Abstract
Selected Issues
Macrofinancial Linkages and Bank Vulnerabilities1
A. Introduction
1. Macrofinancial analysis evaluates vulnerabilities in the financial sector and its linkages to other sectors of the economy. Weak macroeconomic fundamentals typically affect the financial sector, and there may be repercussions from a weak financial sector onto the real economy, especially if hit by a shock. Understanding of macrofinancial linkages permits a forward-looking analysis of the entire economy, with the potential to shape adjustment policies ahead of distress. Ideally, an integrated macrofinancial analysis employs a two-way assessment of macrofinancial risk and macroeconomic stability, with both macro-to-financial and financial-to-macro feedback loops. It is also of interest how the financial sector may magnify or dampen shocks to the economy (IMF, 2017a).
2. More specifically, macrofinancial linkages and thus channels of risk transmission exist between the major economic sectors and the financial sector. Figure 1 illustrates the main transmission channels from the external, fiscal, monetary and real sector onto the financial sector and back, and what type of contagion may be expected.
Mozambique: Macrofinancial Linkages
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Source: IMF Policy Paper, Approaches to Macro-financial Surveillance in Article IV Reports, March 2017.Mozambique: Macrofinancial Linkages
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Source: IMF Policy Paper, Approaches to Macro-financial Surveillance in Article IV Reports, March 2017.Mozambique: Macrofinancial Linkages
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Source: IMF Policy Paper, Approaches to Macro-financial Surveillance in Article IV Reports, March 2017.3. In the case of Mozambique, there are strong macrofinancial linkages between the government on the one hand and state-owned enterprises (SOEs), banks, and external creditors on the other. The following investigation analyzes the vulnerabilities arising from indebtedness of the sovereign versus external and domestic entities, and how the banking system could be affected by the propagation of potential shocks through the economic system, particularly through debt of private enterprises and SOEs to banks. A stress test exercise attempts to quantify the impact of a worsening of private sector performance on loan quality and, hence, bank capitalization. In addition, a debt viability check for the SOE sector calculates the debt overhang of major SOEs that, in the absence of corporate restructuring, could lead to a restructuring of bank debt or debt service and likewise affect bank capital. The analysis suggests the possibility of feedback effects onto the real sector from the banking system reacting to deteriorating asset quality and shrinking capital buffers.
B. Macrofinancial Linkages in Mozambique
4. During 2009–2014, the Mozambican government scaled up public investment, responding to developmental needs in anticipation of natural resource revenues. The exploitation of significant gas reserves is expected to start in 2023. Public investment was carried out through the public investment budget, non-financial public corporations2 and public-private partnerships (PPPs) involving those corporations. Investment projects under the budget were increasingly financed through domestic and external borrowing, increasing central government liabilities. In addition, those projects carried out by non-financial public corporations are a source of contingent liabilities given government guarantees in some cases. Moreover, several large infrastructure projects involving SOEs are financed increasingly by bilateral loans contracted on non-concessional terms. During this period, increasing government and SOEs liabilities built significant macrofinancial risks.
5. During 2015–2016, a sequence of combined shocks led to macroeconomic deterioration. The economic growth was adversely affected by lower commodity prices and adverse weather conditions together with the disclosure of large SOEs liabilities in April 2016 and subsequent freeze of donor support. These shocks slowed growth to 3.8 percent in 2016 (compared to 6.3 percent in 2015). A sizable fiscal slippage in late 2015 together with an overly loose monetary policy resulted in substantial exchange rate depreciation despite sales of foreign exchange by the central bank in 2016. Inflation rose to double digits, peaking at 26.3 percent in October 2016.
6. Increased strain in government accounts and high borrowing costs have affected state-owned enterprises (SOEs). In some cases, government investment in SOEs (e.g. telecom firm) has dwindled, contributing to a loss in market share. There has also been an accumulation of arrears to many SOEs, and the government has delayed injecting new capital, leading to higher SOEs leverage. At the same time, SOEs have accumulated tax arrears to the government. Some SOEs have also run up arrears to other SOEs and to the private sector (in individual cases even to their own private-sector subsidiaries).
7. The government disclosed sizable external loans contracted by SOEs in 2014 and 2016,3 part of which are held by banks. These loans were contracted by three companies from international creditors and backed by government guarantees. The Mozambican Tuna Company (EMATUM) contracted a loan of $850 million for fishing tuna activities and maritime security. Proindicus borrowed $622 million for purchase of maritime surveillance vessels and equipment, while Mozambique Asset Management (MAM) contracted a loan of $535 million for the purchase of floating docks to maintain vessels. The first loan disclosed was EMATUM in 2014 while Proindicus and MAM followed in 20164. Four local banks held participatory notes on Proindicus and EMATUM loans. The exposure to the loans has been fully provisioned.5 The EMATUM loan was swapped in April 2016 into an Eurobond maturing in 2023.
8. Public debt is in distress, and restructuring discussions have not progressed. In October 2016, the government announced to creditors its inability to service the disclosed external loans and its resolve to initiate restructuring discussions. The stock of public sector debt-to-GDP reached 128.3 percent at end-2016, of which 103.7 percent of GDP is external. Several payments on external borrowing were missed.6 If talks fail, external budget financing will remain tight, restricting fiscal space.
9. Balance sheet analysis of the economy reveals strong linkages that the public sector has with banks, the private sector and the rest of the world.7 The balance sheet matrix for 2017Q3 estimates intersectoral assets and liabilities (Table 1). To build the matrix, we use sectoral balances sheets for the central bank and other depositary corporation, international investment position statistics, and partial information on the stock of government securities. The matrix, presented in terms of net asset or creditor position, shows that the public sector is a major player with important linkages to other sectors through two channels: bank exposure to government securities (with a net position of 10.7 percent of GDP) and government and SOEs access to external financing (with the rest of the world holding assets close to 95 percent of GDP). The second link is a source of liability dollarization.8
Mozambique: Balance Sheet Matrix, Net Position, 2017Q31
(Percent of GDP)
IIP does not discriminate between corporations and households that held a net asset position of 46 percent of GDP against the rest of the world.
