## cr18107 ## Source details **Canonical URL:** [cr18107](https://www.imf.org/-/media/files/publications/cr/2018/cr18107.pdf) ## Other formats - [Markdown version](/-/media/files/publications/cr/2018/cr18107.pdf.md) - [Structured JSON version](/-/media/files/publications/cr/2018/cr18107.pdf.json) --- ### Growth experience and benchmarking - WAEMU experienced its longest episode of rapid economic growth since 2012; since 2012, WAEMU’s real GDP has grown by more than 6 percent a year. - Long-run averages (1960-2017): - WAEMU: average growth = 3.4 percent; growth volatility = 4.1 percent. - LIDCs: average growth = 4.0 percent; growth volatility = 3.6 percent. - Real GDP per capita (PPP) path: - WAEMU’s income per capita was 108 percent of the LIDC group in the early 1960s and dropped to 65 percent in 2017. - Heterogeneity across members since the 1960s: - Per capita income increased in Benin, Burkina Faso, and Mali. - Improved slightly in Guinea-Bissau and Senegal. - Decreased in Côte d’Ivoire, Niger, and Togo. - Relative country size in the union: - Côte d’Ivoire accounts for more than 40 percent of the currency union’s GDP. - Senegal accounts for about 15 percent. - Mali, Burkina Faso, and Benin each contribute approximately 10 percent. ### Benchmarking approach and key fundamentals (1997-1999 vs 2012-2017) - Benchmark periods: 1997-1999 (reference) and 2012-2017 (recent acceleration). - Fundamentals examined: (i) macroeconomic stability, (ii) trade performance, (iii) investment, (iv) polity score, (v) terms of trade, (vi) total factor productivity (TFP). Macroeconomic stability (selected table evidence; difference = 2012-2017 minus 1997-1999): - Inflation rate (WAEMU): 1997-1999 = 4.2; 2012-2017 = 1.0; Difference = -3.2** - Inflation rate (LIDCs): 1997-1999 = 12.6; 2012-2017 = 5.9; Difference = -6.7*** - Inflation volatility (WAEMU): 1997-1999 = 5.7; 2012-2017 = 1.7; Difference = -4.0*** - Exchange rate depreciation (WAEMU): 1997-1999 = 39.1; 2012-2017 = 24.2; Difference = -14.9 - Exchange rate depreciation (LIDCs): 1997-1999 = 71.8; 2012-2017 = 58.9; Difference = -12.9 - Exchange rate volatility (WAEMU): 1997-1999 = 70.4; 2012-2017 = 37.1; Difference = -33.3*** Trade performance: - Trade openness = (exports + imports) / GDP. - WAEMU: 1997-1999 = 52.7; 2012-2017 = 65.3; Difference = 12.6*** - LIDCs: 1997-1999 = 81.2; 2012-2017 = 96.0; Difference = 14.8*** Investment: - Total investment-to-GDP: - WAEMU: 1997-1999 = 15.9; 2012-2017 = 22.6; Difference = 6.7*** - LIDCs: 1997-1999 = 23.5; 2012-2017 = 24.1; Difference = 0.6 - Public investment-to-GDP: - WAEMU: 1997-1999 = 5.7; 2012-2017 = 9.1; Difference = 3.4*** - LIDCs: 1997-1999 = 5.3; 2012-2017 = 6.7; Difference = 1.4*** - Country highlights: investment increased particularly in Niger, Togo, Senegal and Burkina Faso well above WAEMU average of 6.7 percent. Polity score: - WAEMU: 1997-1999 = 0.1; 2012-2017 = 4.1; Difference = 4.1*** - LIDCs: 1997-1999 = 3.2; 2012-2017 = 4.6; Difference = 1.4*** Terms of trade: - WAEMU average change in terms of trade (2012-17) = 1.2 percent; LIDCs (2012-17) = -0.3 percent. - Improvements concentrated in Guinea-Bissau, Burkina Faso, Mali, and Côte d’Ivoire. Total Factor Productivity (TFP) contribution (points; growth accounting): - WAEMU TFP: 1997-1999 = 0.57; 2012-2017 = 1.03; Difference = 0.46 - LIDCs TFP: 1997-1999 = 1.03; 2012-2017 = -0.45; Difference = -1.48*** - Country-level TFP increases notable in Togo, Côte d'Ivoire, Niger, and Guinea-Bissau. ### Growth acceleration identification and empirical approach - Growth acceleration definition (modified Hausmann, Pritchett, Rodrik (2005)): 1. Real GDP growth rate is at least 2 percent; 2. Post-acceleration growth rate is at least 3.5 percent per year, and last for at least 5 years; 3. Level of post-acceleration output must exceed the pre-episode peak level of income. - Model: Probit baseline with independent variables lagged one year and year-fixed effects; robustness checks with Logit and Linear Probability Models. - Episodes identified (1960-2017) among WAEMU members: - Côte d’Ivoire: 2 episodes (1964, 2012) - Burkina Faso: 2 episodes (2001, 2009) - Togo: 2 episodes (1964, 2008) - Benin: 1 episode (1994) - Niger: 1 episode (2011) - Senegal: 1 episode (2012) - Guinea-Bissau and Mali: 0 episodes ### Empirical determinants of growth accelerations (Probit estimates, column (1)) - Positive and statistically significant determinants (coefficient; robust standard error): - Investment-to-GDP t-1: 0.0096*** (0.0035) - Trade openness-to-GDP t-1: 0.0043** (0.0019) - Exchange rate depreciation t-1: 0.0364* (0.0188) - Polity score t-1: 0.0229*** (0.0051) - Terms of trade t-1: 0.0049* (0.0030) - Total factor productivity t-1: 0.0303*** (0.0071) - Negative and statistically significant determinants: - Inflation rate t-1: -0.0956** (0.0447) - Exchange rate volatility t-1: -0.0437*** (0.0123) - Insignificant: - Inflation volatility t-1: -0.0342 (0.0380) - Model statistics: - Observations = 1,850 - Number of countries = 60 - R-squared = 0.051 - Robustness: Results persist in Logit and Linear Probability Models; scatter plots corroborate positive correlations with investment, trade openness, exchange rate depreciation, polity score, terms of trade, TFP and negative correlations with inflation and exchange rate volatility. ### Summary conclusion on growth drivers - WAEMU’s recent performance coincides with: - an increase in macroeconomic stability; - an increase in investment (total and public); - improvement in institutional quality (Polity score); - improved terms of trade; and - higher TFP contribution to growth. - Empirical conclusion: investment, trade openness, exchange rate depreciation, political institutions, terms of trade, and TFP increase the likelihood of growth accelerations; higher inflation and greater exchange rate volatility lower that likelihood. --- ### WAEMU banking system: size, vulnerabilities and stress tests Banking system indicators and access: - Broad money grew by about 12 percent per year on average over 2012-17. - Gross assets of commercial banks reached 56 percent of the regional GDP at end-2016 as compared with 43 percent at end-2013. - Number of clients’ accounts increased from 7.8 million to 10.3 million over the period. - Estimated share of households with a bank account rose from about 7 percent in 2007 to 16 percent in 2016. Institutions and ownership (2017): - Total credit institutions: 140 (125 banks including 17 bank branches; 15 non-bank credit institutions including 4 branches). - By ownership (excluding branches): 94 foreign-owned, 12 public, 13 privately held regional credit institutions. - By size: 31 large (accounting for more than 10 percent of banking sector assets in country), 25 average (more than 5 percent of assets), 63 relatively small. Capital and solvency: - Average capital adequacy ratio (CAR) declined to 11.4 percent at mid-2017 from 12.6 percent at end-2015. - As of end-2016, 12 banks — accounting for nearly 9 percent of the system’s assets — did not meet the minimum CAR of 8 percent (on a Basel I basis). - Up to 29 banks and credit institutions did not meet the new minimum capital requirement of FCFA 10 billion effective at mid-2017 and were given until mid-2018 to comply. - Capital shortfall of undercapitalized WAEMU banks equaled about 0.4 percent of the zone’s GDP, of which about 2/3 came from the capital shortfall in public banks. Concentration and NPLs: - At end-September 2017, 12 leading bank groups accounted for about ¾ of the market share, total assets, and clients’ accounts. - Ratio of loans to the top-5 borrowers in total loans: 90 percent at mid-2017 (down from 101 percent at mid-2016). - Gross NPLs to total loans: 13.8 percent at end-2016; 14.6 percent at mid-2017; 15.1 percent at end-September. - Provisions cover about 63 percent of NPLs. Net NPLs to total loans: 5.9 percent at mid-2017; 6.2 percent at end-September. - Net NPLs to capital: 47.2 percent (2016 Dec.) and 49.7 percent (2017 Jun.). Public banks: - Out of 12 public banks in WAEMU, 7 were undercapitalized; some had significantly negative capital. - One public bank was classified as large and another as medium-sized in their countries of operations. Liquidity and central bank interaction: - Net aggregate liquidity of the banking system has been negative since 2014 and reached about FCFA 3,000 billion at end-2016 (about 5 percent of the zone’s GDP). - BCEAO limited banks’ access to the refinancing facility to twice their own capital effective from June 2017. - BCEAO lowered the obligatory reserve requirement ratio from 5 percent to 3 percent in March 2017. - BCEAO actions resulted in the reduction of the banking system’s net liquidity position by about FCFA 600 billion over the first half of 2017. - Banking system’s net liquidity deficit covered by the BCEAO refinancing came to around FCFA 2,900 billion at end-2017. - In 2015-16, government debt issuances were typically oversubscribed; subscription dipped to about 66 percent in February 2017 and later fell to 41 percent in December 2017. Stress test outcomes (Data as of June 30, 2017) — Credit Portfolio Tests (Table 2): - Before stress tests: Average CAR 11.42; Undercapitalized banks (CAR < 8 percent) 15; Capital shortfall, FCFA billion 268.0; Public banks undercapitalized banks 7; Average CAR -1.88. - Largest borrower default: Average CAR 3.11; Undercapitalized banks 76; Capital shortfall, FCFA billion 995.9; Public banks undercapitalized banks 9; Average CAR -10.79. - Two largest borrowers default: Average CAR -2.57; Undercapitalized banks 91; Capital shortfall, FCFA billion 1,667.4; Public banks undercapitalized banks 10; Average CAR -17.16. - NPLs increase by 50 percent: Average CAR 6.06; Undercapitalized banks 57; Capital shortfall, FCFA billion 684.0; Public banks undercapitalized banks 9; Average CAR -16.51. - NPLs increase by 75 percent: Average CAR 3.13; Undercapitalized banks 67; Capital shortfall, FCFA billion 1,042.8; Public banks undercapitalized banks 12; Average CAR -25.52. - 15 percent of standard loans become NPLs: Average CAR 5.14; Undercapitalized banks 62; Capital shortfall, FCFA billion 790.3; Public banks undercapitalized banks 11; Average CAR -13.26. Liquidity stress (deposit run over five days): - Demand deposits outflows: 5 percent, 5 percent, 5 percent, 10 percent, 10 percent; term deposits outflows: 3 percent, 3 percent, 3 percent, 5 percent, 5 percent. - Banks needing liquidity support by day: Day 1: 20; Day 2: 26; Day 3: 33; Day 4: 56; Day 5: 79. - In a sustained