## Devarajan–O'Connell final report on the MRLIC project (2012–end of calendar 2024)

## Source details

**Canonical URL:** [Devarajan–O'Connell final report on the MRLIC project (2012–end of calendar 2024)](https://www.imf.org/-/media/files/topics/lics/macro-research-for-development/related-links/devarajan-oconnell-final-report-on-mrlic-project-final.pdf)

## Other formats

- [Markdown version](/-/media/files/topics/lics/macro-research-for-development/related-links/devarajan-oconnell-final-report-on-mrlic-project-final.pdf.md)
- [Structured JSON version](/-/media/files/topics/lics/macro-research-for-development/related-links/devarajan-oconnell-final-report-on-mrlic-project-final.pdf.json)

---

### 1. Introduction — overview and purpose
- In 2012, the United Kingdom’s Department for International Development (DFID) began a strategic partnership with the International Monetary Fund (IMF) to strengthen research on macroeconomic issues affecting low-income countries (LICs).
- The Macroeconomic Research in Low Income Countries (MRLIC) project aimed to improve macroeconomic policies to increase the pace of economic growth, job creation and poverty reduction in LICs.
- The final external review covers the period from MRLIC’s inception in 2012 to the end of calendar 2024.
- Purpose of the review:
  - Assess overall effectiveness and impact.
  - Extract lessons for future macroeconomic research programming by the Foreign, Commonwealth and Development Office (FCDO) and the IMF.
- Theory of change and scope:
  - Funding → enhanced research and capacity building → improved IMF policy engagement → stronger policies and policy institutions in LICs.
  - The review focuses on links from funding to research and capacity building and onward to IMF policy engagement; it does not perform an end-to-end assessment of MRLIC’s impact on growth, job creation and poverty reduction.
- Methodology:
  - Statistical analysis of project outputs and interviews with researchers, present and former senior IMF and FCDO officials, and policymakers in low-income countries.
  - Annual reports documented MRLIC research and training outputs and generated consistent DFID/FCDO ratings of A or above (A+ and A++).

### 2. Context: the IMF and low-income countries — membership, mission, and resource use
- LIC representation and composition:
  - LICs represent about 40 percent of the IMF’s membership but, because of their small economic size, do not represent systemic risk to the global economy.
  - The IMF’s low-income category refers to PRGT-eligible countries (Poverty Reduction and Growth Trust) eligible to borrow on concessional terms.
  - PRGT eligibility is determined by:
    - Incomes below the IDA operational cutoff used for concessional World Bank resources, or below twice that cutoff if population < 1 million; or
    - Lack of access to global financial markets.
  - Group sizes:
    - In 2012: 72 (PRGT eligible) > 60 (LIDC) > 35 (World Bank low income) (total IMF membership 188).
    - In 2024: 70 (PRGT eligible) > 58 (LIDC) > 26 (World Bank low income) (total IMF membership 191).
  - Regional/composition notes:
    - 53 percent located in Sub-Saharan Africa (SSA) in 2012.
    - Composition shifted toward SSA and fragile and conflict affected states (FCS).
  - Specific country movements (2012–2024):
    - Graduations: Armenia, Bolivia, Georgia, Guyana, Nigeria and Vietnam.
    - New PRGT-eligible entries: Marshall Islands, Federated States of Micronesia, Syrian Arab Republic and Tuvalu.
- IMF mission in LICs:
  - The Fund’s LIC mandate incorporates development dimensions—promotion of policies for economic diversification, inclusive growth and poverty reduction—in addition to macroeconomic stability and external adjustment.
  - The IMF developed concessional lending facilities and technical assistance tailored to LIC needs; historical multilateral debt relief participation (HIPC, Enhanced HIPC, MDRI) and new lending facilities for natural disasters and environmental resilience are noted.
  - Development trajectory: strong LIC performance mid-1990s; slowdown after 2014; reversal since COVID with per-capita incomes in fully half of LIC economies diverging from wealthy countries (Chrimes et al. 2024).
  - Fund engagement intensified since 2020, including concessional lending and coordination in external debt restructurings.
- Use of Fund resources by LICs (empirical 2000–2024 evidence, LIC-2012 group):
  - The LIC-2012 group initiated 5.3 new lending programs per member on average over 2000-2024.
  - Non-PRGT-eligible countries initiated 1.9 new programs per member over the same period.
  - Fewer than 10 percent of the PRGT-2012 group made no use at all of IMF lending facilities during 2000-2024, compared with nearly half of non-PRGT-eligible members.
  - Of 27 non-lending conditionality programs initiated since 2005, more than 80 percent (22) were undertaken by PRGT-2012 group countries.
  - The LIC-2012 group made intensive use of the Fund’s free-standing technical assistance and capacity development resources.
- PRGT share in IMF credits and loans outstanding:
  - PRGT loans averaged 13.5 percent of IMF credits over the referenced period despite LICs accounting for 63.5 percent of IMF programs in Table 1.
  - PRGT credit outstanding fell sharply in real terms with MDRI debt relief in 2005 and remained below pre-2005 levels as lending to advanced and emerging-market economies surged.
  - As of 2011, PRGT loans constituted less than 6 percent of the Fund’s outstanding portfolio.
  - Aggregate ratio notes: for PRGT-eligible countries, the ratio of outstanding non-PRGT credits to total outstanding IMF claims averaged 16.4 percent over 2010-2024; calculated as an average across PRGT-eligible countries it was 10.1 percent.