Mozambique: Balance Sheet Matrix, Net Position, 2017Q31
(Percent of GDP)
| Private sector1 | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Central bank | Central government | State and local government | Public nonfinancial corporations | Other depository corporations | Other financial corporations | Corporations | Households | External | |
| Central bank | 5.0 | 0.2 | 0.0 | 7.9 | 0.0 | 0.0 | −0.2 | −15.9 | |
| Central government | −5.0 | … | … | 10.7 | … | … | … | 72.8 | |
| State and local government | −0.2 | … | … | 0.1 | … | … | … | … | |
| Public nonfinancial corporations | 0.0 | … | … | −0.6 | … | … | … | −5.9 | |
| Other depository corporations | −7.9 | −10.7 | −0.1 | 0.6 | 2.1 | −0.5 | 11.3 | −3.8 | |
| Other financial corporations | 0.0 | … | … | … | −2.1 | … | … | … | |
| Corporations | 0.0 | … | … | … | 0.5 | … | … | ||
| Households | 0.2 | … | … | … | −11.3 | … | … | ||
| External | 15.9 | −72.8 | 0.0 | −5.9 | 3.8 | … | 46.0 | ||
IIP does not discriminate between corporations and households that held a net asset position of 46 percent of GDP against the rest of the world.
Mozambique: Balance Sheet Matrix, Net Position, 2017Q31
(Percent of GDP)
| Private sector1 | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Central bank | Central government | State and local government | Public nonfinancial corporations | Other depository corporations | Other financial corporations | Corporations | Households | External | |
| Central bank | 5.0 | 0.2 | 0.0 | 7.9 | 0.0 | 0.0 | −0.2 | −15.9 | |
| Central government | −5.0 | … | … | 10.7 | … | … | … | 72.8 | |
| State and local government | −0.2 | … | … | 0.1 | … | … | … | … | |
| Public nonfinancial corporations | 0.0 | … | … | −0.6 | … | … | … | −5.9 | |
| Other depository corporations | −7.9 | −10.7 | −0.1 | 0.6 | 2.1 | −0.5 | 11.3 | −3.8 | |
| Other financial corporations | 0.0 | … | … | … | −2.1 | … | … | … | |
| Corporations | 0.0 | … | … | … | 0.5 | … | … | ||
| Households | 0.2 | … | … | … | −11.3 | … | … | ||
| External | 15.9 | −72.8 | 0.0 | −5.9 | 3.8 | … | 46.0 | ||
IIP does not discriminate between corporations and households that held a net asset position of 46 percent of GDP against the rest of the world.
10. Data gaps constrained the scope of the analysis. A partial balance sheet matrix was derived from IMF’s Standardized Reports Forms (SRF). The empty cells reflect lack of data. To conduct a comprehensive economy-wide balance sheet analysis, key data to be compiled are: government (GFS) balance sheet data covering general government (central government, state, and local government) and nonfinancial public corporations, as well as coverage of private non-financial corporations and households.
11. Significant domestic government arrears have accumulated. The stock of arrears reached about 3.7 percent of GDP at end 2016.9 A breakdown by sector and the data for 2017 were not available. In 2018, the government plans to undertake an audit of arears including amounts incurred in 2017 and a sectoral breakdown.
12. Mozambique is facing specific risks that may have a high impact on the economy. Relevant downside risks include: 10 (i) a prolonged uncertainty related to the hidden debt issue and lack of progress in restructuring SOE debt, (ii) increasing fiscal risks, including non-performing SOEs and government expenditure arrears; and (iii) deterioration in asset quality of banks and liquidity shortfalls due to protracted low growth and high real interest rates. If these risks materialize, macrofinancial linkages—especially involving sectors with lopsided net debtor positions—may play out adversely, propagating shocks through the balance sheet of the economy (Figure 2) and ultimately affecting the banking system. More specifically, materialization of risks (i) and/or (ii) would restrict the fiscal space further and make the restructuring and recapitalization of major SOEs as well as arrears clearance with suppliers less feasible. If the macro shock reflected by risk (iii) occurred, both the government and the private sector would suffer, leading to even higher loan defaults in the banking sector.11
Mozambique: Network Map of Sectoral Linkages, Net Position, 2017Q31
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Sources: Mozambique authorities; and IMF staff calculations.1 Arrow goes from debtor to creditor: The thickness of the arrow represents the size of the net liability.Mozambique: Network Map of Sectoral Linkages, Net Position, 2017Q31
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Sources: Mozambique authorities; and IMF staff calculations.1 Arrow goes from debtor to creditor: The thickness of the arrow represents the size of the net liability.Mozambique: Network Map of Sectoral Linkages, Net Position, 2017Q31
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Sources: Mozambique authorities; and IMF staff calculations.1 Arrow goes from debtor to creditor: The thickness of the arrow represents the size of the net liability.C. Banking Sector Developments and Assessment of Vulnerabilities
13. Despite a period of severe instability in 2016 and sluggish growth, the banking system remains steady, holding significant capital and liquidity buffers, but vulnerabilities persist. Banks remain liquid, with stable deposits, and are well capitalized and profitable on average (Figure 3), but there is heterogeneity across institutions, and a few small banks have weaker capital positions. In addition, nonperforming loans (NPLs) have doubled to 11.4 percent between end-2016 and September 2017. Several factors underlie this deterioration: (i) more focused credit risk inspections by the BM, which led to stricter loan classification, in part of exposures with state-owned enterprises (SOEs); (ii) slower economic growth; (iii) high real interest rates; and (iv) substantial government and SOE arrears with suppliers. Loan loss provisioning, however, remains comfortable at 79 percent of NPLs (58 percent in specific provisions, down from 78 percent at end-2016, 11 percent in generic provisions, and 10 percent prudential provisions12). New regulations to increase the capital adequacy ratio (CAR, from 8 to 12 percent, over a period of three years) and establish minimum liquidity requirements were put in place in 2017.
