Mongolia’s foreign exchange market has declined significantly in recent years.
  - In 2002 commercial banks purchased 22 percent of their foreign exchange needs from the BOM; this share declined to 8 percent in 2005.
- Commercial banks generally only approach the BOM when they are under liquidity constraints, which has led to a decline in the BOM’s market share in recent years.
- Increased foreign exchange inflows in recent years have eased commercial banks’ liquidity constraints, as a result banks rely on the BOM to a lesser extent to meet their liquidity requirements.

### Exchange rate volatility and seasonality
- Exchange rate volatility has increased slightly since 2004.
- Exchange rate volatility (As measured by the standard deviation of the first difference of logarithms of daily exchange rate):
  - 2001: 0.00064
  - 2002: 0.00066
  - 2003: 0.00068
  - 2004: 0.00088
  - 2005: 0.00103
  - 2006 1/: 0.00120
  - 1/ Based on data through end-July.
  - 2/ As measured by the standard deviation of the first difference of logarithms of daily exchange rate.
- The seasonality of the togrog exchange rate against the U.S. dollar appears to have dampened somewhat recently.
  - The togrog usually appreciates in the first half of the year, in response to sales of raw materials (e.g., cashmere, skin and hides) in the spring and inflows of financing for mining operations (usually in the second quarter), and depreciates in the second half of the year, as mining companies pay off loans toward the end of the year.
  - With the increase in foreign exchange inflows in recent years, in particular FDI and tourism receipts (which peaks after June), this pattern seems to have dampened somewhat recently, as the depreciation of togrog did not pick up until August in 2005, and the togrog’s appreciation during 2006 has lasted longer.

### Access constraints and discretionary refusals
- The BOM may reject foreign exchange requests from commercial banks.
- Occasionally, the BOM declines banks’ requests, if it considers they are for arbitrage profits. The BOM may also refuse to sell foreign exchange to banks if the BOM considers the level of its international reserves to be low.
- No explanations are given to commercial banks, however, when their requests for foreign exchange are declined.
  - Footnote: From September 2005 to July 2006, this allegedly happened in three days.

### Impact and effectiveness of interventions
- The impact of the BOM’s intervention on the market rate is limited.
- The BOM’s interventions are atypical of central banks’ interventions, whereby the central banks buy or sell foreign exchange in order to affect the exchange rate level.
- The BOM intervened in the foreign exchange market in 2005 with eight purchases of U.S. dollars totaling $27 million and 74 sales totaling $81 million.
- In general the exchange rate did not move in the anticipated direction (in the near future). This reflects the passive nature of the BOM’s interventions, with U.S. dollar purchases generally made during periods of U.S. dollar depreciation (i.e., when there is an oversupply of U.S. dollars in the market) while U.S. dollar sales generally occurring during periods of U.S. dollar appreciation.
- The interventions may have reduced the magnitude of the appreciation/depreciation (in other words, the volatility), but have not fundamentally changed the trend of exchange rate movements.

*Source: Data provided by the Bank of Mongolia.*

### 92.      The BOM’s midpoint rate seems to affect the market rate. At this rate adjusted by

### _cr0739 - 92.      The BOM’s midpoint rate seems to affect the market rate. At this rate adjusted by

### Key finding
- The BOM’s midpoint rate "seems to affect the market rate."
- Granger Causality tests suggest the midpoint rate "seems to affect the market rate."
- The null hypothesis of no Granger Causality can be rejected in all three tested cases, "with p-values close to zero."

### Conceptual mechanisms described
- At the midpoint rate (adjusted by trading margins), the BOM:
  - Provides the market with an additional source of liquidity.
  - Provides arbitrage opportunities.
  - Conceptually should affect the market rate.
- Because the midpoint rate is "essentially the lagged market rate," it should help reduce market volatility.
- A countervailing factor: the BOM "has the option to decline foreign exchange requests from commercial banks," so the impact of the midpoint rate on the market rate, if any, "could be rather limited."

### Granger Causality test (specification and variants)
- Test equation used:
  y_t = α_0 + α_1 y_{t-1} + ... + α_4 y_{t-4} + β_1 x_{t-1} ... + β_4 x_{t-4} + ε_t.
- Two versions of y_t:
  - (i) the level of the interbank rate;
  - (ii) the first difference of the interbank rate.
- For x_t:
  - When y_t is the level, the level of the BOM rate is used as x_t.
  - When y_t is the first difference, two versions of x_t are used:
    - (i) the first difference of the BOM rate; and
    - (ii) the difference of the BOM rate and the lagged interbank rate.

### Data and implementation
- Frequency: Daily data.
- Sample period: January 2001 to July 2006.
- Lags: 4 lags used in the regressions (results reported as robust to different lag lengths).
- Purpose: "to test if the current and/or lagged BOM rates can help explain current interbank rates, after controlling for lagged interbank rates."

### Interpretation and caveats
- Statistical evidence from Granger tests is "consistent with the hypothesis that the BOM rate affects the interbank rate."
- Caution: "Granger Causality tests are best described as tests of whether x helps forecast y rather than tests of whether x causes y," so one needs to be cautious in concluding causation.
- Institutional constraint: the BOM’s discretion to refuse FX requests may limit the practical impact of the midpoint rate despite statistical associations.

*Box IV.1. Does BOM’s Midpoint Rate Affect the Market Rate?*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2007/_cr0739.pdf_