### 3. Alignment of priorities, structures and processes
- Strategic alignment and genesis:
  - MRLIC aligned FCDO, IMF, and LIC government priorities initially on monetary and fiscal policy and later on diversification, natural resources, gender and inequality.
  - The project was intentionally aligned with the Fund’s own LIC research agenda; IMFs’ priorities were “kept in the lead.”
- Institutional structure and accountability:
  - MRLIC operated as a cross-departmental collaboration between the Research Department (RES) and the Strategy, Policy and Review Department (SPR), led by two co-directors.
  - Project phases: four 3-year phases based on MRLIC research proposals agreed by both institutions.
  - FCDO oversight introduced annual quantitative output targets and detailed performance reviews, leveraging FCDO’s evidence-based project cycle.
- Comparative advantage and impediments:
  - IMF’s comparative advantage: producing macroeconomic research on LICs.
  - Impediments: poor aggregate data, limited systemic relevance, weak domestic research systems, and internal priorities favoring emerging-market and advanced-country research.
- Thematic focus and evolution:
  - Core focus areas: public investment and debt sustainability; fiscal management of primary commodity revenues; monetary policy frameworks including exchange-rate management; structural determinants of diversification and growth.
  - Early incorporation of inequality work and later inclusion of gender, natural disasters and climate resilience in line with evolving LIC priorities.
- Mechanisms amplifying impact:
  - Co-authorships, Visiting Scholars, conversion of MRLIC contractuals to Fund staff, and staff movements broadened internal uptake and application.