Mozambique: Banking Financial Indicators
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Sources: Bank of Mozambique; and IMF staff calculations.Mozambique: Banking Financial Indicators
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Sources: Bank of Mozambique; and IMF staff calculations.Mozambique: Banking Financial Indicators
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Sources: Bank of Mozambique; and IMF staff calculations.14. The monetary and fiscal policy mix fueled fast credit growth in 2014–16 (Figure 4). The Credit-to-GDP ratio was above its trend by more than the lower threshold of 2 percent between June 2014 and September 2016.13 The credit-to-GDP ratio peaked at 40.2 percent (from a 25.4 percent bottom in September 2012). Credit to the economy grew annually by more than 10 percent from May 2013 until December 2015 both in nominal and in real terms.
Mozambique: Credit-to-GDP Gap
(Percentage points of GDP)
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Sources: Bank of Mozambique; and IMF staff calculationsMozambique: Credit-to-GDP Gap
(Percentage points of GDP)
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Sources: Bank of Mozambique; and IMF staff calculationsMozambique: Credit-to-GDP Gap
(Percentage points of GDP)
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Sources: Bank of Mozambique; and IMF staff calculations15. In response to high inflation and exchange rate depreciation, the central bank significantly tightened the monetary policy stance in October 2016. The BM increased its lending rate by 600 basis points to 23.25 percent in October 2016 (Figure 5). This measure increased demand for domestic deposits and helped stabilize the exchange rate by rebalancing the foreign exchange market.14 In late 2016, the tightening and the resolution of two banks marked the end of the cycle of fast credit growth. Credit to the economy started to decline in real terms, and in July 2017 it began to contract nominally year-on-year. In September 2016 Moza Banco, the fourth largest bank with 6 percent of assets, was put under official administration and shortly thereafter Nosso Banco, holding less than 1 percent of assets, was liquidated after failing to comply with capital requirements.
Mozambique: Policy Rates, Inflation and Currency Depreciation
(Percent)
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Mozambique: Policy Rates, Inflation and Currency Depreciation
(Percent)
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Mozambique: Policy Rates, Inflation and Currency Depreciation
(Percent)
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
16. Since mid-2017, inflation has dropped rapidly and policy adjustment lagged, leading to high real interest rates. Broad stability has returned in the foreign exchange market and an expansion in agricultural output accompanied the monetary tightening, leading to a single-digit inflation rate (7 percent year on year in December).15 In response to rapid disinflation but in absence of fiscal consolidation, BM cautiously cut the policy rate by 225 basis points to 19.5 percent during the last three monetary policy committee meetings.16 These elements have led to very high real interest rates and a slow growth environment, causing a contraction domestic currency credit (by 12 percent y-o-y at end 2017) and growth of foreign currency credit (by 6 percent, Figure 6). Overall, the construction, manufacturing, commerce and consumer sector has experienced sharp credit contractions. Since deposits in domestic currency have continued growing, the resulting gap between credit to the private sector and monetary aggregates (Figure 6) has opened financing space for the governments’ significant funding needs to be met, albeit at a higher cost. Private banks reported that SME access to finance has been affected by the contraction in credit in an environment of high interest rates and structural constraints in the banking sector.17
Mozambique: Credit and Aggregate Developments
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Source: Bank of Mozambique.Mozambique: Credit and Aggregate Developments
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Source: Bank of Mozambique.Mozambique: Credit and Aggregate Developments
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Source: Bank of Mozambique.17. Notwithstanding the weakness in SME credit, financial inclusion indicators have improved over the years, and the authorities are executing a detailed inclusion strategy. The authorities’ financial inclusion index showed a steady improvement during 2005–2015, before falling slightly in 2016 for statistical reasons (additional districts with poor financial access now included). The government posits that financial inclusion is, however, still low in non-urban areas. One crucial step toward higher inclusion was the approval of a regime for access spelling out the rules for banks to expand activities through banking agents, expected to help bring services to underbanked rural areas. In addition, the government has elaborated a national financial inclusion strategy for the period of 2016 to 2022 (Republic of Mozambique, 2015). The strategy rests one three pillars—access and use of financial services, strengthening of financial infrastructure, and better consumer protection and financial literacy. It features specific interim and final targets,18 with a monitoring and evaluation mechanism in place.
18. Key vulnerabilities relate to sovereign-financial linkages and the macroeconomic context. Banks are heavily exposed the public sector. The exposure takes different forms: T-bills, T-bonds, direct financing of SOEs, participation in the financing of public projects, among others. As in many other jurisdictions, direct exposures to the central government carry a zero-weight risk on capital, making them attractive but reinforcing risk concentration. Lending to SOEs receives a risk treatment akin to corporate lending, unless it is backed by government guarantees, in which case it also enjoys sovereign classification. Given the high and rising level of public sector exposures (estimated at 25 percent by end September 2017) and the high real interest rates, credit to the private sector has been crowded out, especially in 2017 when a nominal contraction of 12 percent year-on-year was observed. In addition, some institutions are close to their sovereign risk concentration limits, impeding at times their participation in the weekly T-bill market auctions.19
19. Banks liquidity tends to flow to the BM due to the high real returns and low risk. Banks are placing their extra balances in domestic currency at the central bank in the form of excess reserves, or via reverse repos in the overnight market. Such liquidity allocation partly helps cover public sector financing needs. At the same time, it constrains credit normalization, which in the medium term will also be affected negatively by the higher capital requirements.
