## _cr1610 ## Source details **Canonical URL:** [_cr1610](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr1610.pdf) ## Other formats - [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr1610.pdf.md) - [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr1610.pdf.json) --- ### 1. Macroeconomic and Fiscal Implications of the LNG Projects #### A. Background and context - Project location and scale - Rovuma offshore gas fields in Northern Mozambique (near the border with Tanzania) are divided into Area 1 (lead concessionaire: Anadarko) and Area 4 (lead concessionaire: ENI). - Total gas reserves are estimated at about 180 trillion cubic feet. - The total investment for the projects is projected to exceed $100 billion. - Once gas production reaches its peak, Mozambique could become the third largest liquefied natural gas (LNG) exporter in the world after Qatar and Australia. - Project timing and structure - Final Investment Decisions (FID) for the first round of liquefaction and processing facilities (“trains”) are expected to be taken by mid-2016. - Area 1 proposed an integrated upstream–midstream structure; Area 4 proposed segmented upstream and midstream activities. - Proposed project structures include creating Special Purpose Vehicles in other countries for marketing and financing. - Outstanding issues before FID include revision of existing exploration contracts and resettlement of domestic residents. - Expected production profile - Staff assumes production and exports of LNG will start in 2021, with production volume gradually scaled up during the 2020s. - Area 1 initially plans two onshore trains, each producing about 5.5 million tons of LNG per annum. - Area 4 will primarily build a floating liquefaction facility (FLNG) producing about 3 million tons of LNG per annum. - Total expected build-out: 13 onshore trains and 4 floating trains. - Total LNG production could reach 89 million ton per annum by 2028. - Pricing and cost considerations - Reserves are located under the deep ocean floor, which could increase drilling and infrastructure costs relative to onshore fields. - Developers may benefit from economies of scale (fewer wells). - Construction costs of liquefaction facilities could decrease if other projects stop because of low commodity prices. - Operators assume hydrocarbon prices will increase somewhat over the medium term, but abundant ongoing LNG projects worldwide (including Australia and the US) may put downward pressure on prices. - In 2015, East Asian gas prices (primary targeted market) have already dropped substantially. #### B. Macroeconomic and fiscal implications (FARI model results and key projections) - Methodology - Assessment updated using the Fiscal Analysis of Resource Industries (FARI) model developed by the IMF’s Fiscal Affairs Department. - Note on price assumption: the assumption of gas prices is based on oil price projections in the latest World Economic Outlook, and a constant coefficient of 0.14 is applied to calculate gas prices from the oil prices. - Growth - The average real GDP growth rate between 2021 and 2025 could reach 24 percent. - Share of the LNG projects in total nominal output could reach more than 50 percent by the mid-2020s. - After peak LNG production in 2028 (with the final train starting operation), real GDP growth will moderate to 3-4 percent. - This moderation reflects underlying growth of 6-6½ per-cent of the non-LNG economy and no further growth of the LNG sector. - The LNG share in total output would gradually decline starting in the late 2020s. - Fiscal revenues - Total fiscal revenues from the LNG project throughout the entire project period until 2045 could reach about $500 billion. - Main revenue sources: (i) government’s share of profit gas, (ii) corporate income tax on concessionaires, and (iii) dividends paid by the state-owned hydrocarbon company (ENH), which owns a 15 percent share of Area 1 and a 10 percent share of Area 4. - Fiscal revenues during the first few years after production starts will be limited because of large cost recovery for continuous investments in building liquefaction plants. - By the late 2020s, fiscal revenues from the gas projects could account for more than half of total fiscal revenues. - Policy implication highlighted: authorities should be aware of the time lag between gas production and fiscal revenue flows when planning medium-term fiscal strategy. - Balance of payments - Mozambique will experience unprecedented large current account deficits during the late-2010s, peaking at more than 90 percent of GDP in 2020, due to massive investments in LNG trains. - Once gas production begins, the current account could gradually improve and reach surplus in 2025. - During the early 2020s, current account will remain in deficit because of investments in subsequent trains. - By the mid-2020s, LNG will account for about ¾ of total exports. - Staff assume that increased public investments and reforms will improve external competitiveness of the non-mining industries and crowd in private investment, increasing exports of other products from late-2030s onward. - Under the assumption that risks of a Dutch-disease induced exchange rate appreciation can be contained, the overall current account would remain in surplus until the end of LNG production. #### C. Risks and sensitivity - Price risk - Sales prices of gas could drift below baseline assumptions if global gas prices decline further (e.g., due to an enduring global slowdown or if other global LNG projects meet demand). - Fiscal revenues are particularly sensitive to lower gas prices because most revenues are related to project profits. - Fiscal regime and revenue risk - Changes in taxes and fiscal regimes could significantly affect fiscal gains. - The 2014 Decree Law ensures long-term fiscal stability with planned royalty rates (initially 2 percent, gradually raised to 6 percent) that cannot be changed without agreement. - Changes in project structure during negotiations could negatively affect government revenues if profits are shifted abroad or to entities subject to lower tax rates. - Dutch-disease and absorptive capacity risk - Long-term non-LNG growth could be dampened if windfalls are mismanaged—e.g., spent mostly on non-productivity enhancing current spending or less efficient capital investments. - Loss of external competitiveness of non-mining sectors (Dutch Disease) could reduce long-term growth. - Policy avenues to mitigate such risks include setting up a Sovereign Wealth Fund, managing savings and fiscal rules, and increasing competitiveness in non-resource based sectors. #### D. Policy implications and priorities (implied by analysis) - Fiscal strategy and timing - Recognize and plan for the time lag between scaling up LNG production and material fiscal revenue inflows. - Prepare medium-term fiscal strategies that account for large upfront investment-driven deficits and delayed revenue realization. - Revenue management and fiscal regime safeguards - Preserve transparency and resist revenue-shifting that could erode government tax base. - Consider mechanisms to manage fiscal windfalls (e.g., Sovereign Wealth Fund, fiscal rules) to avoid procyclical spending and preserve intergenerational equity. - Macroeconomic management to limit Dutch Disease - Implement measures to maintain external competitiveness of non-resource sectors. - Use public investments strategically to foster productivity-enhancing infrastructure and crowd in private investment. - Contingency for price and project risks - Maintain cautious planning given sensitivity to global commodity price developments and other risk factors. ### 3. The underdeveloped domestic capital market prevents a smooth monetary transmission #### Monetary