deposit run up to 35 percent of demand deposits and 19 percent of term deposits over 5 days, up to 79 banks would need liquidity support. Macro-financial linkages and sectoral exposures (end-2017): - Household deposits: about 11.3 percent of regional GDP. - Credits to households: 6.6 percent of GDP. - Deposits from private companies: about 15.4 percent of GDP. - Credits to private companies: about 18.7 percent of GDP (net claims on private non-financial companies of about 3.4 percent of GDP). - Credits to public non-financial companies: 2.1 percent of GDP; deposits from them: 1.7 percent of GDP. - Commercial banks’ net claims on the central governments: 10.2 percent of GDP. - Net liquidity support from BCEAO: about 4.7 percent of GDP. - Composition of bank credit to economy (June 2017, Percent): Retail and wholesale trade, restaurants and hotels: 31.4; Transportation and communication: 11.9; Other services: 13.3; Insurance, real estate and services: 7.3; Construction: 11.7; Manufacturing: 12.7; Agriculture: 4.7; Extractive industries: 1.3; Electricity, water and gas: 5.6. - Services sector accounted for nearly 2/3 of all bank credit; productive sectors accounted for the remaining 1/3. Policy-relevant priorities: - Enforce compliance with prudential norms. - Recapitalize or resolve problem banks. - Address pockets of vulnerabilities: high exposure to a few largest borrowers, weaknesses in public banks, structural liquidity deficits and reliance on BCEAO refinancing. --- ### Fiscal risks and the need for a comprehensive fiscal risk management framework - Key objective: governments must identify and quantify risks before they can be managed (three-step framework: identify/quantify risks; decide whether to mitigate; decide whether to provision). - GFSM 2014 guidance: record stocks and flows on an accrual basis, report contingent liabilities as memorandum items, draw up Summary Statement of Explicit Contingent Liabilities and Net Implicit Obligations for Future Social Security Benefits. - Full use of GFSM 2014 remained rare; balance sheet data reported in 15 percent of IMF staff reports in early 2017. WAEMU fiscal developments and risks: - Between 2012 and 2017 the gross debt-to-GDP ratio increased by 17 percentage points of GDP. - Region-wide growth rates have been more than 6 percent every year since 2012. - WAEMU Debt-to-GDP convergence criterion = 70% - Most member states have diverged from the deficit convergence criterion of 3 percent of GDP in recent years. - Residual (stock-flow discrepancy) accounted for 5.7 percentage points of GDP of the 16.9 percentage points of GDP increase in gross debt between 2013 and 2017 across eight WAEMU countries. - Unanticipated increases in debt: difference between 2017 debt-to-GDP projections and 2017 projections made in 2012 = 15 percentage points of GDP (unweighted average), with peaks of more than 30 percent in Benin and Togo. - For all WAEMU countries, excluding Guinea-Bissau, the unanticipated share of the total increase was larger than 65 percent of the total increase. Main sources of fiscal risks: - Macro-economic shocks (average fiscal cost when materialized cited as context). - Contingent liabilities: SOEs stand out; PPPs likely to grow in importance; pensions not immediate issue; financial system risks potentially large; natural disasters. - Institutional PFM weaknesses: coverage, timeliness, expenditure control, revenue forecasting and administration. Fiscal risks from SOEs: - SOE interventions costly: context cites average cost of government intervention in SOEs over 1990-2014 about 3 percent of GDP (advanced/emerging). - Quasi-fiscal deficits in electricity sector averaged around 1.3 percent of GDP across 7 WAEMU countries in 2013-2014; hypothetical lower oil-price scenario still implies almost 1 percent of GDP. - Improving SOE reporting and coverage is crucial: classify SOEs, apply GFSM 2014 distinctions, and improve valuation where market tests are difficult. PPPs: - PPPs can create debt-like or contingent obligations often not included in fiscal aggregates. - PPPs are not widespread in WAEMU but feature in development plans; legal frameworks and institutions are being updated and IMF technical assistance provided. - Example country PPP framework status summarized across WAEMU members (laws, missions, structural benchmarks) in source. PFM and stock-flow discrepancies: - Debt stocks reported on cash basis while fiscal balances are reported on commitment basis, contributing to residuals. - Identified drivers of positive residuals by country include pre-financing, coverage, arrears, pensions, data issues, and roll-over of appropriations. - Residual + other identified debt-creating flows: unweighted annual average for all eight countries over 2013-2016 = 1.3 percent of GDP; peaked at 1.6 percent in 2016. - Country examples: - Benin: approved road projects worth 24 percent of GDP approved without budget allocations; project execution about 3 percent of GDP in 2016. - Togo: pre-financed expenditures averaged 3.5 percent of GDP during 2013-16. - Côte d’Ivoire: government bonds issued in 2014 to pay off past domestic arrears worth 1 percent of GDP. - IMF programs include conditionality addressing pre-financing, SOEs, PPPs, and coverage improvements (Benin, Togo, Senegal, Côte d’Ivoire examples). Recommendations for fiscal risk management and PFM improvements: - Move to accrual accounting and audited government financial statements. - Improve fiscal projections and medium-term fiscal frameworks, expenditure allocation and control. - Develop stocks aspect of fiscal tables, report opening/closing public sector balance sheets. - Broaden fiscal table scope to general government. - Use Fiscal Transparency Code and prioritize fiscal transparency and fiscal risk management reforms. - Improve SOE inventories, reporting, and oversight; centralize PPP approval and gatekeeping; apply P-FRAM and other tools. --- ### Public Investment Management and investment efficiency Public Investment Management Assessment (PIMA): - PIMA evaluates 15 key institutions for planning, allocation, and implementation of public investment. - PIMA missions conducted between October 2015–June 2017; initial results cover 23 piloted countries. - The 15 institutions cover planning (Fiscal Rules; National & Sectoral Planning; Central-Local Coordination; Management of PPPs; Company Regulation), allocating (Multi-year Budgeting; Budget Comprehensiveness; Budget Unity; Project Appraisal; Project Selection), and implementing (Protection of Investment; Availability of Funding; Transparency of Execution; Project Management; Monitoring of Assets). WAEMU pilot results: - Regulatory framework slightly better in national and sectoral planning, budget comprehensiveness and protection of investment. - Weaker regulations on average in central-local coordination, management of PPPs, company regulation, monitoring of assets. - Identified implementation gaps: multiyear budgeting, procurement enforcement, monitoring of projects, and asset registration. Public investment efficiency (DEA and regressions; period 2000-2015): - WAEMU efficiency scores (DEA hybrid and physical indicators): - WAEMU: Physical Infrastructure 0.3694; Hybrid Indicator 0.6188. - Sub-Saharan Africa: Physical Infrastructure 0.4597; Hybrid Indicator 0.6417. - CEMAC: Physical Infrastructure 0.3046; Hybrid Indicator 0.5108. - EAC: Physical Infrastructure 0.4875; Hybrid Indicator 0.7351. - Oil exporters: Physical Infrastructure 0.1958; Hybrid Indicator 0.2687. - Non-resource-intensive: Physical Infrastructure 0.4464; Hybrid Indicator 0.6981. - Other resource-intensive: Physical Infrastructure 0.6019; Hybrid Indicator 0.6563. - Interpretation: WAEMU’s performance lags comparator groups; inefficiency magnitude depends on output index choice. Correlation between per capita public capital stock and perceptions of infrastructure quality is positive but weak. Determinants (cross-country regressions; sample 154 countries, results summarized): - Control of Corruption (estimate) positively and significantly associated with hybrid efficiency score: - Control of Corruption: 0.11292 (3.13); Observations: 93; R-squared: 0.216. - Regulatory Quality positively and significantly associated: - Regulatory Quality: 0.11552 (3.130); Observations: 93; R-squared: 0.220. - Natural resource dependency negatively associated: - Natural Resource dependency: -0.12216 (-2.129) and -0.11708 (-2.030) across specifications. - Tobit robustness: - Control of Corruption: 0.11783 (4.588); Observations: 103; R-squared: 0.224. - Regulatory Quality: 0.10568 (4.148); Observations: 103; R-squared: 0.206. - ODA and urban population show mixed or negligible associations. Policy implications: - Institutional improvements (control of corruption, regulatory quality) can substantially improve public investment efficiency; a 10 percent increase in institutional indices could improve efficiency by about 15 percent on average (based on specifications). - Strengthen PIM institutions: better project appraisal, selection, multiyear budgeting, central-local coordination, PPP management, project monitoring, and asset registers. - Address natural resource dependency challenges to mitigate negative association with efficiency. --- ### Revenue administration and PI indicators (selected) - PI-3 (Revenue outturn): measures change in revenue between original approved budget and end-of-year outturn; uses M2 (AV) aggregation. - PI-19 (Revenue administration): assesses entities administering central government revenues, procedures to collect and monitor revenues; contains four dimensions and uses M2 (AV). - PI-20 (Accounting for revenue): assesses procedures for recording and reporting revenue collections, consolidation and reconciliation; contains three dimensions and uses M1 (WL). - Public investment efficiency analysis used inputs: real public capital stock per capita and per capita GDP; outputs: physical and hybrid infrastructure indicators combining roads, electricity, water access, secondary teachers, hospital beds, and survey-based quality measures. *International Monetary Fund — WEST AFRICAN ECONOMIC AND MONETARY UNION (cr18107).