20. Dollarization remains considerable at 22 percent of assets and 28 percent of liabilities. While the FX market is generally less active and volatile than in 2014–15, dollarization ratios have declined somewhat partly due to the accounting effect of the Metical revaluation last year. The new law for the foreign exchange in force since end-2017 introduced, among other features, more flexibility in the management of exports proceeds by eliminating mandatory minimum surrender rules. The added flexibility will improve FX management of exporting companies.20 The BM will also reduce its purchase of FX surpluses in the system,21 and the related sterilization costs. However, by potentially leaving higher FX balances in the real sector, dollarization of domestic transactions may crop up as well as higher intermediation in dollars, especially given the relatively low interest rate in foreign currency and current subdued demand for imports. Macroprudential measures against excessive dollarization are in place, including very high ex-ante provisioning for dollar loans to non-exporting borrowers. However, authorities must remain vigilant and higher reserve requirements on deposits in foreign currency, and minimum dollar liquidity requirements may be warranted.
21. Macroprudential tools aimed at mitigating credit risk have not been used in Mozambique thus far. The incipient mortgage market and the near-term horizon of lending in general have made the introduction of instruments like debt-to-income or loan-to-value limits less pressing. However, the BM could start monitoring those ratios and be ready to establish good practices for the local market, especially for smaller segments of commercial borrowers. Improving macroprudential surveillance, including through upcoming reporting on financial stability,22 would benefit from such more detailed indicators.
Assessment of bank vulnerabilities
22. Considering the sovereign-bank nexus as well as, high credit costs, and dollarization, credit risk and subsequent portfolio deterioration will likely prevail in the system. To gauge possible impacts of NPL growth on provisioning and capital requirements, the mission elaborated a sensitivity analysis using detailed bank by bank data provided by BM.
23. The exercise focuses on the impact of higher provisioning from higher NPLs on solvency ratios. The analysis is based on September 2017 data and covers the 2018–19 period under three scenarios of portfolio deterioration (Table 2). The scenarios include (i) a baseline based on credit growth consistent with the macro-framework and moderate NPL growth, with a gradual increase in provision coverage levels over the next two years; (ii) a continued low credit growth setup with faster loan deterioration; and (iii) a credit recovery scenario with lower NPL growth which allows for full provision coverage. Key assumptions include: provision requirements at the closing of fiscal year 2017 to stay at the level of September 2017, credit growth is homogeneously distributed, and the structure of portfolios is not altered during the period so that risk-weighted assets grow in sync with aggregate credit. Also, while NPLs doubled during 2017, it is assumed the spike was related to a one-off correction and the rates of deterioration range from 5 to 20 percent in the different scenarios.
Mozambique: Non-Performing Loans Sensitivity Analysis
Mozambique: Non-Performing Loans Sensitivity Analysis
| Baseline scenario | NPL deterioration scenario | Credit recovery scenario | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Dec-17 | Dec-18 | Dec-19 | Dec-17 | Dec-18 | Dec-19 | Dec-17 | Dec-18 | Dec-19 | ||
| Assumptions | ||||||||||
| Credit growth | … | 3.0 | 3.0 | … | 0.0 | 10.0 | … | 10.0 | 10.0 | |
| RWA growth | … | 3.0 | 3.0 | … | 0.0 | 10.0 | … | 10.0 | 10.0 | |
| NPL growth | … | 15.0 | 15.0 | … | 20.0 | 20.0 | … | 10.0 | 5.0 | |
| Provision coverage | 59.0 | 80.0 | 100.0 | 59.0 | 80.0 | 80.0 | 59.0 | 100.0 | 100.0 | |
| Results | ||||||||||
| System NPL | 11.5 | 13.2 | 14.7 | 11.5 | 13.8 | 15.0 | 11.5 | 12.6 | 12.0 | |
| System CAR | 14.2 | 13.4 | 10.4 | 14.2 | 13.4 | 11.9 | 14.2 | 10.8 | 10.7 | |
| No. of banks below minimum | 2 | 3 | 6 | 2 | 3 | 6 | 2 | 6 | 6 | |
| Loan market share | 0.6 | 1.6 | 5.1 | 0.6 | 1.6 | 5.1 | 0.6 | 5.1 | 5.1 | |
Mozambique: Non-Performing Loans Sensitivity Analysis
| Baseline scenario | NPL deterioration scenario | Credit recovery scenario | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Dec-17 | Dec-18 | Dec-19 | Dec-17 | Dec-18 | Dec-19 | Dec-17 | Dec-18 | Dec-19 | ||
| Assumptions | ||||||||||
| Credit growth | … | 3.0 | 3.0 | … | 0.0 | 10.0 | … | 10.0 | 10.0 | |
| RWA growth | … | 3.0 | 3.0 | … | 0.0 | 10.0 | … | 10.0 | 10.0 | |
| NPL growth | … | 15.0 | 15.0 | … | 20.0 | 20.0 | … | 10.0 | 5.0 | |
| Provision coverage | 59.0 | 80.0 | 100.0 | 59.0 | 80.0 | 80.0 | 59.0 | 100.0 | 100.0 | |
| Results | ||||||||||
| System NPL | 11.5 | 13.2 | 14.7 | 11.5 | 13.8 | 15.0 | 11.5 | 12.6 | 12.0 | |
| System CAR | 14.2 | 13.4 | 10.4 | 14.2 | 13.4 | 11.9 | 14.2 | 10.8 | 10.7 | |
| No. of banks below minimum | 2 | 3 | 6 | 2 | 3 | 6 | 2 | 6 | 6 | |
| Loan market share | 0.6 | 1.6 | 5.1 | 0.6 | 1.6 | 5.1 | 0.6 | 5.1 | 5.1 | |
24. All three scenarios result in non-negligible provisioning needs and additional capital Two small banks would require a capital infusion at the outset due to losses during 2017., but in 2019 up to six banks with an aggregate market share of 5 percent in the loan market could require significant capital infusions.