### 4. Research outputs — volume, quality, dissemination, and value for money
- Summary assessment:
  - MRLIC generated substantial LIC macroeconomic research, arrested a decline in the share of IMF Working Papers on LICs, and produced high-quality outputs by citation indices; dissemination through capacity development and social media was substantial.
- Research volume and trajectories (aggregated fiscal-year averages and totals by period 2013-16, 2017-20, 2021-24; totals all years):
  - Working papers: averages per fiscal year — 14; 18.5; 15.5. Total, all years — 192.
  - Published papers: averages per fiscal year — 6.3; 13; 9. Total, all years — 113.
  - Books: averages per fiscal year — 0; 2; 1. Total, all years — 3.
  - Toolkits: averages per fiscal year — 3; 7; 6. Total, all years — 16.
  - Courses (the first was in FY2017): averages per fiscal year — 0; 6.7; 8. Total, all years — 52.
  - Instances of uptake by country authorities and IMF teams: averages per fiscal year — 14.3; 26.0; 15.5. Total, all years — 223.
  - High-level policy conferences attended by senior IMF staff: averages per fiscal year — 3.3; 6.8; 4. Total, all years — 56.
  - Research reflected in IMF policy papers, such as Board papers, Staff Discussion Notes, and memos to management: averages per fiscal year — 3.3; 7; 5.3. Total, all years — 62.
  - Thematic areas of IMF policy influenced and made LIC-specific: averages per fiscal year — 1.3; 1.8; 2.3. Total, all years — 21.
  - Trajectory: peak conventional research productivity occurred in 2017-20; every category shows higher output in 2021-24 than in 2013-16.
- Citations and social media:
  - Lens.org citation counts (as of May 5, 2025):
    - MRLIC papers published in regularized form were cited 5,969 times cumulatively.
    - Self-citations by IMF authors excluded: 529.
    - 86 percent of MRLIC papers have been cited at least once.
    - 37 percent of MRLIC papers were cited more frequently than the average of papers published over the same time period in ‘applied economics’ and ‘econometrics’ per Field Citation Ratio by Dimensions.ai.
    - Eleven papers achieved FCR rankings 10 times the average.
    - Only 35 percent of MRLIC publications have reached the age of eight years; social-science research typically takes approximately eight years to reach maximum citations.
  - Policy-institution citations (Overton.io):
    - MRLIC research was cited in 2,786 publications mainly working papers issued by policy institutions.
    - Nine of the citing central banks are located in LICs.
    - Citing organizations composition: 71.5% international governmental organizations; 13.1% government agencies; 15.4% think tanks, legislative bodies, and non-governmental organizations.
  - Social media mentions (Altmetric Explorer):
    - 1,568 social media mentions of MRLIC papers.
    - 130 separate papers were mentioned, primarily on the X platform.
    - Mentions span 78 countries including 16 LICs.
    - Overall sentiment: 58 percent classified as positive or strongly positive.
    - Example press impact: Furceri and others (2019) paper on macroeconomic consequences of tariffs featured in 63 recent news stories from 53 media outlets including CNN, MSN, Yahoo!, and CNBC.
- FCDO funding and value for money:
  - Project budget allocation (MRLIC Annual Report 2024): 70.1 percent of the project budget devoted to producing research papers and capacity-development tools; remainder to country applications and indirect costs.
  - Estimated research-staffing cost per IMF Working Paper: $72,417 per IMF Working Paper (research-staffing cost).
  - Including the remaining 29.9 percent of the project budget, the overall budgetary cost per IMF Working Paper was roughly half the total cost per paper incurred by similar donor-financed research arrangements in a comparable development institution.
  - MRLIC accounted for between 20 and 35 percent of the spending of the Development Macroeconomics Division (RESDM) in recent years.
- MRLIC contribution to IMF Working Papers series:
  - MRLIC produced Working Papers at a pace of over 1 a month, constituting nearly 5 percent of the total flow of Working Papers produced at the Fund since 2012.
  - MRLIC contributions correspond to 13.3 Working Papers contributed per year by the project.
  - Table 2a summary (series totals and proportions):
    - 2000-2011 All: 3,175 published; Proportion with LIC content (%) = 264.6; Proportion with Macro content (%) = 25.4; Total Annual average = 61.4.
    - 2012-2024 All: 3,640 published; Proportion with LIC content (%) = 280; Proportion with Macro content (%) = 25.3; Total Annual average = 59.5.
    - MRLIC project: 173 published; Proportion with LIC content (%) = 13.3; Proportion with Macro content (%) = 73.0; Total Annual average = 71.8.
    - Non-MRLIC: 3,467 published; Proportion with LIC content (%) = 266.7; Proportion with Macro content (%) = 23.0; Total Annual average = 58.9.
  - Counterfactual trend analysis suggests MRLIC had a strongly additional impact on attention devoted to LIC concerns relative to a plausible 2010/2011 counterfactual.
- Capacity building:
  - Institute for Capacity Development (ICD) created in 2012; MRLIC contributed ICD courses starting in 2017.
  - Instances of uptake by country authorities and IMF teams totaled 223 across the project period.
  - Research is reflected in IMF policy papers, Board papers, Staff Discussion Notes, and memos to management (total instances = 62).
  - Thematic policy influence instances = 21.