transmission and interest rate pass-through - The underdeveloped domestic capital market prevents a smooth monetary transmission mechanism, making fine tuning of policy stance difficult. - During periods of large swings in reserve money growth, the impact of monetary developments on inflation outcome is more discernible (example: after the significant monetary tightening in 2011). - For most periods monetary fine tuning does not seem to exert visible impact on inflation, possibly because: - The banking system tends to be structurally over liquid. - Only a small share of the CPI basket is influenced directly by monetary policy (the CPI basket is dominated by food and administered products). - The influence of the policy rates on the interbank market appears to be weak: - Wide corridor between the two policy rates, FPC rate and the FPD rate, provides little guidance to movements of the interbank rate. - The spread between FPC and FPD rates is currently at 575 basis points after the BM narrowed the width of this corridor in recent years, but it is still much wider than in most emerging markets (typically at about 200 basis points) or advanced economies (typically at about 100 basis points). - In practice the FPC is not used frequently by commercial banks. - The interbank rate shows very low day-to-day volatility; anecdotal evidence suggests the interbank market is dominated by one/two banks that are structurally overliquid in meticals and can act as de facto price setters and market makers. - Market deposit rate is sticky; lending rate responds more to the deposit rate than to policy rates: - Between mid-2011 and late 2015, the BM cut the FPC rate by 900 basis points cumulatively, while the one-year deposit rate and lending rate only moved by 450 basis points and 500 basis points, respectively. - Deposit base dominated by a few large institutional investors who can shop around for the best rates, resulting in downward stickiness of the deposit rate. - Commercial banks tend to set lending rates as a markup over their deposit rates, which better reflect their true funding cost than the FPC rate. - Consequences of sticky and high market interest rates: - Hampers proper transmission of monetary policy. - Constrains implementation of BM’s strategy of financial deepening: high and sticky interest rates limit SMEs’ ability to afford bank credit; rapid credit growth tends to concentrate on large borrowers or consumer loans backed by salaries. #### Characteristics of Mozambique’s banking industry - Financial access and banking network expansion: - Commercial bank branches increased from 228 in 2005 to 563 in 2014. - Number of banks increased from 12 to 18 over the same period. - Financial access in Mozambique represented about 50 percent of the access points to formal financial services in 2014 (according to the 2014 FinScope Consumer Survey). - Market concentration: - Top 3 largest commercial banks accounted for 83 percent of total banks’ credit in 2011. - Credit concentration in 2011 was high compared to Mauritius and Botswana, and similar to levels in 1998 for Mozambique. - Funding structure and deposit concentration: - Domestic customers’ deposits constitute the main funding source for commercial banks. - Incentives to use central bank’s lending facility, interbank market, or foreign credit lines are limited because: - Maturity mismatch between central bank lending facility (overnight) and commercial bank loans (longer maturities). - Big and small banks exchange only a fairly low level of liquidity in the interbank market; banks’ risk aversion limits exposure to each other. - Foreign credit lines are limited in size and subject to prudential regulation restrictions that discourage lending to non-exporters and in foreign currency. - Deposit base is narrow and concentrated among a few large institutional clients (public companies and NBFI comprising pension funds and insurance companies): - These clients’ share in total banking system’s deposits grew substantially since 2005, reaching 20 percent in mid-2010, before stabilizing at 15 percent at end-2014. - There are only 14 public companies, 2 large pension funds and not more than 10 large non-bank financial institutions. - Deposit-to-GDP ratio at 33 percent at end-2011 in Mozambique, compared to Mauritius (90 percent) and South Africa (59 percent). - Lending interest rates: - Average lending interest rate for one-year maturity was close to 19 percent in late 2015 despite pronounced cuts in the central bank’s policy rate. #### Assessing the effect of credit and deposit concentration on the lending interest rate - Empirical approach - Staff assessed the impact of credit and deposit concentration on the average lending interest rate over 2002M1-2014M12 through a regression analysis. - Estimation: co-integration and instrumental variable regression; credit concentration instrumented by degree of dollarization in credit and deposit. - Controls: central bank’s lending facility interest rate, reserve money, inflation and reserve requirement. - Lending interest rate measure: banking system average prime interest rate on loans for one-year maturity. - Deposit concentration measure: ratio of public companies’ and NBFI’s deposits to total banking system deposits, excluding foreign currency and demand deposits. - Credit concentration measure: credit of the top 5 banks as a share of total banking system credit. - Key empirical findings - Deposit concentration effect: - Deposit concentration has a significant effect on the lending interest rate: all else being equal, each 10 percentage point increase in public companies’ and NBFI’s deposit concentration ratio is associated with 240 basis points increase in the lending interest rate on average. - Credit concentration effect: - Credit concentration significantly affects the lending interest rate, but the impact depends on the monetary policy stance. - The higher the central bank’s lending interest rate, the lower the effect of credit concentration and vice-versa. - Assessed at the median policy rate over the sampled period (14.5 percent), the coefficient of credit concentration is estimated at 0.19: other things being equal, each 10 percentage point increase in the credit concentration ratio is associated with 190 basis points increase in the lending interest rate. - Model-implied scenario: All else equal, the interest rate could be around 600 basis points lower if Mozambique’s credit concentration level was similar to South Africa’s. - Credit concentration weakens monetary policy effectiveness: - Central bank policy interest rate has a statistically significant and positive effect on the market lending interest rate, but pass-through is constrained by the degree of credit concentration—the higher the credit concentration, the lower the effect of central bank’s lending rate. #### Regression results — key statistics (Table 1) - Dependent variable: average prime interest rate on bank loans for 1-year maturity (percent). - Coefficients and standard errors: - Concentration of public companies' and NBFI deposits (percent): 0.24** (Std. error 0.09) - Concentration of top 5 banks' credit (percent): 1.20*** (Std. error 0.01) - BoM's policy interest rate-FPC (percent): 6.21*** (Std. error 2.01) - Interaction of FPC with credit concentration: -0.07*** (Std. error 0.02) - Reserve money y/y growth (percent): -0.11*** (Std. error 0.04) - 12-month end-of-period inflation rate (percent): -0.06 (Std. error 0.08) - Reserve requirement rate (percent): 0.95*** (Std. error 0.26) - Constant: -99.22 (Std. error 38.35) - Model diagnostics: - Adjusted R2: 0.81 - F test: 102.09 - Observations: 156.00 - Robustness: Newey-West robust standard errors - Significance notation: ***, ** and * denote statistical