* ### References _____________________________________________________________________________ 21 ### cr18107 - References ### Growth experience and stylized facts - WAEMU experienced its longest episode of rapid economic growth since 2012. - Average growth and volatility, 1960-2017: - WAEMU: average growth = 3.4 percent; growth volatility = 4.1 percent. - LIDCs: average growth = 4.0 percent; growth volatility = 3.6 percent. - Since 2012, WAEMU’s real GDP has grown by more than 6 percent a year. - Real GDP per capita (PPP) path and shares: - WAEMU’s income per capita was 108 percent of the LIDC group in the early 1960s and dropped to 65 percent in 2017. - Heterogeneity across WAEMU members since the 1960s: - Per capita income increased in Benin, Burkina Faso, and Mali. - Per capita income improved slightly in Guinea-Bissau and Senegal. - Per capita income decreased in Côte d’Ivoire, Niger, and Togo. - Relative country size in the union: - Côte d’Ivoire accounts for more than 40 percent of the currency union’s GDP. - Senegal accounts for about 15 percent. - Mali, Burkina Faso, and Benin each contribute approximately 10 percent. ### Factors associated with growth accelerations (literature survey) - Early cross-country growth regressions identified key determinants: - Kormendi and Meguire (1985): growth negatively associated with monetary variance, government consumption, and inflation; positively associated with population growth and investment. - Barro (1991): growth of real GDP per capita positively associated with human capital via reduced fertility and increased physical investment. - Mankiw, Romer, and Weil (1992): saving, education, and population growth explain much international variation in income per capita. - Later work added terms of trade, democracy, and investment in transport and communication as controls. - Literature focused on growth acceleration episodes (turning points) finds recurring correlates: - Hausmann, Pritchett, and Rodrik (2005): growth accelerations correlated with increases in investment and trade, exchange rate depreciations, and political-regime changes. - Pattillo, Gupta, and Carey (2005): accelerations associated with improved institutional quality (ICRG indicators), TFP growth, and low debt burdens. - Jones and Olken (2008): TFP rather than factor accumulation correlated with growth accelerations; also associated with international trade and low inflation. - Berg, Ostry, and Zettelmeyer (2012): highlighted income inequality, political institutions, and macroeconomic stability. - Arizala and others (2017): in SSA, increases in investment and trade openness, low inflation, improved fiscal balance, export diversification, and institutional improvements correlate significantly with growth accelerations. ### Approach of the paper - Objective: examine fundamentals behind WAEMU’s recent growth acceleration relative to LIDCs. - Methodology overview: - Present stylized facts on WAEMU growth performance. - Survey literature on growth regressions and growth acceleration models. - Benchmark changes in key macroeconomic variables in the WAEMU and peers during the last six years relative to an earlier reference period in the 1990s. - Explore empirical relationship between those variables and average GDP growth during growth acceleration episodes. - Peer group definition: - LIDCs defined following the World Economic Outlook (WEO) classifications; the group constitutes 60 countries, including WAEMU countries. *Source: cr18107 - References. March 13, 2018.* ### 11. The benchmarking analysis looks at changes in macroeconomic fundamentals ### 11. The benchmarking analysis looks at changes in macroeconomic fundamentals ### Benchmarking approach - WAEMU member countries are benchmarked against LIDCs peers during two time periods: 1997-99 and 2012-17. - The period 2012-17 corresponds to the recent growth acceleration episode in the WAEMU. The period 1997-99 avoids capturing effects of civil conflict and the 1994 CFA franc devaluation. - Key fundamentals examined: (i) macroeconomic stability, (ii) trade performance, (iii) investment, (iv) polity score, (v) terms of trade, and (vi) total factor productivity (TFP). ### Macroeconomic stability - Stability proxied by: inflation rate and volatility, exchange rate depreciation and volatility. - Data show WAEMU has benefitted from greater macroeconomic stability than LIDCs (Table 1). - Selected table evidence (1997-1999 vs 2012-2017; difference = 2012-2017 minus 1997-1999): - Inflation rate (WAEMU): 1997-1999 = 4.2; 2012-2017 = 1.0; Difference = -3.2** - Inflation rate (LIDCs): 1997-1999 = 12.6; 2012-2017 = 5.9; Difference = -6.7*** - Inflation volatility (WAEMU): 1997-1999 = 5.7; 2012-2017 = 1.7; Difference = -4.0*** - Exchange rate depreciation (WAEMU): 1997-1999 = 39.1; 2012-2017 = 24.2; Difference = -14.9 - Exchange rate depreciation (LIDCs): 1997-1999 = 71.8; 2012-2017 = 58.9; Difference = -12.9 - Exchange rate volatility (WAEMU): 1997-1999 = 70.4; 2012-2017 = 37.1; Difference = -33.3*** ### Trade performance - Trade openness measured as (exports + imports) / GDP. - Levels and changes: - WAEMU average trade openness: 1997-1999 = 52.7; 2012-2017 = 65.3; Difference = 12.6*** - LIDCs average trade openness: 1997-1999 = 81.2; 2012-2017 = 96.0; Difference = 14.8*** - Comparative points: WAEMU ~65 percent in 2012-17; LIDCs ~96 percent in 2012-17. WAEMU 1997-99 = 52 percent; LIDCs 1997-99 = 81 percent. ### Investment - Growth accelerations correlated with an increase in investment. - Total investment-to-GDP (averages and differences): - WAEMU: 1997-1999 = 15.9; 2012-2017 = 22.6; Difference = 6.7*** - LIDCs: 1997-1999 = 23.5; 2012-2017 = 24.1; Difference = 0.6 - Public investment-to-GDP: - WAEMU: 1997-1999 = 5.7; 2012-2017 = 9.1; Difference = 3.4*** - LIDCs: 1997-1999 = 5.3; 2012-2017 = 6.7; Difference = 1.4*** - Country highlights: investment increased particularly in Niger, Togo, Senegal and Burkina Faso well above WAEMU average of 6.7 percent. ### Political institutions (Polity score) - Polity Score improved significantly in the WAEMU between the two periods. - Polity score averages and differences: - WAEMU: 1997-1999 = 0.1; 2012-2017 = 4.1; Difference = 4.1*** - LIDCs: 1997-1999 = 3.2; 2012-2017 = 4.6; Difference = 1.4*** ### Terms of trade - Terms of trade change: - WAEMU average change in terms of trade (2012-17) = 1.2 percent; LIDCs (2012-17) = -0.3 percent. - The terms of trade improved between 1997-99 and 2012-17 in the WAEMU; improvements particularly in Guinea-Bissau, Burkina Faso, Mali, and Côte d’Ivoire. ### Productivity gains (TFP) - Growth accounting using Penn World Table (1960-2014) to decompose growth into TFP, physical capital, and labor. - TFP contribution (points) and differences: - WAEMU TFP: 1997-1999 = 0.57; 2012-2017 = 1.03; Difference = 0.46 - LIDCs TFP: 1997-1999 = 1.03; 2012-2017 = -0.45; Difference = -1.48*** - Country-level TFP increases: Togo, Côte d'Ivoire, Niger, and Guinea-Bissau reported notable TFP contribution increases. ### From benchmarking to regressions — methodology - Dependent variable: dummy = 1 in year of growth acceleration, 0 otherwise. - Growth acceleration definition (modified Hausmann, Pritchett, Rodrik (2005)): 1. Real GDP growth rate is at least 2 percent; 2. Post-acceleration growth rate is at least 3.5 percent per year, and last for at least 5 years; 3. Level of post-acceleration output must exceed the pre-episode peak level of income. - Differences from Hausmann et al. (2005): post-acceleration window shortened to five years (vs seven); acceleration defined using real GDP growth (not real GDP per capita). - Empirical approach: Probit model (baseline), with independent variables lagged one year and year-fixed effects; robustness checks with Logit and Linear Probability Models. ### Episodes of rapid growth (1960-2017) — WAEMU members - Summary from identification exercise: - Côte d’Ivoire: 2 episodes (1964, 2012) - Burkina Faso: 2 episodes (2001, 2009) - Togo: 2 episodes (1964, 2008) - Benin: 1 episode (1994) - Niger: 1 episode (2011) - Senegal: 1 episode (2012) - Guinea-Bissau and Mali: 0 episodes ### Empirical results — determinants of growth accelerations (Table 8, Probit estimates, column (1)) - Statistically significant positive determinants (coefficients and robust standard errors in parentheses): - Investment-to-GDP t-1: 0.0096*** (0.0035) - Trade openness-to-GDP t-1: 0.0043** (0.0019) - Exchange rate depreciation t-1: 0.0364* (0.0188) - Polity score t-1: 0.0229*** (0.0051) - Terms of trade t-1: 0.0049* (0.0030) - Total factor productivity t-1: 0.0303*** (0.0071) - Statistically significant negative determinants: - Inflation rate t-1: -0.0956** (0.0447) - Exchange rate volatility t-1: -0.0437*** (0.0123) - Insignificant (Probit column (1)): - Inflation volatility t-1: -0.0342 (0.0380) - Model statistics: - Observations = 1,850 - Number of countries = 60 - R-squared = 0.051 ### Robustness and overall findings - Baseline Probit results are robust to Logit and Linear Probability specifications (columns 2 and 3); signs and statistical significance of key variables persist. - Preliminary scatter-plot evidence: growth accelerations positively correlated with exchange rate depreciation, trade openness, investment-to-GDP, polity score, favorable terms of trade, and TFP; negatively correlated with inflation rate, inflation volatility, and exchange rate volatility. ### Concluding remarks (summary) - Compared to the pre-acceleration period, WAEMU’s recent better performance coincides with: - an increase in macroeconomic stability; - an increase in investment (total and public); - improvement in institutional quality (Polity score); - improved terms of trade; and - higher TFP contribution to growth. - Empirical analysis conclusion: investment, trade openness, exchange rate depreciation, political institutions, terms of trade, and TFP are key favorable drivers of growth accelerations in LIDCs; higher inflation rate and greater exchange rate volatility lower the likelihood of experiencing a growth acceleration. *International Monetary Fund — WEST AFRICAN ECONOMIC AND MONETARY UNION (excerpt).