Mozambique: NPL Sensitivity Scenarios
(Percent)
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Sources: Bank of Mozambique; and IMF staff calculations and projections.Mozambique: NPL Sensitivity Scenarios
(Percent)
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Sources: Bank of Mozambique; and IMF staff calculations and projections.Mozambique: NPL Sensitivity Scenarios
(Percent)
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Sources: Bank of Mozambique; and IMF staff calculations and projections.D. Excess Leverage of State-Owned Enterprises and Implications for Banks
25. In the following section, the debt sustainability of major SOEs and the resilience of banks to potential debt restructuring are assessed. Standard metrics of corporate debt sustainability analysis are employed to determine the debt overhang and excessive debt service burden of individual SOEs. The impact of the necessary adjustments following from this analysis—additional provisions on loan exposures and/or reduction of the interest bill—is then computed for each bank.
26. The financial situation of the largest SOEs has been deteriorating. As the latest available audited financial statements at end-2016 show, the ten largest non-financial SOEs that accounting for about 90 percent of sector assets, experienced rising pre-tax losses compared to 2015. This was due to rising net interest costs that almost doubled on the back of higher loan rates and rising liabilities. Indeed, the top-10 SOEs’ bank debt grew by 47 percent in 2016, lifting the ratio of bank debt to total assets from 42 to 49 percent.
27. Essentially cash-strapped, SOEs increasingly turn to bank financing to defray operational costs, with the potential of bank debt ballooning further. Half of the 18 SOEs that reportedly carry bank debt23 experienced both rising bank debt and net interest payments during 2016 (Figure 8). Many SOEs have a negative operational cash flow even after adding back non-cash expenses. In covering the cash shortfall, they need to turn to banks for loan funding or to suppliers for short-term trade credit. By safeguarding cash, some SOEs show cash flows that exceed their operational income.24 SOEs’ bank debt has increased substantially, with gross loans rising by 40 percent during 2016. Not all this debt is owed to local banks; in fact, a few SOEs are almost exclusively indebted with foreign lenders. The additional debt together with rising loan rates caused interest paid by SOEs to almost double in 2016.
Mozambique: SOEs Changes in Bank Debt and Interest Payments in 2016
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Source: SOE’s audited financial statements, 2015 and 2016.Mozambique: SOEs Changes in Bank Debt and Interest Payments in 2016
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Source: SOE’s audited financial statements, 2015 and 2016.Mozambique: SOEs Changes in Bank Debt and Interest Payments in 2016
Citation: IMF Staff Country Reports 2018, 066; 10.5089/9781484345634.002.A003
Source: SOE’s audited financial statements, 2015 and 2016.28. A standard sustainability metric is used to determine the debt overhang of SOEs. The analysis applies the most common metric for corporate debt sustainability, the net debt-to-EBITDA ratio. The ratio divides net debt, which is gross debt minus cash and cash equivalents, by earnings before interest, taxes, depreciation and amortization (EBITDA). Net debt is used because cash holdings could be used to pay down gross debt, while EBITDA serves as proxy for cash flow across industries by excluding cost drivers that depend on the structure of individual firms.25 The net debt-to-EBITDA ratio shows how many years it would take a company to pay back its debt from operations. Analysts typically use a critical threshold of 5 times EBITDA for determining the debt overhang but thresholds vary across industries, with firms in some capital-intensive sectors (e.g. utility firms) able to sustain a higher ratio and thus higher debt levels.
29. Based on the sustainability metric most SOEs show a substantial debt overhang. In the empirical exercise, the net debt-to-EBITDA ratios were computed for 18 SOEs with bank debt at end-2016. Eleven SOEs (i.e. almost two-thirds) had ratios above the critical threshold of 5 or undefined ratios because of negative EBITDA.26 The remaining seven firms had ratios of below 5 or debt that after subtracting cash went into negative territory (i.e. no relevant indebtedness). Ratios ranged from 0.1 to 75.5 for the nine SOEs for which they could be calculated.27 The debt overhang—domestic and external debt combined—was then calculated as net debt minus 5 times EBITDA, amounting to MZN 30.7 billion for the overindebted SOEs, and the corresponding haircuts ranged from 47.7 to 99.9 percent of gross debt.
30. The calculated debt overhang would need to be addressed by either equity injections or debt restructuring. In principle, the SOEs’ owners, primarily the government, should inject fresh capital into the firms with which the excess debt would be repaid. However, in many cases this has not happened because the government lacks the resources to provide an appropriate capital cushion and thus sustainable leverage. As it happens, SOEs need to accumulate more debt to carry on operations and investment. At some point, the debt burden becomes excessive, and the debt or debt service need to be restructured. According to market participants, in Mozambique banks and SOEs tend to agree a lengthening of maturities, at times with periods of interest-only payments, or in some cases a modification of the loan terms (e.g. lower rates). Adjusting the principal is considered a measure of last resort. Therefore, in practice restructuring benefits the debt service but not the debt level.
31. As the with the debt stock, metrics exist for assessing the sustainable debt service. A prominent measure for a firm’s capacity to service its debt is the Interest Coverage Ratio (ICR) that relates a company’s earnings before interest and taxes (EBIT) to its interest expense. The lower the ratio, the more a firm is burdened by the debt service, with a ratio of 1 implying that it is not generating sufficient operational revenue to pay interest without making adjustments. Typically, if the ratio falls below a critical threshold of 1.5 a firm is seen as potentially encountering payment difficulties (Chow, 2015). In the exercise, excess debt service (using 2016 data) is determined as: SOE Interest Expense—(EBIT/1.5). The impact of a reduction of the contractual interest rate to make debt service viable would be felt in banks’ net interest income. If instead, the maturity is extended (and the interest rate not lowered), the bank’s liquidity position is impacted by the delay in repayment of the loan principal. In the exercise, eight out of 18 SOEs show an ICR of below 1.5 or negative EBIT (i.e. ICR undefined), with the corresponding necessary haircuts to the interest bill ranging from 74 percent to 100 percent.