significance at 1%, 5% and 10%, respectively. #### Conclusions and policy implications - Main conclusions: - High deposit and credit concentration could contribute to the high lending interest rate in Mozambique. - Higher credit and deposit concentration is associated with higher lending interest rates. - Higher credit concentration appears to weaken the effectiveness of the central bank’s policy rate adjustment. - Findings are consistent with Structure-Conduct-Performance theory: highly concentrated banking industries tend to enjoy lower competition, leading to higher interest rates and reducing banks’ incentives to respond to monetary policy stimulus. - Policy implications (three main recommendations) I. Promote an environment for greater competition in banking. - Strengthen the regulatory framework to limit monopoly powers. - Implement financial literacy programs aimed at empowering the public to assess and compare available financial products across banks. II. Promote private savings and competition in the deposit base. - Conduct campaigns to bring into the banking system funds currently lying in the huge (partly unbanked) informal sector to widen the deposit base and reduce banks’ dependence on a limited number of large depositors with oligopolistic market power. III. Deepen the domestic capital markets and help SMEs obtain financial services. - Recognize that given shallow domestic markets and intrinsic SME risks in a low income country, simply promoting more credit growth is unlikely to greatly increase SME access to credit. - Use financial literacy programs to assist SMEs to produce better loan application packages that better demonstrate the viability of their businesses. ### 7. Income inequality in Mozambique and fuel import/subsidy reform #### Spatial dimension of inequality - Most of the country’s wealth is located in the southern area, and especially around the capital Maputo. - Higher rates of poverty are highly concentrated in the central and northern regions, particularly in rural areas (Alfani et al. 2012). - Fiscal transfers to rural areas could help address spatial inequality, but experience with fiscal transfers to date, especially to municipalities, has shown mixed results. #### Three Key Policy priorities — fiscal policy mechanisms - Fiscal policy can reduce income inequality through three mechanisms: tax policy, public investment, and social policies. Tax policy - Direct taxes, especially personal income taxes, are often preferable for redistributive purposes than consumption taxes. - In Mozambique, the share of direct taxes has increased over time, and income taxes now account for about 40 percent of tax collections (excluding capital gain taxes). - Property tax revenue is progressive and considered less disruptive for economic growth, but its current share in Mozambique is negligible. - The efficiency of direct taxes is low: - The corporate income tax efficiency is low due to the system of fiscal incentives (2009 Fiscal Benefits Code), which tends to favor large, capital intensive projects. - Higher income brackets enjoy larger tax credits. - There is a special regime for small taxpayers but its equity could be improved by increasing the exemption threshold. - Reducing consumption tax rates, with broadening of the base and improvements in efficiency to avoid a loss of revenues, can help redistribute income. - The Mozambican VAT is crippled with an extensive list of exempt and zero-rated items which need to be reviewed and rationalized. - VAT efficiency in Mozambique is well below the average for Southern African Development Community (SADC). - Fiscal incentives are costly and reduce fiscal space for other social spending: - Tax expenditure decreased from 4 percent of GDP in 2013 to 3.3 percent in 2014. - In 2014 tax expenditures still represented 3.3 percent of GDP, with an increasingly high concentration of VAT import exemptions for large investment projects. - A cited study suggests large mining projects account for up to 12 percent of GDP but contribute less than 3 percent of tax revenues and represent 3 percent of employment. #### PFM and Public Investment Management - Public spending has increased rapidly, but is of limited efficiency and redistributive capacity. - Total expenditures and net lending reached over 42 percent of GDP in 2014. - Public spending per capita is $283 (one of the lowest in Sub-Saharan Africa), while the Mozambican economy is just 5 percent of South Africa. - Drivers of growing public spending include the wage bill, goods and services, and domestically financed capital expenditures, with insufficient controls to ensure value-for-money. - Reforms to increase efficiency and redistributive potential: - Align public spending with a robust medium-term fiscal framework to ensure fiscal sustainability; adopt a fiscal rule embedded in fiscal responsibility legislation. - Improve public investment management by: - Approving a new legal and institutional framework for public investment management. - Introducing mandatory evaluations by a centralized evaluation committee using feasibility studies and well-defined rules. - Developing a comprehensive project database. - Avoid negative stop-and-go spending patterns that hit the poor through across-the-board cuts. - Eliminate expensive fuel subsidies taking advantage of lower oil prices: - In 2014 fuel subsidies were equivalent to 1.4 percent of GDP and less than 2 percent of the subsidy accrued to the bottom quintile of the population. #### Priority Spending - The current definition of priority spending is too broad: over 70 percent of the government budget is classified as priority spending under the current budget classification. - The administrative classification means entire ministry budgets are classified as “priority” irrespective of specific programs (e.g., purchase of new cars for the Ministry of Health classified equally with procurement of vaccines). - Recommendation: identify specific programs as priority rather than classifying whole ministry budgets. #### Conclusion: fiscal policy role and key reforms - Mozambique has experienced strong and sustained economic growth over the last two decades, but growth has not been sufficiently inclusive. - Income inequality has increased over the last decade despite high rates of economic growth; the elasticity of poverty to growth has been relatively low. - Geographical inequality (much higher income levels in the Southern provinces) could become a source of political tension given that most natural resource wealth is located in the Northern provinces. - Key fiscal reforms to help reduce income inequality: - Expansion of the tax base, reduction of exemptions (which have accrued to large corporations), and increasing reliance on direct taxes. - Greater focus on efficiency and appropriate sectoral and geographical distribution of public investment. - Reduction in the scope of priority spending to focus on the most critical social sectors and programs. #### Fuel import and subsidy reform — introduction and current challenges - Objectives of reform: (i) budgetary savings to generate fiscal space for social programs, (ii) greater transparency and efficiency, and (iii) reduction in balance of payment pressures. - Fuel subsidy system observations: - International oil prices declined from $108 to around $50 per barrel from June 2014 to mid-October 2015, but retail prices in Mozambique did not change, allowing government to offset part of the fiscal cost associated with past subsidies. - In May 2015 the government had to securitize about $100 million (0.7 percent of GDP) of debt due to fuel distributors to pay off part of the subsidies accrued in 2014, due in large part to inefficiencies in the import system. - The