* ### Annex I. List of Countries Included in the Analysis (LIDCs) ### Annex I. List of Countries Included in the Analysis (LIDCs) ### List of Countries Included in the Analysis (LIDCs) - Afghanistan - Bangladesh - Benin - Bhutan - Bolivia - Burkina Faso - Burundi - Cambodia - Cameroon - Central African Republic - Chad - Comoros - Congo, Democratic Republic of - Congo, Republic of - Côte d'Ivoire - Djibouti - Eritrea - Ethiopia - Gambia, The - Ghana - Guinea - Guinea-Bissau - Haiti - Honduras - Kenya - Kiribati - Kyrgyz Republic - Lao People's Democratic Republic - Lesotho - Liberia - Madagascar - Malawi - Mali - Mauritania - Moldova - Mongolia - Mozambique - Myanmar - Nepal - Niger - Nigeria - Papua New Guinea - Rwanda - São Tomé and Príncipe - Senegal - Sierra Leone - Solomon Islands - Somalia - South Sudan - Sudan - Tajikistan - Tanzania - Togo - Uganda - Uzbekistan - Vietnam - Yemen, Republic of - Zambia - Zimbabwe - WEST AFRICAN ECONOMIC AND MONETARY UNION ### List of Variables Used, Definitions and Sources - NGDP_RPCH — Gross domestic product, constant prices in national currency, percent change. Source: World Economic Outlook. - NGDP_R_PPP — Gross domestic product, constant prices, in PPP international dollars. - NGDP_R — Gross domestic product, constant prices in national currency. - PCPI_PCH — Consumer Prices, period average, percent change. - ENDA — National currency units per U.S. dollar, period average. - TX_GDP — Value of exports of goods and services, U.S. dollars, percent of GDP. Source: World Economic Outlook. - NID_GDP — Gross capital formation, current prices in U.S. dollars, percent of GDP. Source: World Economic Outlook. - TT_PCH — Terms of trade of goods and services, U.S. dollars, percent change. Source: World Economic Outlook. - polity2 — Combined polity score of autocracy/democracy. The unified polity scale ranges from +10 (strongly democratic) to -10 (strongly autocratic). Source: Polity IV Project: Center for Systemic Peace. - rgdpna — Real GDP at constant 2011 national prices (in millions US$). - ck — Capital stock at current PPPs (in millions 2011 US$). - emp — Number of persons engaged (in millions). - labsh — Share of labour compensation in GDP at current national prices. - Table and figure attributions: - Figure 1 — World Economic Outlook. - Figures 2, 3, 4 — World Economic Outlook; Penn World Table version 9.0 where indicated. - Tables 1–6 — World Economic Outlook; Penn World Table version 9.0 where indicated. - Notes: - The unified polity scale explanation is repeated: ranges from +10 (strongly democratic) to -10 (strongly autocratic). - Some variable definitions (rgdpna, ck, emp, labsh) repeat across tables and figures as noted in the source. ### Selected WAEMU Banking System Findings and Vulnerabilities (excerpted) - Banking system size and access: - Broad money grew by about 12 percent per year on average over 2012-17. - Gross assets of commercial banks reached 56 percent of the regional GDP at end-2016 as compared with 43 percent at end-2013. - Number of clients’ accounts increased from 7.8 million to 10.3 million over the same period. - Estimated share of households with a bank account rose from about 7 percent in 2007 to 16 percent in 2016. - Institutions and ownership (2017): - Total credit institutions: 140 (125 banks including 17 bank branches; 15 non-bank credit institutions including 4 branches). - By ownership (excluding branches): 94 foreign-owned, 12 public, 13 privately held regional credit institutions. - By size: 31 large (accounting for more than 10 percent of banking sector assets in country), 25 average (more than 5 percent of assets), 63 relatively small. - Capital and solvency (selected measures and trends): - Average capital adequacy ratio (CAR) declined to 11.4 percent at mid-2017 from 12.6 percent at end-2015. - As of end-2016, 12 banks — accounting for nearly 9 percent of the system’s assets — did not meet the minimum CAR of 8 percent (on a Basel I basis). - Up to 29 banks and credit institutions did not meet the new minimum capital requirement of FCFA 10 billion effective at mid-2017 and were given until mid-2018 to comply. - Capital shortfall of undercapitalized WAEMU banks equaled about 0.4 percent of the zone’s GDP, of which about 2/3 came from the capital shortfall in public banks. - Concentration and nonperforming loans: - At end-September 2017, 12 leading bank groups accounted for about ¾ of the market share, total assets, and clients’ accounts. - Ratio of loans to the top-5 borrowers in total loans: 90 percent at mid-2017 (down from 101 percent at mid-2016). - Gross NPLs to total loans: 13.8 percent at end-2016; 14.6 percent at mid-2017; 15.1 percent at end-September. - Provisions cover about 63 percent of NPLs. Net NPLs to total loans: 5.9 percent at mid-2017; 6.2 percent at end-September. - Net NPLs to capital: 47.2 percent (2016 Dec.) and 49.7 percent (2017 Jun.) as per Table 1 entries. - Public banks: - Out of 12 public banks in WAEMU, 7 were undercapitalized; some had significantly negative capital. - One public bank was classified as large and another as medium-sized in their countries of operations. - Liquidity and central bank interaction: - Net aggregate liquidity of the banking system has been negative since 2014 and reached about FCFA 3,000 billion at end-2016 (about 5 percent of the zone’s GDP). - BCEAO limited banks’ access to the refinancing facility to twice their own capital effective from June 2017. - BCEAO lowered the obligatory reserve requirement ratio from 5 percent to 3 percent in March 2017. - BCEAO actions resulted in the reduction of the banking system’s net liquidity position by about FCFA 600 billion over the first half of 2017. - Banking system’s net liquidity deficit covered by the BCEAO refinancing came to around FCFA 2,900 billion at end-2017. - In 2015-16, government debt issuances were typically oversubscribed; subscription dipped to about 66 percent in February 2017 and later fell to 41 percent in December 2017. - Stress test outcomes and vulnerability scenarios: - Stress tests modeled shocks including deterioration of loan portfolio, default of the largest borrower, and deposit run. - In the scenario of the largest borrower default in each bank: - Number of undercapitalized banks would increase to 76 (from 15 before the shocks). - Combined capital shortfall would be 1.6 percent of the zone’s GDP. - In the scenario of default by the two largest borrowers: - 91 banks would fall below the minimum capital requirement of 8 percent (further capital shortfall implied in the source). ### Policy-relevant observations and priorities (as stated) - Enforce compliance with prudential norms. - Recapitalize or resolve problem banks. - Address pockets of vulnerabilities in the banking system, notably: - High exposure to a few largest borrowers. - Weaknesses in public banks. - Structural liquidity deficits and reliance on BCEAO refinancing. *Source: cr18107 - Annex I. List of Countries Included in the Analysis (LIDCs).* ### 2.7 percent of the zone’s GDP. ### 2.7 percent of the zone’s GDP. ### Stress test outcomes and banking-sector risks - Deterioration of banks’ loan portfolios is a major source of risk: the number of undercapitalized banks would rise sharply under extreme scenarios. - Stress test summary (Data as of June 30, 2017) — Credit Portfolio Tests (Table 2): - Before stress tests: Average CAR 11.42; Undercapitalized banks (CAR < 8 percent) 15; Capital shortfall, FCFA billion 268.0; Of which: Public banks undercapitalized banks 7; Average CAR -1.88. - Largest borrower default: Average CAR 3.11; Undercapitalized banks 76; Capital shortfall, FCFA billion 995.9; Of which: Public banks undercapitalized banks 9; Average CAR -10.79. - Two largest borrowers default: Average CAR -2.57; Undercapitalized banks 91; Capital shortfall, FCFA billion 1,667.4; Of which: Public banks undercapitalized banks 10; Average CAR -17.16. - NPLs increase by 50 percent: Average CAR 6.06; Undercapitalized banks 57; Capital shortfall, FCFA billion 684.0; Of which: Public banks undercapitalized banks 9; Average CAR -16.51. - NPLs increase by 75 percent: Average CAR 3.13; Undercapitalized banks 67; Capital shortfall, FCFA billion 1,042.8; Of which: Public banks undercapitalized banks 12; Average CAR -25.52. - 15 percent of standard loans become NPLs: Average CAR 5.14; Undercapitalized banks 62; Capital shortfall, FCFA billion 790.3; Of which: Public banks undercapitalized banks 11; Average CAR -13.26. - Liquidity stress (Table 3) — Deposit run over five days (demand deposits outflows: 5 percent, 5 percent, 5 percent, 10 percent, and 10 percent; term deposits outflows: 3 percent, 3 percent, 3 percent, 5 percent and 5 percent): - Banks cope with liquidity shortages: Day 1: 20; Day 2: 26; Day 3: 33; Day 4: 56; Day 5: 79. - In the case of a sustained deposit run with progressive withdrawal of up to 35 percent of demand deposits and 19 percent of term deposits over 5 days, up to 79 banks would need liquidity support. - Public banks, which exhibited significant weaknesses before the tests, are strongly affected by the stress test scenarios. ### Macro-financial linkages and sectoral exposures - Bank intermediation and stocks (end-2017): - Household deposits: about 11.3 percent of regional GDP. - Credits to households: 6.6 percent of GDP. - Deposits from private companies: about 15.4 percent of GDP. - Credits to private companies: about 18.7 percent of GDP — resulting in net claims on private non-financial companies of about 3.4 percent of GDP. - Credits to public non-financial companies: 2.1 percent of GDP; deposits from them: 1.7 percent of GDP. - Commercial banks’ net claims on the central governments: 10.2 percent of GDP. - Net liquidity support from BCEAO: about 4.7 percent of GDP (provided largely under collateral of government debt securities). - Composition of bank credit to the economy (June 2017, Percent): - Retail and wholesale trade, restaurants and hotels: 31.4 - Transportation and communication: 11.9 - Other services: 13.3 - Insurance, real estate and services: 7.3 - Construction: 11.7 - Manufacturing: 12.7 - Agriculture: 4.7 - Extractive industries: 1.3 - Electricity, water and gas: 5.6 - Services sector concentration: nearly 1/3 of outstanding bank credit in retail and wholesale trade, restaurants and hotels; banking exposure to services sectors accounted for nearly 2/3 of all bank credit to the economy. Productive sectors (agriculture, manufacturing, electricity, water and gas, extractive industries and construction) accounted for the remaining 1/3. ### Evolution of macro-financial flows in 2016–17 - 2016 developments: - Consolidated central government’s fiscal deficit: about 4½ percent of GDP. - Regional banks held about 90 percent of the outstanding government debt securities. - BCEAO injected liquidity equivalent to about 2.7 percent of GDP in the banking system through its refinancing window. - Banks extended net credit to government by 3.6 percent of GDP in 2016. - Official international reserves declined significantly during 2016. - 2017 developments and policy tightening: - BCEAO’s December 2016 tightening limited refinancing that banks could get from BCEAO to twice their own capital; banks were given until June 2017 to comply. - Commercial banks reduced net