### 5. How MRLIC research affected IMF policies and practices — channels and country uptake
- Summary uptake and channels:
  - Substantial uptake in IMF policy advice and program conditions on:
    - alternative monetary policy frameworks (“inflation targeting lite”),
    - debt-investment-growth strategies,
    - diversification,
    - natural resource management.
  - Uptake channels: policy papers, MRLIC researchers on missions, contractuals joining Fund staff, extensive capacity development.
- Quantitative country-level uptake:
  - Project records list 122 IMF country reports (Article IV reports and project documents) that explicitly incorporate MRLIC research over the twelve-year period.
  - At least three IMF policy documents distilled lessons: two IMF Policy Papers (2014 and 2015) and Gurara et al. (2019).
  - Uptake by country group:
    - 46 separate countries recorded at least one uptake.
    - 29 countries in SSA, including 10 fragile states.
    - 9 non-SSA small island states.
    - 8 other PRGT-eligible countries (including Myanmar and Yemen, also fragile states).
  - Uptake intensity:
    - SSA countries average 3.1 uptakes each.
    - small island states average 2.1 uptakes each.
    - SSA average is nearly twice the average for the ‘other PRGT-eligible’ subset.
- How working papers influenced policy:
  - Working papers translated into policy principally when distilled into synthesis or policy documents (e.g., 2014, 2015 policy papers; Gurara et al. 2019).
  - FPAS adaptations and DIG family models contributed to emerging consensus on inflation targeting in LICs; most LICs with inflation-targeting avoided high inflation during the 2022 global surge.
  - Working papers informed central-bank researchers in LICs leading to country adoption requests (example: one country adopted full-fledged inflation-targeting and maintained low inflation post-pandemic).
  - Dissemination gap: limited direct familiarity of some senior African policymakers with MRLIC country studies and limited bibliographical overlap with leading African policy researchers.
- Participation of MRLIC researchers on missions and country teams:
  - Largest operational uptake occurred when MRLIC researchers joined country teams and missions applying FPAS and DIG models in real time, conducting training, and facilitating dialogue.
  - Illustrative impacts:
    - Ghana: DIG analysis showed more ambitious investment could be sustained with modest debt given gradual user-fee increases and investment efficiency constraints.
    - Senegal: model application designed a program focusing on energy-sector modernization while reducing unproductive subsidies.
  - MRLIC presence broadened policy dialogue beyond stabilization toward investment prioritization and distributional implications.
- MRLIC contractuals, visiting scholars, and staff impacts:
  - MRLIC enabled LIC divisions to hire visiting scholars and contractual staff; about one-fifth of researchers brought in under MRLIC joined regular IMF staff.
  - Contractuals introduced methods (heterogeneous-agent models for inequality and gender) and embedded skills as they advanced into senior roles.
- Capacity building and training:
  - FPAS and DIG+ material represented in ICD curricula, Africa Training Institute in Mauritius, and regional CD centers including AFRITAC-East in Tanzania.
  - Observed gaps: less activity at AFRITAC-West 2 (Accra).
  - Recommended collaborations with AERC and MEFMI to strengthen regional research and demand for CD services.
- Four primary impact channels synthesized:
  - Working papers synthesized into policy papers and staff guidance.
  - Participation of MRLIC researchers in country teams and missions.
  - Hiring of MRLIC contractuals and visiting scholars (about one-fifth joined IMF staff).
  - Capacity building via ICD and regional centers embedding FPAS and DIG+.