cost of importing fuel is higher than in most other countries in the region due to inefficiencies, generating pressures on the budget and international reserves. - Since 2009 the central bank agreed to provide up to 100 percent of the foreign exchange needed for fuel imports; fuel imports account for the majority of foreign exchange sales by the central bank on average. - Structure of fuel imports: - Fuel imports are centralized through a consortium (Imopetro). Membership in Imopetro is compulsory and no operator is authorized to import fuel outside this system. - Petromoc holds a 51 percent stake and has de facto control of Imopetro. - Contract signature is supposed to follow an international tender supervised by an ad-hoc inter-ministerial commission (CACL), but there were complaints in 2014 about possible irregularities. - Main problems in the current system: - Transparency: previous contract extensions (2013 and 2014) did not follow principle of lowest price; formula for conversion from barrels to metric tons unclear; temperature-related changes affected quantities received. - Fuel import system inefficiencies: - Weak supervision and control of key parameters affecting fuel prices (e.g., verification of shipment date in bill of lading, delivery delays up to three months, weak control of import quantities). - Imopetro has limited capacity to monitor contract execution and impose penalties; CALC has no dedicated technical staff or sanctioning ability. - Forcing distributors to mobilize financing through Imopetro and a bank syndicate increases costs and links fuel imports to international reserves; central bank provided until November 2015 a more favorable rate than the interbank market. - Fuel pricing: - The pricing formula with monthly adjustments has not been implemented since July 2011 and is used only to calculate the size of fuel subsidies. - Problems include lack of clarity on subsidy calculation, use of CIF prices inflated by import inefficiencies with no detailed breakdown, higher direct import costs than efficient levels, a poorly understood price correction factor, and distribution and retail margins not regularly updated. - Cost-effectiveness: - In 2014 fuel subsidies reached 1.1 percent of GDP on an accrual basis. - An estimated 50 percent of this subsidy ($73 million dollars) was due to inefficiencies in the import system that created large gaps between formula CIF price and international FOB price without benefit for the population. - Part of the subsidy compensates Petromoc for quasi-fiscal losses; much of the remainder was captured disproportionately by the top income quintile in urban areas, which received about 48 percent of the subsidy. #### Reforms: rationale, institutional changes and expected benefits - Rationale and objectives - Fuel subsidies paid by the State could have been lower (especially in 2014) through a more efficient import system. This would have reduced the amount of fuel imports and the volume of FX sales by the central bank. - Petromoc operates an extensive network of fuel stations, including in remote areas where transportation costs are higher, and where other private sector distributors would have no incentive to operate. As a result, Petromoc incurs operational losses (due to the social objective of ensuring fuel availability throughout the country) that should be compensated to ensure that the company is managed with a commercial orientation. - Best international practice is to record transparently these expected quasi-fiscal losses/activities in budget documents. - Specific reform steps (high level) - Allow fuel product distributors and large natural resource companies (megaprojects) to import fuel and mobilize financing directly, in line with their market needs. - Key points supporting decentralization: - The current centralization implies that the company with the weakest balance sheet (Petromoc) in practice controls the process through its majority stake in the monopoly importer (Imopetro), generating reluctance by banks to provide dollar liquidity to Imopetro. - In a decentralized system, each company could mobilize its own foreign exchange independently. Petromoc’s balance sheet would also be strengthened if the Treasury provided timely compensation for its quasi-fiscal activities. - Until November 2015, the Central Bank sold foreign exchange at a more favorable rate than the effective interbank rate to help reduce the import bill and offset some system inefficiencies. In a liberalized system, main fuel importers and fuel distributors could try to mobilize financing to pay for fuel shipments directly out of their export proceeds, through offshore loans (subject to central bank authorization), or financing from parent companies. - The domestic banking system had about $1.8 billion in October 2015 in dollar customer deposits that could finance fuel imports, conditional on banks buying dollar deposits from customers at a sufficiently attractive foreign exchange rate and on the Central Bank no longer selling dollars at a discount. - Companies would have greater incentives to supervise shipments and audit quantities received. - Imopetro should then be dismantled or transformed into an organization where participation by the fuel distributors is voluntary. - Institutional changes required to support the new system: - Use a reference price: the CIF price in the formula should be based on a benchmark international reference price increased by a standardized margin reflecting an efficient importer, rather than on actual import costs. - Regular application of the price-setting formula: avoid long retail price freezes which make the formula unmanageable and leave the system vulnerable to manipulation and mismanagement. Consider establishing an independent institution responsible for data collection, implementation of the automatic pricing mechanism, verification of the tender process and execution of fuel imports contracts. - Reinforce government regulation and supervision to avoid market collusion and ensure quality, safety and regular access to fuel products in all areas of the country. - Expected benefits and distributional considerations - Eliminating fuel subsidies via reactivation of the fuel price-setting mechanism would: - Permanently eliminate the need for a fuel subsidy, which could generate savings of around $65 million per year, on average, if we consider the annual average subsidy disbursed over the last five years. - Help activate a market adjustment mechanism as rising prices would help reduce import volumes. - Help reduce "leakage" or "smuggling" of fuel imports to neighboring countries. - Distributional impact and compensatory measures: - A fuel price increase of 20 percent is estimated to decrease income by 20 percent in the two lowest quintiles of the income distribution. - The government could study targeted subsidies to the public transportation system and/or expansion of the existing social safety net programs. *Prepared by IMF staff — Keiichiro Inui, Leandro Medina, and Christian Henn; based on FARI model estimates and project information.