credit from the central bank by about 0.2 percent of GDP in 2017. - BCEAO was able to replenish its FX reserves in 2017. - Banks provided new net credit to governments of only 0.8 percent of GDP in 2017 (compared with 3.6 percent of GDP in 2016). - Domestic credit dynamics: - Pace of domestic credit expansion decelerated from 15-20 percent prevailing in 2016 and previous years to less than 10 percent at end-2017. - Contribution of net credit to government to domestic credit growth declined to 2 percentage points in 2017 (from 11.2 percentage points in 2016). - Contribution of credit to the economy remained steady at about 8 percentage points at end-2017 (about the same as at end-2016). ### Addressing banking-sector vulnerabilities — regulatory and supervisory measures - Enforcing compliance with prudential regulations is key: - Supervisor’s sanctions range from warnings, fines, limits on certain operations, provisional administration, to withdrawing banking licenses. - Supervisor can prohibit weak banks from expanding business (providing new loans and accepting new deposits) and distributing profits until identified weaknesses are resolved. - New prudential rules aligned with Basel II/III phased in from 2018: - Reforms include gradually increasing minimum capital requirements, introducing new accounting rules, and moving to consolidated supervision of bank groups. - Initially, the CAR will increase from 8.0 percent (under Basel I rules) to 8.6 percent of risk-weighted assets by end-2018. - For problem loans, lowering the threshold for classifying loans as NPLs from 180 days of payment delay to the internationally accepted 90 days would increase the reported NPLs. - Supervision reforms: - Banking Commission adopting new supervision criteria aligned with Pillar 2 of Basel II/III standards. - Consolidated supervision of cross-border groups has been launched and tools readied. - Bank resolution powers: - Banking Commission gained new bank resolution powers in late 2017 through an amendment to its Statutes, ensuring member countries commit to cooperate based on its independent resolution decision-making — expected to accelerate resolution of problem banks with persistent negative capital. ### Stress test scenarios (Annex I description) - Scenarios modelled using bank-by-bank data as of end-June 2017. Scenario parameters (Table 4): - Credit risk — Increase in non-performing loans (NPLs): - A: Increase by 50 percent — additional NPLs assumed to be 70 percent provisioned on average; impacts own funds and risk-weighted assets. - A: Increase by 75 percent — additional NPLs assumed to be 70 percent provisioned on average; impacts own funds and risk-weighted assets. - Credit concentration risk: - C: Default of the largest borrower — migration of all loans to him to doubtful or disputed categories; additional NPLs assumed to be 70 percent provisioned on average. - C: Default of the two largest borrowers — migration of all loans to them to doubtful or disputed categories; additional NPLs assumed to be 70 percent provisioned on average. - Deterioration of the credit portfolio: - B: 15 percent of standard loans become unpaid or immobilized; outstanding unpaid or immobilized loans require provisioning of 50 percent and doubtful or disputed loans require a 100 percent provisioning level. The 15 percent migration leads to additional NPLs assumed to be 20 percent provisioned, while originally outstanding unpaid or immobilized loans and doubtful or disputed loans require 50 percent or 100 percent provisioning, respectively. - Liquidity risk — Deposit flight for five days: - Deposit run: demand deposits outflows: 5 percent, 5 percent, 5 percent, 10 percent, and 10 percent; term deposits outflows: 3 percent, 3 percent, 3 percent, 5 percent and 5 percent. - Liquid assets realization rate estimated at 90 percent (i.e., 10 percent discount); availability rate of illiquid assets estimated at 1 percent. ### Fiscal risks context (summary) - Fiscal risks in WAEMU are mainly related to macro-economic shocks and contingent liabilities linked to state-owned enterprises and public-private partnerships, compounded by weak public financial management systems. - Recent research cited: - Public finances are typically hit by a macroeconomic shock (sharp decline in nominal GDP growth) once every 12 years, with an average fiscal cost of 9 percent of GDP. - A dataset of 80 advanced and emerging market countries identifies 174 episodes of materialization of contingent liabilities, with an average fiscal cost of 6 percent of GDP. - Sources of fiscal risks listed (Table 1) include macro-economic shocks (Real GDP growth, Primary balance, Terms-of-trade, Nominal exchange rate, Interest rate, Aid), contingent liabilities (State-owned enterprises, Large infrastructure projects, PPPs, Pensions, Financial system, Natural disasters), and institutional factors (quality/timeliness/coverage of fiscal data, expenditure control, oversight of local governments, revenue forecasting and administration, risks related to natural resource management). *Source: IMF staff estimates and BCEAO data as presented in the chapter.* ### 6. Developing a comprehensive fiscal risk management framework can help ### 6. Developing a comprehensive fiscal risk management framework can help ### Fiscal risk management framework (reference: IMF, 2016) - Key objective: governments must identify and quantify risks before they can be managed. - Table 2 — Fiscal Risk Management Framework (steps preserved as in source): - Step 1: Identify and quantify risks - Identify sources of risk - Calculate fiscal exposure - Estimate likelihood of realization - Step 2: Decide whether to mitigate risks - Direct controls and limits on exposure - Regulation, incentives and other indirect measures - Transfer and risk sharing mechanisms - Step 3: Decide whether to provision for funds - Expense fiscal costs in the budget - Budget contingencies - set aside financial assets in buffer funds ### Need for forward-looking strategies and relevance to WAEMU - Measuring past realized fiscal risks gives a sense of potential fiscal costs, but cannot substitute for forward-looking strategies because the nature and size of fiscal risks changes rapidly. - Example: PPP-related risks in the WAEMU context have been negligible in the past, but could become more important going forward as recent development plans outline an increased reliance on PPP financing. - A fiscal risk toolkit translates the three-step framework into concrete policy options for contingent liabilities. ### Fiscal reporting and GFSM 2014 guidance - GFSM 2014 is a holistic system covering all general government transactions and incorporates the general government's balance sheet to link stocks and flows. - GFSM 2014 recommends recording stocks and flows on an accrual basis, while maintaining cash-flow data for liquidity assessment. - Contingent liabilities are typically not reported as liabilities until called; GFSM 2014 advises reporting contingent liabilities as memorandum items and drawing up a Summary Statement of Explicit Contingent Liabilities and Net Implicit Obligations for Future Social Security Benefits. - Full use of GFSM 2014 remains rare: balance sheet data was only reported in 15 percent of IMF staff reports in early 2017. ### Monitoring, tracking, and reporting fiscal risks - Proper monitoring, tracking, and reporting when risks materialize would provide a fuller picture of their fiscal costs. - Incomplete accounting can improve headline fiscal numbers without underlying structural improvements; signs include stock-flow discrepancies where debt increases much faster than implied by fiscal deficits. - Typical causes of such discrepancies: - Deficiencies in public financial management (PFM) systems and incomplete accounting techniques that keep operations outside the headline budget deficit but require financing. - Materialization of fiscal risks not well captured in the central government budget, adding to the debt burden. - These issues often occur together: realized fiscal risks lead to larger deficits, which may then be masked via creative accounting. ### Recent WAEMU fiscal developments and importance of fiscal risks - WAEMU debt developments: - Between 2012 and 2017 the gross debt-to-GDP ratio increased by 17 percentage points of GDP. - Region-wide growth rates have been more than 6 percent every year since 2012. - WAEMU Debt-to-GDP convergence criterion = 70% - Fiscal deficits: - Most member states have diverged from the deficit convergence criterion of 3 percent of GDP in recent years. - Decomposition of debt increase attributes changes to: primary balance, exchange rate, interest/growth differential, and a residual (stock-flow discrepancy) that includes privatization receipts, recognition of contingent liabilities, debt relief, bank recapitalizations, etc. - Across the eight WAEMU countries, the residual accounted for 5.7 percentage points of GDP of the 16.9 percentage points of GDP increase in gross debt between 2013 and 2017. - Unanticipated increases in debt (proxy for materialized fiscal risks): - Difference between 2017 debt-to-GDP projections and 2017 projections made in 2012: 15 percentage points of GDP (unweighted average), with peaks of more than 30 percent in Benin and Togo. - For all WAEMU countries, excluding Guinea-Bissau, the unanticipated share of the total increase was larger than 65 percent of the total increase. - Main reasons: larger-than-expected spending as new development plans got under way (Côte d’Ivoire, Senegal), shocks to commodity prices (cocoa in Côte d’Ivoire, fuel imports elsewhere), reconstruction after conflict and security-related spending (Côte d’Ivoire, Mali), pre-financing arrangements (Togo, Benin), calling of contingent liabilities, and PFM issues. ### Main sources of fiscal risks in WAEMU - Primary categories identified: - Macro-economic shocks (covered in staff reports and DSAs; play a significant role). - Contingent liabilities: - SOEs (State-Owned Enterprises) stand out among contingent liabilities. - PPPs have potential to become more important in the near future. - Pensions: not an immediate issue due to relatively young populations. - Financial system risks: potentially large, but to be dealt with at WAEMU regional level. - Natural disasters: discussed elsewhere. - Institutional risks related to fiscal management and PFM quality (coverage, timeliness, expenditure allocation and control, revenue forecasting and administration). - Note: It is not immediately possible to