32. In the exercise, banks were assumed to build provisions for the calculated debt overhang, with direct impact on capital ratios. To estimate the impact a matrix of claims of each bank on each SOE is used. The individual loans to overindebted SOEs were provisioned one by one by multiplying the carrying amount by the calculated haircut. The provisions were then aggregated for each bank and subtracted from bank capital (see Box 1).28 Provision data compiled by BM illustrate that most banks have not provided for loans to distressed SOEs (other than a general provision of 2 percent), and they are not required to do so thanks to government guarantees.29 Still, for prudence and in line with the introduction of IFRS 9 as of January 2018 requiring forward-looking provisioning,30 banks are here assumed to build a provision for the cumulative debt overhang that would need to be forgiven in a debt restructuring exercise for SOE debt to become sustainable again.
33. The results of the exercise illustrate that while the banking system could withstand the value adjustments, some banks’ capital would drop to critical levels. The total additional provisions amounting to MZN 15.5 billion would lower banks’ capital to MZN 44.8 billion. Set in relation to risk-weighted assets, this reduction corresponds to a drop in the system’s capital adequacy ratio (CAR) from 21.3 percent to 15.8 percent. One larger bank’s CAR would fall below the current minimum required (8 percent) and that of smaller bank below 11 percent. The largest drop in CAR experienced by any bank is 12 percentage points. The outcome of the sensitivity analysis clearly depends on the assumption of the critical threshold for EBITDA, but varying it to 7 times EBITDA (more lenient) or 3 times EBITDA (stricter) does not change the outcome materially (Table 3).
Debt Viability Check for SOEs and Impact on Banks
Step 1: Determine critical value of debt of each SOE based on the net debt-to-EBITDA ratio and using data from SOEs’ 2016 audited financial statements.
Step 2: Calculate debt overhang (debt minus critical amount) and, hence, needed haircut on the debt level (1—(sustainable debt/current debt)).
Step 3: Multiply each bank’s current loan exposure to an SOE (as of November 2017) by the respective haircut (%) and obtain required additional loan loss provision on the exposure.
Step 4: Sum up additional provisions for each bank.
Step 5: Reduce the total additional provisions by the amount of existing provisions and subtract resulting net additional provisions from each bank’s capital.
Step 6: Recalculate capital adequacy ratios.
Mozambique: Impact of Debt Restructuring Exercise on the Banking System
Mozambique: Impact of Debt Restructuring Exercise on the Banking System
| Threshold EBITDA=5 | Threshold EBITDA=7 | Threshold EBITDA=3 | |
|---|---|---|---|
| Provisions required | MZN 15.5 billion | MZN 12 billion | MZN 18.9 billion |
| CAR end-Sept. 2017 | 21.3 percent | 21.3 percent | 21.3 percent |
| CAR adjusted | 15.8 percent | 17.0 percent | 14.6 percent |
| No. banks CAR<8% | 1 | 1 | 1 |
| Largest drop in CAR | 12.0 pp | 9.8 pp | 14.2 pp |
Mozambique: Impact of Debt Restructuring Exercise on the Banking System
| Threshold EBITDA=5 | Threshold EBITDA=7 | Threshold EBITDA=3 | |
|---|---|---|---|
| Provisions required | MZN 15.5 billion | MZN 12 billion | MZN 18.9 billion |
| CAR end-Sept. 2017 | 21.3 percent | 21.3 percent | 21.3 percent |
| CAR adjusted | 15.8 percent | 17.0 percent | 14.6 percent |
| No. banks CAR<8% | 1 | 1 | 1 |
| Largest drop in CAR | 12.0 pp | 9.8 pp | 14.2 pp |
34. As mentioned, it is more likely that banks would accept a modification of the loan terms than outright debt forgiveness, yet this could still have a significant impact. The haircuts to the interest bill of the eight SOEs with an Interest Coverage Ratio of below 1.5 (or negative EBIT) amount to MZN 17.1 billion. This reduction would be applied to both domestic and external debt. It is clear that net interest income of banks would be materially reduced by any such debt service restructuring and the currently comfortably high return on assets of 2.5 percent lowered materially.
E. Conclusions and Recommendations
35. This paper has assessed the macrofinancial linkages playing a central role in the propagation of adverse shock between the fiscal/real sector and the banking sector. High indebtedness of the government sector, including major SOEs, and tight financial conditions affecting the real sector have been creating macrofinancial vulnerabilities that, if unraveling abruptly in the face of external shocks, could lead rising loan delinquencies, eroding banks’ capital buffers. The two sensitivity analyses quantify the impact of private sector loan defaults and necessary restructuring of excessive SOE debt onto the banking sector. The quantitative assessment suggest that while the banking system would be able to absorb the shocks, individual institutions would become undercapitalized, highlighting the need to take preventive action.
36. Were the contemplated macrofinancial risks to materialize, banking sector fragility could cause feedback effects onto the real economy and government. Negative feedback effects onto the real sector would result from banks’ absorption of additional losses from exposures to ailing SOEs and private firms. As banks repair their balance sheets, adjusting to the riskier credit environment, they could restrict the supply of new credit supply or raise loan rates. As a result, the private sector could be deprived of credit for working capital on sustainable terms in the short term, which in turn would likely harm economic growth and in the process, affect tax revenue. The reduced fiscal space would, in turn, make arrears clearance and the overdue restructuring of major SOEs less viable, potentially triggering a vicious circle.