* ### 1. Macroeconomic and Fiscal Implications of the LNG Projects _____________________________ 7 ### 1. Macroeconomic and Fiscal Implications of the LNG Projects ### A. Background and context - Project location and scale - Rovuma offshore gas fields in Northern Mozambique (near the border with Tanzania) are divided into Area 1 (lead concessionaire: Anadarko) and Area 4 (lead concessionaire: ENI). - Total gas reserves are estimated at about 180 trillion cubic feet. - The total investment for the projects is projected to exceed $100 billion. - Once gas production reaches its peak, Mozambique could become the third largest liquefied natural gas (LNG) exporter in the world after Qatar and Australia. - Project timing and structure - Final Investment Decisions (FID) for the first round of liquefaction and processing facilities (“trains”) are expected to be taken by mid-2016. - Area 1 proposed an integrated upstream–midstream structure; Area 4 proposed segmented upstream and midstream activities. - Proposed project structures include creating Special Purpose Vehicles in other countries for marketing and financing. - Outstanding issues before FID include revision of existing exploration contracts and resettlement of domestic residents. - Expected production profile - Staff assumes production and exports of LNG will start in 2021, with production volume gradually scaled up during the 2020s. - Area 1 initially plans two onshore trains, each producing about 5.5 million tons of LNG per annum. - Area 4 will primarily build a floating liquefaction facility (FLNG) producing about 3 million tons of LNG per annum. - Total expected build-out: 13 onshore trains and 4 floating trains. - Total LNG production could reach 89 million ton per annum by 2028. - Pricing and cost considerations - Reserves are located under the deep ocean floor, which could increase drilling and infrastructure costs relative to onshore fields. - Developers may benefit from economies of scale (fewer wells). - Construction costs of liquefaction facilities could decrease if other projects stop because of low commodity prices. - Operators assume hydrocarbon prices will increase somewhat over the medium term, but abundant ongoing LNG projects worldwide (including Australia and the US) may put downward pressure on prices. - In 2015, East Asian gas prices (primary targeted market) have already dropped substantially. ### B. Macroeconomic and fiscal implications (FARI model results and key projections) - Methodology - Assessment updated using the Fiscal Analysis of Resource Industries (FARI) model developed by the IMF’s Fiscal Affairs Department. - Note on price assumption: the assumption of gas prices is based on oil price projections in the latest World Economic Outlook, and a constant coefficient of 0.14 is applied to calculate gas prices from the oil prices. - Growth - The average real GDP growth rate between 2021 and 2025 could reach 24 percent. - Share of the LNG projects in total nominal output could reach more than 50 percent by the mid-2020s. - After peak LNG production in 2028 (with the final train starting operation), real GDP growth will moderate to 3-4 percent. - This moderation reflects underlying growth of 6-6½ per-cent of the non-LNG economy and no further growth of the LNG sector. - The LNG share in total output would gradually decline starting in the late 2020s. - Fiscal revenues - Total fiscal revenues from the LNG project throughout the entire project period until 2045 could reach about $500 billion. - Main revenue sources: (i) government’s share of profit gas, (ii) corporate income tax on concessionaires, and (iii) dividends paid by the state-owned hydrocarbon company (ENH), which owns a 15 percent share of Area 1 and a 10 percent share of Area 4. - Fiscal revenues during the first few years after production starts will be limited because of large cost recovery for continuous investments in building liquefaction plants. - By the late 2020s, fiscal revenues from the gas projects could account for more than half of total fiscal revenues. - Policy implication highlighted: authorities should be aware of the time lag between gas production and fiscal revenue flows when planning medium-term fiscal strategy. - Balance of payments - Mozambique will experience unprecedented large current account deficits during the late-2010s, peaking at more than 90 percent of GDP in 2020, due to massive investments in LNG trains. - Once gas production begins, the current account could gradually improve and reach surplus in 2025. - During the early 2020s, current account will remain in deficit because of investments in subsequent trains. - By the mid-2020s, LNG will account for about ¾ of total exports. - Staff assume that increased public investments and reforms will improve external competitiveness of the non-mining industries and crowd in private investment, increasing exports of other products from late-2030s onward. - Under the assumption that risks of a Dutch-disease induced exchange rate appreciation can be contained, the overall current account would remain in surplus until the end of LNG production. ### C. Risks and sensitivity - Price risk - Sales prices of gas could drift below baseline assumptions if global gas prices decline further (e.g., due to an enduring global slowdown or if other global LNG projects meet demand). - Fiscal revenues are particularly sensitive to lower gas prices because most revenues are related to project profits. - Fiscal regime and revenue risk - Changes in taxes and fiscal regimes could significantly affect fiscal gains. - The 2014 Decree Law ensures long-term fiscal stability with planned royalty rates (initially 2 percent, gradually raised to 6 percent) that cannot be changed without agreement. - Changes in project structure during negotiations could negatively affect government revenues if profits are shifted abroad or to entities subject to lower tax rates. - Dutch-disease and absorptive capacity risk - Long-term non-LNG growth could be dampened if windfalls are mismanaged—e.g., spent mostly on non-productivity enhancing current spending or less efficient capital investments. - Loss of external competitiveness of non-mining sectors (Dutch Disease) could reduce long-term growth. - Policy avenues to mitigate such risks include setting up a Sovereign Wealth Fund, managing savings and fiscal rules, and increasing competitiveness in non-resource based sectors. ### D. Policy implications and priorities (implied by analysis) - Fiscal strategy and timing - Recognize and plan for the time lag between scaling up LNG production and material fiscal revenue inflows. - Prepare medium-term fiscal strategies that account for large upfront investment-driven deficits and delayed revenue realization. - Revenue management and fiscal regime safeguards - Preserve transparency and resist revenue-shifting that could erode government tax base. - Consider mechanisms to manage fiscal windfalls (e.g., Sovereign Wealth Fund, fiscal rules) to avoid procyclical spending and preserve intergenerational equity. - Macroeconomic management to limit Dutch Disease - Implement measures to maintain external competitiveness of non-resource sectors. - Use public investments strategically to foster productivity-enhancing infrastructure and crowd in private investment. - Contingency for price and project risks - Maintain cautious planning given sensitivity to global commodity price developments and other risk factors. *Prepared by IMF staff — Keiichiro Inui, Leandro Medina, and Christian Henn; based on FARI model estimates and project information.