quantify the relative importance of various sources due to correlation between risks and data limitations. ### Fiscal risks from SOEs (detailed) - SOEs are an important source of fiscal risk; SOE liabilities are often explicitly or implicitly guaranteed by the state. - SOEs frequently carry out quasi-fiscal operations that worsen their financial position relative to a commercial profit-maximizing level. - For advanced economies and emerging markets, the average cost of government intervention in SOEs over 1990-2014 amounted to about 3 percent of GDP (cited as context). - In WAEMU, fiscal risks related to SOEs in the energy sector are important: - Quasi-fiscal deficits in the electricity sector averaged around 1.3 percent of GDP across 7 WAEMU countries in 2013-2014. - Even with much lower oil prices than in the 2013-14 reference years, quasi-fiscal deficits would still average almost 1 percent of GDP (lined-bars hypothetical). - Improving reporting and coverage of SOE fiscal transactions is crucial (equivalent to Step 1 in Table 2). This includes: - Properly classifying SOEs and clearly defining what constitutes an SOE. - GFSM 2014 distinction between public corporations (market producers) and other SOEs (noncommercial government entities) should guide classification, though practical difficulties exist. ### Classification guidance (GFSM 2014 / Box 2 excerpt) - Definition and classification of public corporations: - Corporations are entities capable of generating a profit or other financial gain, recognized by law as separate legal entities, and set up for market production. - A public corporation is a corporation controlled by the government; control interpreted broadly as the ability to determine general corporate policy. - Entities controlled by government should be classified as public corporations only if they are market producers. - Market test: a unit is a market producer if the value of its sales averages at least half of production costs over a period of at least three years. - Practical valuation difficulties can arise in sectors like utilities and nuclear energy where prices are difficult to determine or not strongly linked with production costs. *Source: cr18107 - 6. Developing a comprehensive fiscal risk management framework can help* ### 17. WAEMU fiscal indicators refer mostly to central government. Collection of data ### 17. WAEMU fiscal indicators refer mostly to central government. Collection of data ### SOE data collection and monitoring - The GFSM 2014 concept of general government does not include public enterprises, but does include “extra-budgetary units”. - WAEMU countries are mainly focused on improving the availability of SOE data; several countries face challenges with SOE monitoring, as evidenced by low scores on the relevant PEFA indicator. - Recent country developments and technical assistance: - Benin: authorities are collecting and analyzing the financial statements and questionnaire responses of all 22 SOEs and 140 autonomous agencies to assess their indebtedness. - Côte d’Ivoire: a database has been put together with data on debt servicing from 23 public enterprises. Aggregate debt figures for these SOEs are being reported. - Guinea-Bissau: comprehensive audits, covering both financial and management aspects, for 8 SOEs completed in 2017. - Senegal: as part of the process to qualify for the IMF’s Special Data Dissemination Standard (SDDS) which demands general government coverage, authorities have started to produce data related to SOEs in TOFE format. Senegal also reports some summary statistics related to public enterprises and semi-autonomies agencies (like universities and hospitals) in its annual budget framework paper. - Togo: SOE debt is routinely reported as part of wider debt statistics and included in debt sustainability analyses. - Institutional incentives and IMF role: - The IMF is playing a prominent role in standard setting and program-related incentives. As of February 2018, most disbursing IMF programs in WAEMU countries had some structural benchmark related to SOEs. - Countries like Côte d’Ivoire have benefited from IMF technical assistance on SOE-related matters. ### Contingent liabilities: Public Private Partnerships (PPPs) - Definition and fiscal risk: - PPPs refer to long-term contracts between a private party and a government entity for providing a public asset or service, in which the private party bears risks and management responsibility and remuneration is linked to performance. - PPPs allow for deferred spending compared with standard public procurement and can create debt-like or contingent obligations for the government; frequently these obligations are not included in the government’s fiscal aggregates. - Main PPP risks: delays in construction, litigation, increases in financing costs, and optimistic implementation assumptions. - WAEMU context: - PPPs are not widespread in WAEMU despite large infrastructure needs and limited fiscal space; development plans often aim to make use of PPPs. - Legal frameworks and institutional capacity are being updated across the region; countries are updating legal texts and setting up institutions (e.g., dedicated PPP units within the Ministry of Finance) to monitor PPP fiscal risks. - International institutions provide technical assistance and program-related benchmarks to ensure PPPs do not jeopardize debt sustainability. - At the WAEMU level, work is ongoing on a region-wide directive on PPPs. - Example: Côte d’Ivoire plans to include all PPP projects in its 2018-20 multiyear investment plan. - Table 4 (summary of PPP frameworks): - Benin: New law adopted in 2016; PIMA IMF mission, Oct-17; Yes—structural benchmark: Setting up Strategic Support Unit responsible i.a. for PPP evaluation (done in June 2017). - Burkina Faso: Revision of PPP law ongoing; PIMA IMF mission, Mar-17; PPP IMF mission, Nov-17; A potential new ECF arrangement is likely to feature PPP-related Structural Benchmarks. - Côte d’Ivoire: New law adopted in 2012; PPP IMF mission, January 2017; Yes—Develop PPP database (set for December 2017). - Guinea-Bissau: New law adopted in 2009; None; No. - Mali: New law adopted in 2016; PIMA IMF mission, June 2017; No. - Niger: PPP law adopted in 2011, being amended in 2018; PPP IMF mission, April 2017; Yes—New PPP law consistent with the investment code and the 2012 budget law, is to be sent to National Assembly. - Senegal: New law adopted in 2014; Peer-learning event organized by IMF, April 2017; Yes, but only in current PSI first and second review. - Togo: No dedicated PPP law; None; No. ### PFM: Quality and coverage of reported fiscal data - Reporting basis and coverage: - Most WAEMU countries still use elements of the cash-based GFSM 1986 system and a mixed reporting system where some items are reported on a cash basis while most spending items are recorded at the commitment (‘ordonnancement’) stage. - Coverage is mostly limited to central government, implying that information related to SOEs and local governments is often not part of standard budgetary presentation, creating the risk that fiscal pressures can develop outside decision-makers’ visibility. - PEFA findings: - On the PEFA indicators ‘Comprehensiveness of information included in the budget’ and ‘Transparency of intergovernmental fiscal relations’, WAEMU countries like Burkina Faso, Côte d’Ivoire and Mali score better than the global average. - Most WAEMU countries for which data are available score poorly on ‘extent of unreported government operations’, indicating fiscal risks linked to incompleteness of data coverage. - Expenditure control: more than 50 percent of the PEFA data points show the lowest levels (‘D’ or ‘D+’) for expenditure control indicators; particular weaknesses include ‘stock and monitoring of arrears’ and ‘effectiveness of payroll controls’. - Revenue forecasting: measured by PEFA, forecasting is reasonably accurate in WAEMU, with more than half of countries above the global average and only 1 country-system deemed inadequate (‘D’ score). - Revenue administration: more than half of WAEMU countries’ systems assessed as inadequate across the two relevant PEFA indicators. ### Detecting fiscal risks outside the fiscal balance: stock-flow exercise - Non-structural fiscal adjustments: - Statistical treatments and accounting choices can mask underlying fiscal challenges; terms used include “non-structural adjustments”, “fiscal gimmicks”, or “fiscal devices”. - Such practices can temporarily improve fiscal indicators while leaving government net worth unchanged or worsened later. - The use of fiscal gimmicks has been documented in countries with fiscal rules; deficit rules are more subject to rapid adjustments and thus to gimmicks than debt rules. - Measurement via stock-flow residuals: - A proxy for non-structural fiscal adjustment is the size of stock-flow discrepancies: when the debt stock increases faster than implied by fiscal deficits the residual is larger than zero. - Calculating the ‘residual + other identified debt-creating flows’ variable for WAEMU countries: the (unweighted) annual average for all eight countries over 2013-2016 is 1.3 percent of GDP. - Residuals have tended to increase over the period, peaking at 1.6 percent in 2016. - Analysis of positive stock-flow residuals — identified sources: - Table 6 summary (sources identified by country; checkmark implies presence or relevance): - Benin: Pre-financing; Coverage; Roll-over of appropriations. - Burkina Faso: Pre-financing; Coverage. - Côte d’Ivoire: Coverage. - Guinea-Bissau: Pre-financing; Coverage. - Mali: Pre-financing. - Niger: Not applicable as stock-flow discrepancies small. - Senegal: Pre-financing; Coverage; Pensions; Arrears; Data issues. - Togo: Pre-financing; Coverage; Pensions; Arrears; Data issues. - Country-specific explanations: - Pre-financing: - Typical pre-financing: a private company obtains a loan from domestic commercial banks for public works; the Finance Ministry guarantees and agrees to service principal and interest from the Treasury account, making full liability rest with the government. - Benin: approved road projects worth 24 percent of GDP were approved without corresponding budgetary allocations through pre-financing schemes; some project execution totaled about 3 percent of GDP in 2016, adding to the debt stock. - Togo: pre-financed expenditures averaged 3.5 percent of GDP during 2013-16. - Arrears: - Arrears build-up followed by audits and clearance plans add to gross debt but do not add to the deficit. - Côte d’Ivoire: government bonds were issued in 2014 to pay off past domestic arrears incurred prior to 2011 (worth 1 percent of GDP); these were not captured in debt numbers until verified in 2014. - Changes in government deposits and reporting practices: - Withdrawals from government bank deposits are included in domestic bank financing but are not reflected in the debt stock (Côte d’Ivoire and Togo). - Data reconciliation issues: - In Burkina Faso and Guinea-Bissau, reconciliation between budgetary data and financing data remains a work in progress. - Note on interpretation: - The identity underlying the stock-flow residual can be affected by valuation changes, cash versus accrual accounting, privatization, below-the-line treatment of past liabilities (including arrears clearance/accumulation), and other country-specific factors. Large residuals may indicate muddled fiscal positions even when individually explainable. *International Monetary Fund — WEST AFRICAN ECONOMIC AND MONETARY UNION (chapter excerpt).