37. The following recommendations are aimed at dealing with the identified vulnerabilities:
Banks should provision adequately given the debt overhang of major SOEs and potential debt restructuring. As the empirical analysis shows, more than half of the SOEs with bank debt are overindebted by standard metrics and require reduction of the debt level or at least of the debt service burden. While at present a mix of forbearance and light loan restructuring prevails, more strident measures may have to be taken to restore SOE’s debt sustainability, notably corporate restructuring and financial restructuring (reduction of the loan rate and/or the principal). It is recommended that BM urge the banks to build adequate loan loss provisions for this eventuality following the expected credit loss approach now required by IFRS 9.
Supervisors should monitor closely the impact of the updated regulation concerning regulatory capital, minimum liquidity requirements and the FX market changes. They should gauge the impact of the market liquidity stance on the stability of the sector and its capacity to support the rest of the economy. NPLs are likely to continue growing and banks need to protect their buffers. Supervisors should remain watchful of loan classification, collateral valuation, and provisioning. Preparedness for early remedial action is warranted. BM’s intention to reform the banking law is a step in the right direction to allow implementation of better resolution tools together with the financial safety net.
Central Bank empowerment with an explicit mandate for macroprudential policy together with capacity building and adequate resources is warranted. BM is encouraged to make efforts to deliver macroprudential policies by enhancing its capacities in macroprudential analysis, and produce a financial stability report to communicate on financial stability issues. Also, publication of financial stability indicators is recommended as is elaboration of the balance sheet matrices and compilation of real estate indicators including housing prices.31
Exposures of the financial system, SOEs loans, government securities and exposure of firms to SOEs should be closely monitored. Also, it is recommended to monitor carefully the spillovers of arrears across economic sectors and implement an action plan to prevent and manage arrears within a fiscal consolidation strategy. Validating these arrears and gradually clearing them should be a priority after reaching agreement on clearance modalities with creditors.
Reinvigorate improvements in debt management. The debt unit’s capacity needs to be strengthened to exercise effective oversight over the entire public debt portfolio, including SOEs and loans that are part of cooperation agreements signed by other line ministries. Moreover, implementing a strong action plan to strengthen governance, improve transparency and ensure accountability in debt management is crucial.
Broadening statistical coverage to non-bank financial institutions and general government are essential to progress in diagnosing macro-financial linkages and risks. Developing an integrated sectoral balance sheet information is fundamental for national surveillance. Statistical efforts needed are:32 (i) increase coverage of monetary and financial statistics to other depositary corporations, (ii) complete the compilation of fiscal statistics under GFSM 201433 including reconciliation of government domestic securities by instrument and holder and producing the government balance sheet, and (iii) determine the stock of government expenditures arrears and distribution across sectors and the stock of SOEs arrears with private sector suppliers. Also, disaggregating assets and liabilities by maturity (e.g. short and long term) is needed.34
References
Chow, J.T.S. (2015): “Stress Testing Corporate Balance Sheets in Emerging Economies”, IMF Working Paper No. 15/216.
Cohen, B.H. and G.A. Edwards, Jr. (2017): “The new era of expected credit loss provisioning”, BIS Quarterly Review, March, pp. 39–56.
Drehmann, M., C. Borio, L. Gambacorta, G. Jiménez, and C. Trucharte, (2010): “Countercyclical Capital Buffers: Exploring Options,” BIS Working Papers, No. 298 (Basel: Bank for International Settlements).
IMF (2011): “The Quest for Leading Indicators of Financial Distress” in Global Financial Stability Report. September.
IMF (2014): “Staff Guidance Note on Macroprudential Policy—Detailed Guidance on Instruments”, IMF Policy Paper, November.
IMF (2015): “Balance Sheet Analysis in Fund Surveillance”, IMF Policy Paper, June.
IMF (2017): “Approaches to Macrofinancial Surveillance in Article IV Reports”, IMF Policy Paper, March.
Osano, H.M and H. Languitone. (2016): “Factors influencing access to finance by SMEs in Mozambique: case of SMEs in Maputo central business district”, Journal of Innovation and Entrepreneurship pp. 5–13.
Republic of Mozambique (2015): “National Financial Inclusion Strategy 2016 – 2022”.
Simione, F. and Y. Xiao. (2016): “Monetary Policy, Banking Structure and Interest Rates in Mozambique” in Selected Issues, IMF country report No.16/10.
World Bank (2016a): “Mozambique Economic Update: Navigating Low Prices”, Mozambique Economic Update Series, March.
World Bank (2016b): “Mozambique Economic Update: Facing Hard Choices”, Mozambique Economic Update Series, December.
By Mario Mansilla, Torsten Wezel, and Harold Zavarce.
Non-financial Public Corporations (NFPC) include 14 public corporations fully owned by the government and, in addition, private companies in which the government is a shareholder. Indirect and direct state shareholdings are not fully disclosed. A NFPC is considered an State-Owned Enterprise (SOE) if the government holds a share of 50 percent or more.
The 2016 disclosure triggered a sovereign debt downgrade from B- to CC by S&P.
The maturity of Proindicus and MAM loans are March 2021 and May 2019, respectively.
By end-2017 sovereign external arrears have been incurred including the Mozam Eurobond coupon, debt service of Proindicus and MAM, and two loans of the state-owned airports company, Aeroportos de Moçambique (AdM), for which a state guarantee has been called.
Balance Sheet Analysis (BSA) compiles all the main balance sheets in an economy using aggregate data by sector into a balance sheet matrix that shows asset and liability positions between key sectors. In Figure 2, the balance sheet matrix has as columns the creditor sector (i.e. the holder of the liability) and in the rows debtor sector. Each cell contains a net claim position (asset minus liability). The matrix serves as a starting point to diagnose risks and potential transmission channels of shocks, and sets the stage for deeper analysis (IMF 2015a).