* ### 3. The underdeveloped domestic capital market prevents a smooth monetary ### _cr1610 - 3. The underdeveloped domestic capital market prevents a smooth monetary ### Monetary transmission and interest rate pass-through - The underdeveloped domestic capital market prevents a smooth monetary transmission mechanism, making fine tuning of policy stance difficult. - During periods of large swings in reserve money growth, the impact of monetary developments on inflation outcome is more discernible (example: after the significant monetary tightening in 2011). - For most periods monetary fine tuning does not seem to exert visible impact on inflation, possibly because: - The banking system tends to be structurally over liquid. - Only a small share of the CPI basket is influenced directly by monetary policy (the CPI basket is dominated by food and administered products). - The influence of the policy rates on the interbank market appears to be weak: - Wide corridor between the two policy rates, FPC rate and the FPD rate, provides little guidance to movements of the interbank rate. - The spread between FPC and FPD rates is currently at 575 basis points after the BM narrowed the width of this corridor in recent years, but it is still much wider than in most emerging markets (typically at about 200 basis points) or advanced economies (typically at about 100 basis points). - In practice the FPC is not used frequently by commercial banks. - The interbank rate shows very low day-to-day volatility; anecdotal evidence suggests the interbank market is dominated by one/two banks that are structurally overliquid in meticals and can act as de facto price setters and market makers. - Market deposit rate is sticky; lending rate responds more to the deposit rate than to policy rates: - Between mid-2011 and late 2015, the BM cut the FPC rate by 900 basis points cumulatively, while the one-year deposit rate and lending rate only moved by 450 basis points and 500 basis points, respectively. - Deposit base dominated by a few large institutional investors who can shop around for the best rates, resulting in downward stickiness of the deposit rate. - Commercial banks tend to set lending rates as a markup over their deposit rates, which better reflect their true funding cost than the FPC rate. - Consequences of sticky and high market interest rates: - Hampers proper transmission of monetary policy. - Constrains implementation of BM’s strategy of financial deepening: high and sticky interest rates limit SMEs’ ability to afford bank credit; rapid credit growth tends to concentrate on large borrowers or consumer loans backed by salaries. ### Characteristics of Mozambique’s banking industry - Financial access and banking network expansion: - Commercial bank branches increased from 228 in 2005 to 563 in 2014. - Number of banks increased from 12 to 18 over the same period. - Financial access in Mozambique represented about 50 percent of the access points to formal financial services in 2014 (according to the 2014 FinScope Consumer Survey). - Market concentration: - Top 3 largest commercial banks accounted for 83 percent of total banks’ credit in 2011. - Credit concentration in 2011 was high compared to Mauritius and Botswana, and similar to levels in 1998 for Mozambique. - Funding structure and deposit concentration: - Domestic customers’ deposits constitute the main funding source for commercial banks. - Incentives to use central bank’s lending facility, interbank market, or foreign credit lines are limited because: - Maturity mismatch between central bank lending facility (overnight) and commercial bank loans (longer maturities). - Big and small banks exchange only a fairly low level of liquidity in the interbank market; banks’ risk aversion limits exposure to each other. - Foreign credit lines are limited in size and subject to prudential regulation restrictions that discourage lending to non-exporters and in foreign currency. - Deposit base is narrow and concentrated among a few large institutional clients (public companies and NBFI comprising pension funds and insurance companies): - These clients’ share in total banking system’s deposits grew substantially since 2005, reaching 20 percent in mid-2010, before stabilizing at 15 percent at end-2014. - There are only 14 public companies, 2 large pension funds and not more than 10 large non-bank financial institutions. - Deposit-to-GDP ratio at 33 percent at end-2011 in Mozambique, compared to Mauritius (90 percent) and South Africa (59 percent). - Lending interest rates: - Average lending interest rate for one-year maturity was close to 19 percent in late 2015 despite pronounced cuts in the central bank’s policy rate. ### Assessing the effect of credit and deposit concentration on the lending interest rate - Empirical approach: - Staff assessed the impact of credit and deposit concentration on the average lending interest rate over 2002M1-2014M12 through a regression analysis. - Estimation: co-integration and instrumental variable regression; credit concentration instrumented by degree of dollarization in credit and deposit. - Controls: central bank’s lending facility interest rate, reserve money, inflation and reserve requirement. - Lending interest rate measure: banking system average prime interest rate on loans for one-year maturity. - Deposit concentration measure: ratio of public companies’ and NBFI’s deposits to total banking system deposits, excluding foreign currency and demand deposits. - Credit concentration measure: credit of the top 5 banks as a share of total banking system credit. - Key empirical findings: - Deposit concentration effect: - Deposit concentration has a significant effect on the lending interest rate: all else being equal, each 10 percentage point increase in public companies’ and NBFI’s deposit concentration ratio is associated with 240 basis points increase in the lending interest rate on average. - Credit concentration effect: - Credit concentration significantly affects the lending interest rate, but the impact depends on the monetary policy stance. - The higher the central bank’s lending interest rate, the lower the effect of credit concentration and vice-versa. - Assessed at the median policy rate over the sampled period (14.5 percent), the coefficient of credit concentration is estimated at 0.19: other things being equal, each 10 percentage point increase in the credit concentration ratio is associated with 190 basis points increase in the lending interest rate. - Model-implied scenario: All else equal, the interest rate could be around 600 basis points lower if Mozambique’s credit concentration level was similar to South Africa’s. - Credit concentration weakens monetary policy effectiveness: - Central bank policy interest rate has a statistically significant and positive effect on the market lending interest rate, but pass-through is constrained by the degree of credit concentration—the higher the credit concentration, the lower the effect of central bank’s lending rate. ### Regression results — key statistics (Table 1) - Dependent variable: average prime interest rate on bank loans for 1-year maturity (percent). - Coefficients and standard errors: - Concentration of public companies' and NBFI deposits (percent): 0.24** (Std. error 0.09) - Concentration of top 5 banks' credit (percent): 1.20*** (Std. error 0.01) - BoM's policy interest rate-FPC (percent): 6.21*** (Std. error 2.01) - Interaction of FPC with credit concentration: -0.07*** (Std. error 0.02) - Reserve money y/y growth (percent): -0.11*** (Std. error 0.04) - 12-month end-of-period inflation rate (percent): -0.06 (Std. error 0.08) - Reserve requirement rate (percent): 0.95*** (Std. error 0.26) - Constant: -99.22 (Std. error 38.35) - Model diagnostics: - Adjusted R2: 0.81 - F test: 102.09 - Observations: 156.00 - Robustness: Newey-West robust standard errors - Significance notation: ***, ** and * denote statistical significance at 1%, 5% and 10%, respectively. ### Conclusions and policy implications - Main conclusions: - High deposit and credit concentration could contribute to the high lending interest rate in Mozambique. - Higher credit and deposit concentration is associated with higher lending interest rates. - Higher credit concentration appears to weaken the effectiveness of the central bank’s policy rate adjustment. - Findings are consistent with Structure-Conduct-Performance theory: highly concentrated banking industries tend to enjoy lower competition, leading to higher interest rates and reducing banks’ incentives to respond to