* ### 33. Debt stocks are reported on a cash basis, while the fiscal balance in WAEMU ### 33. Debt stocks are reported on a cash basis, while the fiscal balance in WAEMU countries is reported on a commitment basis ### Stock-flow discrepancies: nature and causes - Debt stocks are reported on a cash basis, while the fiscal balance in WAEMU countries is reported on a commitment basis. - In practice, most WAEMU countries still report certain items on a cash basis, but most spending items are reported on a commitment basis, and this can create non-zero stock-flow residuals. - Main factors explaining debt stocks growing much faster than implied by fiscal deficits (stock-flow discrepancies): - Pre-financing; - SOE-related fiscal transactions; - Multi-year investment projects; - Coverage and data issues, including arrears. - The stock-flow discrepancy captures the increase in public debt that cannot be explained by the fiscal deficit (adjusted for exchange rate valuations and GDP growth). ### Senegal: specific sources of discrepancies and improved transparency - Sources of stock-flow discrepancies in Senegal are mainly related to: - La Poste (an example of fiscal operations difficult to classify); - Pensions (civil service pensions have recently gone into deficit, i.e. payments have been larger than contributions – on a flow basis); - Roll-over/financing of unutilized appropriations from past budgets (comptes de depôt). - Senegal’s so-called treasury operations contributed substantially to the government’s gross financing needs. - Thanks to improved data availability and transparency, a large part of Senegal’s 2015-16 stock-flow discrepancy can be explained. - Technical note: once arrears are identified, they should be accounted for in the year where they were supposed to have been committed; however, previous fiscal accounts are often not changed and settlement of arrears is mostly accounted for through below-the-line operations in current and future years. ### IMF programs and conditionality addressing stock-flow residuals - IMF programs are playing a key role in improving aspects related to positive stock-flow residuals. - Program examples and measures: - Benin and Togo: IMF programs have set conditionality related to pre-financing. - Senegal (PSI): program conditionality to help reduce sources of stock-flow discrepancies described above. - Côte d’Ivoire: program includes conditionality on SOEs and PPP risks. - Most programs include incentives to improve coverage and transactions of SOEs. - Overall starting point for reducing positive stock-flow residuals: improving data coverage and transparency. ### Improving Fiscal Risk Management and PFM priorities - A PFM system can be a source of fiscal risks, but when functioning properly, can reduce other types of fiscal risks. - Non-exhaustive list of PFM issues important for fiscal risk management: - Moving to accrual accounting and audited financial statements at the government level to enhance disclosure of fiscal risks and contingent liabilities; PPPs would appear on the balance sheet. - Improving fiscal projections and medium-term fiscal frameworks, expenditure allocation and controls. - Developing the stocks aspect of the fiscal table, notably through reporting opening and closing public sector balance sheets. - Moving to 2001-2014 GFS table, and broadening the scope of the fiscal table (at least to the general government level). ### WAEMU regional PFM directives and implementation - WAEMU developed six regionally agreed directives reflecting international standards to improve PFM: 1. Directive dealing with PFM Transparency Code; 2. Directive dealing with the State’s Finance Laws; 3. Directive dealing with the general regulations regarding the State’s Public Accounting; 4. Directive dealing with the State’s Budget Nomenclature; 5. Directive dealing with the State’s Public Accounting Plan; 6. Directive dealing with the State’s Table of Financial Transactions (TOFE). - By the end of 2012, all 6 directives had been incorporated into member countries’ legislative texts. - Implementation will require capacity upgrades and technical assistance from development partners. - The application of PFM principles is an on-going task and is not linear (PEFA indicators can worsen over time in certain countries). ### Oversight of SOEs and management of PPP fiscal risks (Box 3) - SOEs: short-term priorities to build a good oversight regime include: - Develop an inventory of SOEs; - Establish basic reporting framework for SOEs that present fiscal risks; - Clarify the role of other public entities (Ministry of Finance, line ministries, Presidency, etc.) in SOE oversight. - PPPs: main elements of fiscal risk management for PPPs (IMF, 2016, p.38): - Ensure central control over PPP approvals, ceilings on overall fiscal exposure, and appropriate gatekeeping role for the government; - Systematically subject projects to appraisal, undertake independent review of project feasibility and charge risk-related guarantee fees where they are provided; - Ensure the risk-allocation framework allocates risks to the parties best able to influence or control the risk, and introduce upper bounds where payments are linked to demand; - Ensure that the full lifetime costs and the potential fiscal exposure are accurately identified and budgeted for during the decision-making process. - Tools and initiatives noted: WAEMU PPP directive, Public Fiscal Risk Assessment Model (P-FRAM), IMF fiscal risk management framework. ### Role of international organizations and IMF instruments - International organizations, including the IMF, play a key role in setting and disseminating fiscal transparency standards and helping countries apply them. - The IMF’s roles cited: - Standard-setting (Fiscal Transparency Code); - Technical Assistance delivery (Fiscal Transparency Evaluation); - Engagement through ECF (disbursing) and PSI (non-disbursing) programs in WAEMU countries; - Revamped debt sustainability framework to help assess contingent liabilities and fiscal risks, including broadening debt coverage (IMF, 2018b, p.33). ### Conclusions and policy implications - Fiscal risks in the WAEMU region are mainly related to macro-economic shocks and contingent liabilities linked to state-owned enterprises and public-private partnerships. - Weak public financial management systems often contribute to the size of the problem once fiscal risks materialize. - Progress noted: - Analysis and monitoring of fiscal risks is improving in most WAEMU countries. - Information availability regarding SOEs is improving. - Substantial progress on frameworks governing PPPs. - Côte d’Ivoire completed a stock-taking report on fiscal risks. - Senegal improved public finance statistics and adhered to SDDS. - Remaining needs and recommendations: - Assess overall risks to fiscal frameworks in most countries. - Use the Fiscal Transparency Code as the guiding framework to improve reporting of fiscal risks. - Prioritize ongoing reform programs related to fiscal transparency and fiscal risk management among regional decision-makers. - Improve PFM systems further, including reducing discrepancies between debt stock dynamics and annual deficits, to help improve fiscal management and attract private investment. *Source: IMF staff report excerpt (chapter/section as provided).* ### 3. Revenue Administration in WAEMU ### 3. Revenue Administration in WAEMU ### PI-3. Revenue outturn; PI-19. Revenue administration; PI-20. Accounting for revenue - PI-3. Revenue outturn - Description: This indicator measures the change in revenue between the original approved budget and end-of-year outturn. It contains two dimensions and uses the M2 (AV) method for aggregating dimension scores. - PI-19. Revenue administration - Description: This indicator relates to the entities that administer central government revenues, which may include tax administration, customs administration, and social security contribution administration. It also covers agencies administering revenues from other significant sources such as natural resources extraction. These may include public corporations that operate as regulators and holding companies for government interests. In such cases the assessment will require information to be collected from entities outside the government sector. The indicator assesses the procedures used to collect and monitor central government revenues. It contains four dimensions and uses M2 (AV) method for aggregating dimension scores. - PI-20. Accounting for revenue - Description: This indicator assesses procedures for recording and reporting revenue collections, consolidating revenues collected, and reconciling tax revenue accounts. It covers both tax and nontax revenues collected by the central government. This indicator contains three dimensions and uses M1 (WL) for aggregating dimension scores. ### Public investment efficiency: overview and methodology - Purpose - The chapter assesses the infrastructure gap in WAEMU and empirically evaluates public investment efficiency using efficiency frontier analysis and panel regressions. - Methodology - Data envelopment analysis (DEA) output-oriented model, following IMF (2015), Grigoli and Kapsoli (2013), and Albino‑War and others (2014). - Period: 2000-2015. - Inputs: - Real public capital stock per capita. - Per capita GDP (proxy for private sector contribution to infrastructure services). - Outputs: - Physical indicator combining length of road network, electricity production, access to water, number