In Mozambique, the increase in foreign currency borrowing by public non-financial corporates poses a substantial currency mismatch risk, if indeed unhedged. Also, public guaranteed borrowing is a source fiscal risk.
This information is preliminary and includes arrears incurred during 2014–16.
Risks correspond to the identified sources in the Risk Assessment Matrix (Staff Report for the Article IV 2017).
Risk may propagate through a recurrence of banking stress, public and private corporates liquidity issues, and household weak household demand.
This corresponds to the so-called prudential filter: voluntary provision above IFRS requirements.
One-sided Hodrick-Prescott filter with a smoothing factor of 400,000 appropriate for monthly data. A credit-to-GDP ratio exceeding ten percentage points or more above trend (upper threshold) issues the strongest signal of an impending crisis in terms of the noise-to-signal ratio (Drehmann and others, 2010). The lower threshold of 2 percent above the trend signals excessive credit growth or overheating (IMF 2014).
The Metical nominally has appreciated by almost 24 percent against the USD since October 2016.
The weight of imported products in the consumer basket and food are important. Annualized food inflation peaked at 42 percent in November 2016 against lows of 5.5 percent in December 2017.
In April 2017, in April 2017, the BM changed its operational target to a short-term interest rate (MIMO) as part of a longer-term transition to inflation targeting. The MIMO rate was set initially at 21.75 percent and the lending rate cut to 22.75 percent. See Selected Issues paper on transitioning the monetary policy regime.
Simeone and Xiao (2016, pp. 9–11) argue that despite Mozambique’s rapid banking sector expansion experienced in 2005–2014, interest rates on bank loans remained prohibitively high for SMEs, partly related to high market concentration. In addition, Osano and Languitone (2016) show that access to finance is also constrained by collateral requirements, structure of the financial sector (market concentration) and lack of awareness of funding opportunities among banks and entrepreneurs.
The main targets are: (i) 60 percent of the adult population with physical or electronic access to financial services; (ii) 100 percent of the districts with at least one formal access point; and (iii) 75 percent of the population with one access point within five kilometers. A 2018 interim target postulating that 40 percent of the adult population have access to electronic money has already been met.
T-bill auctions are conducted every Wednesday mainly for monetary purposes; recently some auctions were deserted. Some institutions have binding constraints (set internally or with their parent banks) for increasing their exposure to sovereign risk, which have been further tightened by the downgrade of the country’s risk rating.
Eliminating unnecessary costs in the process, especially for exporters which products have higher import content
Tight net open position regulations prompt banks to regularly sell the BM their surplus FX balances, which have increased given the low FX demand for imports.
The BM has been reorganized internally to implement greater financial surveillance through specialized divisions.
Two supervisory entities, the Ministry of Finance and Economy (MEF) and the State-owned Equity Holdings Management Institute (IGEPE) each control nine bank-indebted SOEs. While for most SOEs the existence of bank debt was verified by BM data as at end-November 2017, the bank debt of a few IGEPE-controlled firms that is not reported in the data set provided by BM is based on a survey reportedly conducted in the first half of 2017.
To what extent potential capital expenditure is limited by the financing constraint would have to be determined in each case. It also is not clear whether individual SOEs have run into lending limits notwithstanding the availability of state guarantees and comfort letters.
Interest payments are excluded because they depend on the financing structure of a firm (debt vs. equity financing); taxes vary across firms due to various idiosyncrasies, and depreciation and amortization costs depend on historical investments that firms have made and not necessarily on the current operating performance of the business. Still, EBITDA is not a perfect proxy for cash flow because actual cash flow depends critically on required capital expenditure to preserve the firm as a going concern; this expenditure may vary significantly across firms and industries, which is why analysis of free cash flow is conceptually superior. This said, for analyzing debt sustainability across the board, the net debt-to-EBITDA ratio has become analysts’ preferred measure.
In a few cases, cash flow was taken as the denominator if it was positive as opposed to a negative EBITDA.
For SOEs with negative EBITDA the ratio was undefined but would clearly be above the critical threshold if those firms instead showed small positive earnings or cash flows.
Risk weights on SOE exposures are ordinarily 100 percent (as with other corporate loans) and are unaffected by the increase in provisions. Therefore, only the numerator of the capital adequacy ratio needs to be recalculated.
Six out of 13 banks with SOE exposures have provisioned exactly 2 percent of the exposure (i.e. the general provision) and another three banks even less than that. Two banks have a provisions coverage ratio of 5 percent, although one bank has sizable impaired loans that arguably could have been provisioned more already. The remaining two banks have a coverage ratio of 69 and 100 percent, with provisions actually exceeding the impaired loan amount.
Under IFRS 9, expected credit losses have to be recognized to reflect the credit risk of financial assets. It is no longer necessary for a “trigger event” to have occurred before credit losses are recognized through provisions (or write-offs). Whenever credit quality is deemed to have deteriorated significantly and can no longer be deemed low risk, the loan would require a provision for the lifetime expected credit loss (Stage 2 of the three-stage classification system under IFRS 9); see Cohen and Edwards (2017). This treatment may conflict with provisioning regulation imposed by the central bank but the issue can be handled by adjusting regulatory capital for any difference between IFRS 9-based provisions and regulatory provisions.
Measuring housing wealth, mainly related to property values, will serve to assess balance sheet effect on household consumption and its feedbacks with mortgage financing.
The compilation was initiated in 2011. In June 2017, the central bank broadened the investor base for primary issuance of T-bills to include non-banks financial institutions. Producing the government balance sheet Table 6 of GFSM2014, including data on government security by holder and maturity, is needed. Technical assistance in this area is scheduled for 2018–2019.
This information is key to diagnose an increase in the maturity mismatch between assets and liabilities that could lead to a liquidity spiral creating financial instability and a bank run.