monetary policy stimulus. - Policy implications (three main recommendations): I. Promote an environment for greater competition in banking. - Strengthen the regulatory framework to limit monopoly powers. - Implement financial literacy programs aimed at empowering the public to assess and compare available financial products across banks. II. Promote private savings and competition in the deposit base. - Conduct campaigns to bring into the banking system funds currently lying in the huge (partly unbanked) informal sector to widen the deposit base and reduce banks’ dependence on a limited number of large depositors with oligopolistic market power. III. Deepen the domestic capital markets and help SMEs obtain financial services. - Recognize that given shallow domestic markets and intrinsic SME risks in a low income country, simply promoting more credit growth is unlikely to greatly increase SME access to credit. - Use financial literacy programs to assist SMEs to produce better loan application packages that better demonstrate the viability of their businesses. *Source: _cr1610 - 3. The underdeveloped domestic capital market prevents a smooth monetary (IMF document).* ### 7. Income inequality in Mozambique also ### 7. Income inequality in Mozambique also ### Spatial dimension of inequality - Most of the country’s wealth is located in the southern area, and especially around the capital Maputo. - Higher rates of poverty are highly concentrated in the central and northern regions, particularly in rural areas (Alfani et al. 2012). - Fiscal transfers to rural areas could help address spatial inequality, but experience with fiscal transfers to date, especially to municipalities, has shown mixed results. ### Three Key Policy priorities — fiscal policy mechanisms - Fiscal policy can reduce income inequality through three mechanisms: tax policy, public investment, and social policies. Tax policy - Direct taxes, especially personal income taxes, are often preferable for redistributive purposes than consumption taxes. - In Mozambique, the share of direct taxes has increased over time, and income taxes now account for about 40 percent of tax collections (excluding capital gain taxes). - Property tax revenue is progressive and considered less disruptive for economic growth, but its current share in Mozambique is negligible. - The efficiency of direct taxes is low: - The corporate income tax efficiency is low due to the system of fiscal incentives (2009 Fiscal Benefits Code), which tends to favor large, capital intensive projects. - Higher income brackets enjoy larger tax credits. - There is a special regime for small taxpayers but its equity could be improved by increasing the exemption threshold. - Reducing consumption tax rates, with broadening of the base and improvements in efficiency to avoid a loss of revenues, can help redistribute income. - The Mozambican VAT is crippled with an extensive list of exempt and zero-rated items which need to be reviewed and rationalized. - VAT efficiency in Mozambique is well below the average for Southern African Development Community (SADC). - Fiscal incentives are costly and reduce fiscal space for other social spending: - Tax expenditure decreased from 4 percent of GDP in 2013 to 3.3 percent in 2014. - In 2014 tax expenditures still represented 3.3 percent of GDP, with an increasingly high concentration of VAT import exemptions for large investment projects. - A cited study suggests large mining projects account for up to 12 percent of GDP but contribute less than 3 percent of tax revenues and represent 3 percent of employment. ### PFM and Public Investment Management - Public spending has increased rapidly, but is of limited efficiency and redistributive capacity. - Total expenditures and net lending reached over 42 percent of GDP in 2014. - Public spending per capita is $283 (one of the lowest in Sub-Saharan Africa), while the Mozambican economy is just 5 percent of South Africa. - Drivers of growing public spending include the wage bill, goods and services, and domestically financed capital expenditures, with insufficient controls to ensure value-for-money. - Reforms to increase efficiency and redistributive potential: - Align public spending with a robust medium-term fiscal framework to ensure fiscal sustainability; adopt a fiscal rule embedded in fiscal responsibility legislation. - Improve public investment management by: - Approving a new legal and institutional framework for public investment management. - Introducing mandatory evaluations by a centralized evaluation committee using feasibility studies and well-defined rules. - Developing a comprehensive project database. - Avoid negative stop-and-go spending patterns that hit the poor through across-the-board cuts. - Eliminate expensive fuel subsidies taking advantage of lower oil prices: - In 2014 fuel subsidies were equivalent to 1.4 percent of GDP and less than 2 percent of the subsidy accrued to the bottom quintile of the population. ### Priority Spending - The current definition of priority spending is too broad: over 70 percent of the government budget is classified as priority spending under the current budget classification. - The administrative classification means entire ministry budgets are classified as “priority” irrespective of specific programs (e.g., purchase of new cars for the Ministry of Health classified equally with procurement of vaccines). - Recommendation: identify specific programs as priority rather than classifying whole ministry budgets. ### Conclusion: fiscal policy role and key reforms - Mozambique has experienced strong and sustained economic growth over the last two decades, but growth has not been sufficiently inclusive. - Income inequality has increased over the last decade despite high rates of economic growth; the elasticity of poverty to growth has been relatively low. - Geographical inequality (much higher income levels in the Southern provinces) could become a source of political tension given that most natural resource wealth is located in the Northern provinces. - Key fiscal reforms to help reduce income inequality: - Expansion of the tax base, reduction of exemptions (which have accrued to large corporations), and increasing reliance on direct taxes. - Greater focus on efficiency and appropriate sectoral and geographical distribution of public investment. - Reduction in the scope of priority spending to focus on the most critical social sectors and programs. ### Fuel import and subsidy reform — introduction and current challenges - Objectives of reform: (i) budgetary savings to generate fiscal space for social programs, (ii) greater transparency and efficiency, and (iii) reduction in balance of payment pressures. - Fuel subsidy system observations: - International oil prices declined from $108 to around $50 per barrel from June 2014 to mid-October 2015, but retail prices in Mozambique did not change, allowing government to offset part of the fiscal cost associated with past subsidies. - In May 2015 the government had to securitize about $100 million (0.7 percent of GDP) of debt due to fuel distributors to pay off part of the subsidies accrued in 2014, due in large part to inefficiencies in the import system. - The cost of importing fuel is higher than in most other countries in the region due to inefficiencies, generating pressures on the budget and international reserves. - Since 2009 the central bank agreed to provide up to 100 percent of the foreign exchange needed for fuel imports; fuel imports account for the majority of foreign exchange sales by the central bank on average. - Structure of fuel imports: - Fuel imports are centralized through a consortium (Imopetro). Membership in Imopetro is compulsory and no operator is