of secondary teachers, and hospital beds. - Hybrid indicator combining physical and survey-based indicators into a synthetic index of coverage and quality of infrastructure networks. ### Key findings on WAEMU public investment efficiency - Infrastructure gap and quality - WAEMU has substantial infrastructure needs, particularly in electricity supply, paved road density and telecommunication infrastructure. - WAEMU’s infrastructure level and quality remain relatively low despite high public investment over the past decade. - World Economic Forum Global Competitiveness Indicators rank WAEMU behind the sub-Saharan African average and regional comparator groups; quality of electricity supply, railroads and roads scored below regional comparator groups. - Efficiency scores (Table 1; scale as estimated by DEA) - Sub-Saharan Africa: Physical Infrastructure 0.4597; Hybrid Indicator 0.6417. - CEMAC: Physical Infrastructure 0.3046; Hybrid Indicator 0.5108. - EAC: Physical Infrastructure 0.4875; Hybrid Indicator 0.7351. - WAEMU: Physical Infrastructure 0.3694; Hybrid Indicator 0.6188. - Oil exporters: Physical Infrastructure 0.1958; Hybrid Indicator 0.2687. - Non-resource-intensive: Physical Infrastructure 0.4464; Hybrid Indicator 0.6981. - Other resource-intensive: Physical Infrastructure 0.6019; Hybrid Indicator 0.6563. - Interpretation - WAEMU’s performance lags all comparator groups; magnitude of inefficiency depends on chosen output index. - Relationship between per capita public capital stock and perceptions of infrastructure quality is positive but weak. - Correlation between real GDP growth and investment has been weaker in WAEMU than in regional comparator groups. ### Determinants of public investment efficiency: empirical results - Regression setup - Sample: 154 countries including 45 SSA countries; period 2000-2015. - Dependent variable: Efficiency scores (hybrid). - Explanatory variables: quality of institutions (Control of Corruption and Regulatory Quality), Official Development Assistance (ODA), percentage of urban population, natural resource dependency (dummy for LIC or LMIC rich in non-renewable natural resources). - Main regression results (Table 2; t-statistics in parentheses) - Column 1 (Control of Corruption): - Control of Corruption: 0.11292 (3.13) — positive and statistically significant. - ODA: -0.00000 (-0.225). - Proportion of Urban Population: -0.00002 (-0.021). - Natural Resource dependency: -0.12216 (-2.129) — negative and statistically significant. - Constant: 0.7533 (11.065). - Observations: 93; R-squared: 0.216. - Column 2 (Regulatory Quality): - Regulatory Quality: 0.11552 (3.130) — positive and statistically significant. - ODA: -0.00000 (-0.696). - Proportion of Urban Population: -0.00045 (-0.416). - Natural Resource dependency: -0.11708 (-2.030) — negative and statistically significant. - Constant: 0.77387 (11.043). - Observations: 93; R-squared: 0.220. - Significance notation: *** p<0.01, ** p<0.05, * p<0.1. - Robustness and alternative estimation (Table 3; Tobit model) - Column 1 (Control of Corruption): - Control of Corruption: 0.11783 (4.588). - ODA: 0.00000 (1.443). - Proportion of Urban Population: -0.00146 (-2.066). - Natural Resource Receipts: -0.02733 (-0.748). - Constant: 0.89319 (19.280). - Observations: 103; R-squared: 0.224. - Column 2 (Regulatory Quality): - Regulatory Quality: 0.10568 (4.148). - ODA: 0.00000 (0.934). - Proportion of Urban Population: -0.00159 (-2.180). - Natural Resource Receipts: -0.02801 (-0.750). - Constant: 0.88955 (18.940). - Observations: 103; R-squared: 0.206. - Interpretation - Quality of institutions (Control of Corruption and Regulatory Quality) is positively and robustly correlated with public investment efficiency. - Natural resource dependency is negatively associated with public investment efficiency. - ODA and proportion of urban population show mixed or negligible associations depending on specification. - Different indicators of institutional quality are highly correlated; results corroborated using alternative measures including CPIA, with caution due to perception-based measures and data limitations. ### Policy implications and recommendations - Institutional improvements matter - A 10 percent increase in the control of corruption index or the regulatory quality index could improve public investment efficiency in WAEMU by about 15 percent on average (based on various specifications). - Focus areas to reduce the efficiency gap - Strengthen public investment management (PIM) institutions to raise returns on existing public capital stock and close the efficiency gap with comparator regions. - Address governance weaknesses—specifically control of corruption and regulatory quality—to lift investment efficiency. - Manage natural resource dependency to mitigate its negative association with investment efficiency. *Prepared from “3. Revenue Administration in WAEMU” (cr18107).* ### 11. Fund staff has developed a new Public Investment Management Assessment ### 11. Fund staff has developed a new Public Investment Management Assessment (PIMA) ### Overview of PIMA - PIMA evaluates 15 key institutions for planning, allocation, and implementation of public investment. - The 15 PIM institutions are a subset of the broader framework of “budget institutions” that govern the public financial management process (Figure 6). - PIMA missions were conducted between October 2015–June 2017. - Initial PIMA results cover 23 piloted countries. ### The 15 PIM institutions assessed - Planning - 1. Fiscal Rules - 2. National & Sectoral Planning - 3. Central-Local Coordination - 4. Management of PPPs - 5. Company Regulation - Allocating - 6. Multi-year Budgeting - 7. Budget Comprehensiveness - 8. Budget Unity - 9. Project Appraisal - 10. Project Selection - Implementing - 11. Protection of Investment - 12. Availability of Funding - 13. Transparency of Execution - 14. Project Management - 15. Monitoring of Assets ### WAEMU pilot results and comparators - WAEMU countries for which the PIMA has been conducted show: - Slightly better regulatory framework in: - National and sectoral planning - Budget comprehensiveness and protection of investment - Weaker regulations on average in: - Central-local coordination - Management of PPPs - Company regulation - Monitoring of assets - Comparison notes: - Figure 7 shows average scores for regulatory framework for WAEMU and comparators. - Figure 8 compares PIMA scores for WAEMU on regulatory framework and effectiveness; in several areas regulations exist but are not used effectively to achieve public investment efficiency. - Example: procurement law requires competitive and transparent tendering in principle, but in practice most projects are not tendered competitively. ### Sample composition (as presented in source) - WAEMU sample consists of Burkina Faso, Côte D’Ivoire, and Togo. - The SSA sample consists of Cameroon, Ghana, Liberia, Madagascar, Mozambique, and Zambia (developing SSA), and Botswana and Mauritius (emerging SSA). - Other countries in the piloted sample include Thailand, Maldives, Timor Leste and Mongolia (emerging and developing Asia), Kosovo, Serbia, Ukraine and Albania and Kyrgyz (emerging and developing Europe). ### Identified weaknesses and implementation gaps - Regulations that exist but are not effectively used in: - Multiyear budgeting - Budget comprehensiveness and unity - Protection of investment - Availability of funding - Transparency of execution, including procurement - Specific practical shortfall: - Procurement law’s competitive tendering requirement is often not applied in practice. ### Potential improvements and policy priorities - WAEMU countries could improve public investment efficiency by strengthening: - Planning and selection of PPPs - Effectiveness of central-local coordination - Credibility of multiyear budgeting - Project appraisal and selection - Monitoring of projects during implementation - Registration and monitoring of infrastructure assets ### Empirical conclusions - Regression results reported in the source indicate: - Public investment efficiency is positively correlated with the quality of institution. - Public investment efficiency is negatively correlated with dependency on natural resource. - There is scope for institutional improvements to deliver greater returns on planned capital investments. - WAEMU countries exhibit relatively good regulatory framework in: - National and sectoral planning - Multiyear budgeting - Project management - Nonetheless, efficiency gains are possible by addressing the weaknesses listed above. ### Data and methodological notes (Annex I & II highlights) - Key variables used in analysis include: - General government capital stock, per capita (2011 PPP$-adjusted, per capita) - Gross domestic product, current prices, in U.S. dollars, per capita (US$ dollars per capita) - Public gross fixed capital formation, current prices, percent of GDP (Percent of GDP) - Quality of overall infrastructure (Scale (1 - 7), 7 = best) - Mobile cellular subscriptions (Percent of population) - Access to electricity (Percent of population) - Physical Infrastructure Index (Units (100 = AE average)) - Physical Public Investment Efficiency Indicator (PIE-X) (Scale (0 - 1), 1 = highest levels of infrastructure coverage and quality) - Hybrid Public Investment Efficiency Indicator (PIE-X) (Scale (0 - 1), 1 = highest levels of infrastructure coverage and quality) - CPIA macroeconomic management (Rating (1=low to 6=high)) - Control of Corruption (Estimate) - Regulatory Quality (Estimate) - LICs and LMICs Rich in Non-Renewable Natural Resources (Dummy, 1=Rich or 0=Non-Rich) - Urban population (Percent of population) - Net official development assistance received (US Dollars) - DEA methodology: - Efficiency measured relative to a frontier using Data Envelopment Analysis (DEA) with variable return to scale. - Efficiency scores indicate the proportional amount by which countries could raise the quality of their infrastructure while leaving public capital (and other inputs) unchanged. - Two infrastructure output indicators used: - A physical indicator: quantitative index capturing use of public investment, standardized and combined. - A hybrid indicator: the mean of the physical and survey-based indicators to measure both coverage and quality. *Source: cr18107 - 11. Fund staff has developed a new Public Investment Management Assessment (PIMA), PIMA missions (October 2015–June 2017).* --- _Source: https://www.imf.org/-/media/files/publications/cr/2018/cr18107.pdf_