authorized to import fuel outside this system. - Petromoc holds a 51 percent stake and has de facto control of Imopetro. - Contract signature is supposed to follow an international tender supervised by an ad-hoc inter-ministerial commission (CACL), but there were complaints in 2014 about possible irregularities. - Main problems in the current system: - Transparency: previous contract extensions (2013 and 2014) did not follow principle of lowest price; formula for conversion from barrels to metric tons unclear; temperature-related changes affected quantities received. - Fuel import system inefficiencies: - Weak supervision and control of key parameters affecting fuel prices (e.g., verification of shipment date in bill of lading, delivery delays up to three months, weak control of import quantities). - Imopetro has limited capacity to monitor contract execution and impose penalties; CALC has no dedicated technical staff or sanctioning ability. - Forcing distributors to mobilize financing through Imopetro and a bank syndicate increases costs and links fuel imports to international reserves; central bank provided until November 2015 a more favorable rate than the interbank market. - Fuel pricing: - The pricing formula with monthly adjustments has not been implemented since July 2011 and is used only to calculate the size of fuel subsidies. - Problems include lack of clarity on subsidy calculation, use of CIF prices inflated by import inefficiencies with no detailed breakdown, higher direct import costs than efficient levels, a poorly understood price correction factor, and distribution and retail margins not regularly updated. - Cost-effectiveness: - In 2014 fuel subsidies reached 1.1 percent of GDP on an accrual basis. - An estimated 50 percent of this subsidy ($73 million dollars) was due to inefficiencies in the import system that created large gaps between formula CIF price and international FOB price without benefit for the population. - Part of the subsidy compensates Petromoc for quasi-fiscal losses; much of the remainder was captured disproportionately by the top income quintile in urban areas, which received about 48 percent of the subsidy. *Source: _cr1610 - 7. Income inequality in Mozambique also_* ### 7. Reforms should aim at generating fiscal savings, reducing pressures on international ### 7. Reforms should aim at generating fiscal savings, reducing pressures on international reserves and achieving greater transparency and efficiency in the sector ### Rationale and objectives - Fuel subsidies paid by the State could have been lower (especially in 2014) through a more efficient import system. This would have reduced the amount of fuel imports and the volume of FX sales by the central bank. - Petromoc operates an extensive network of fuel stations, including in remote areas where transportation costs are higher, and where other private sector distributors would have no incentive to operate. As a result, Petromoc incurs operational losses (due to the social objective of ensuring fuel availability throughout the country) that should be compensated to ensure that the company is managed with a commercial orientation. - Best international practice is to record transparently these expected quasi-fiscal losses/activities in budget documents. ### Specific reform steps (high level) - Allow fuel product distributors and large natural resource companies (megaprojects) to import fuel and mobilize financing directly, in line with their market needs. Key points supporting decentralization: - The current centralization implies that the company with the weakest balance sheet (Petromoc) in practice controls the process through its majority stake in the monopoly importer (Imopetro), generating reluctance by banks to provide dollar liquidity to Imopetro. - In a decentralized system, each company could mobilize its own foreign exchange independently. Petromoc’s balance sheet would also be strengthened if the Treasury provided timely compensation for its quasi-fiscal activities. - Until November 2015, the Central Bank sold foreign exchange at a more favorable rate than the effective interbank rate to help reduce the import bill and offset some system inefficiencies. In a liberalized system, main fuel importers and fuel distributors could try to mobilize financing to pay for fuel shipments directly out of their export proceeds, through offshore loans (subject to central bank authorization), or financing from parent companies. - The domestic banking system had about $1.8 billion in October 2015 in dollar customer deposits that could finance fuel imports, conditional on banks buying dollar deposits from customers at a sufficiently attractive foreign exchange rate and on the Central Bank no longer selling dollars at a discount. - Companies would have greater incentives to supervise shipments and audit quantities received. - Imopetro should then be dismantled or transformed into an organization where participation by the fuel distributors is voluntary. - Institutional changes required to support the new system: - Use a reference price: the CIF price in the formula should be based on a benchmark international reference price increased by a standardized margin reflecting an efficient importer, rather than on actual import costs. - Regular application of the price-setting formula: avoid long retail price freezes which make the formula unmanageable and leave the system vulnerable to manipulation and mismanagement. Consider establishing an independent institution responsible for data collection, implementation of the automatic pricing mechanism, verification of the tender process and execution of fuel imports contracts. - Reinforce government regulation and supervision to avoid market collusion and ensure quality, safety and regular access to fuel products in all areas of the country. Note: "This reform would require strong government regulation and supervision to avoid market collusion and ensure quality, safety and regular access to fuel products in all areas of the country. The formula might have to be adjusted to replace actual cost of imports by a benchmark price based on the Platts and a mark-up reflecting the cost of an efficient supplier. Oil companies would have different import costs and import schedules, and transparent pricing would be key to avoid collusion in vertically integrated companies." ### Expected benefits and distributional considerations - Eliminating fuel subsidies via reactivation of the fuel price-setting mechanism would: - Permanently eliminate the need for a fuel subsidy, which could generate savings of around $65 million per year, on average, if we consider the annual average subsidy disbursed over the last five years. - Help activate a market adjustment mechanism as rising prices would help reduce import volumes. - Help reduce "leakage" or "smuggling" of fuel imports to neighboring countries. - Distributional impact and compensatory measures: - A fuel price increase of 20 percent is estimated to decrease income by 20 percent in the two lowest quintiles of the income distribution. - The government could study targeted subsidies to the public transportation system and/or expansion of the existing social safety net programs. Footnote context: "However, at the international prices prevailing in late October 2015, the current system did not generate any explicit subsidies (it actually involves a fairly modest negative subsidy). These market conditions provide a good opportunity for reform, even though the recent depreciation of the metical over the last two weeks of November may require another reassessment of the situation." *REPUBLIC OF MOZAMBIQUE INTERNATIONAL MONETARY FUND* --- _Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr1610.pdf_