### 6. How specific MRLIC projects affected IMF policy and country policies — monetary policy, DIG, and thematic extensions
- Monetary policy frameworks (FPAS and variants):
  - FPAS and modernized monetary frameworks took root in a number of LICs including Uganda, Tanzania, and Ghana.
  - All three countries have adopted, or are in the process of adopting, inflation targeting and more flexible exchange rate regimes.
  - MRLIC produced working papers, two Board papers (2014 and 2015), and supported missions and workshops aiding adoption of “IT-lite”.
  - Empirical notes:
    - Flexible exchange rate regimes performed better than attempts to keep exchange rates stable (IMF 2021).
    - Inflation targeting regimes fared a little better than reserve-money targeting counterparts, though sample sizes are small.
    - Morazumi et al. (2020) found inflation targeting did not reduce inflation in LICs by as much as in emerging markets.
    - A recent dataset of inflation targeting countries has only one low-income country, Kenya.
  - Implementation challenges:
    - Historical Fund expertise centered on reserve-money programming; friction with MCM department emerged.
    - Need for greater attention to monetary-fiscal interactions and deeper analysis of multiple-anchor regimes.
    - Central bankers expressed concern about AI implications for central bank research, operations, and financial stability.
- Fiscal policy: Debt-Investment-Growth (DIG) family:
  - DIG framework shows non-concessional foreign borrowing can be superior to domestic borrowing when financing productive, efficiently implemented public investment and when inflation anchor maintained.
  - DIG collapses detailed debt profiles into a single interest rate compared with LIC-DSF’s creditor-specific detail.
  - MRLIC applied DIG and DIGNAR models to over 65 countries; notable country applications in Ghana and Senegal.
  - Limits and critiques:
    - DIG not systematically adopted by the Fiscal Affairs Department (FAD); perceived as delivering a message already understood by practitioners.
    - DIG does not address mechanisms to ensure productive and efficiently implemented investments.
    - Key data problems:
      - Hidden or misreported debt undermined modeled trajectories (examples: Mozambique’s secret borrowing ~ $2 billion (12% of GDP); Senegal’s debt-to-GDP revision from 74 percent to 111 percent).
      - Buffie et al. (2012) baseline assumed steady-state public capital stock; deterioration in public capital from inadequate recurrent spending may over-estimate marginal product of public capital.
    - Risk modeling shortcomings:
      - DIG uses perfect-foresight deterministic assumptions; shock modeling limited to an introduced commodity price shock with subsequent full anticipation.
      - Lack of stochastic or scenario-based extensions limits usefulness under “exceptionally high uncertainty.”
    - Optimism bias concerns: DIG may have provided excessive flexibility; several DIG-applied countries deviated from modeled trajectories; evidence on fiscal multipliers is mixed.
  - Forward uses and recommendations:
    - Run DIG models in reverse to model fiscal consolidation and debt restructuring.
    - Retrospective empirical assessment of DIG and LIC-DSF performance is high priority.
- Diversification, inequality, gender, and climate change:
  - Diversification: MRLIC developed a database on diversification in LICs; maintaining and refreshing it is critical.
  - Inequality and gender: DIG extensions with heterogeneous agents allow anticipating distributional impacts and designing compensating measures.
  - Climate change: DIGNAD extensions address adaptation and mitigation; relevance is high for small island economies but questions remain about DIG’s ability to handle extreme uncertainty in climate impacts.
- Key recommendations from section 6:
  - Strengthen data transparency and integration (debt statistics; public capital stock).
  - Incorporate risk and uncertainty: undertake retrospective assessments; explore stochastic or scenario-based extensions.
  - Incorporate political economy and implementation realism in model applications.
  - Broaden analytical focus to fiscal consolidation, debt restructuring, and continued work on diversification, inequality, gender, and climate change.

### 7. Going forward — overall assessment, priorities, gaps, and sustainability
- Overall assessment and signature impacts:
  - MRLIC produced 192 working papers.
  - MRLIC research was used in 122 IMF country reports and several Board papers.
  - Signature themes—more flexible monetary policy frameworks and dynamically consistent debt-investment-growth strategies—are routinely applied in many LICs.
  - Recent extensions incorporate diversification, gender, inequality and climate change.
- Institutional benefits and capacity building:
  - MRLIC enhanced the profile of LIC research at the IMF by providing analytical frameworks and datasets (e.g., diversification, inflation targeting).
  - Active dissemination through training courses, toolkits, and dialogue supported demand and supply of LIC research.
  - Strategic leadership by co-directors Andrew Berg and Catherine Pattillo and institutional mechanisms supported sustained research.
  - Collaboration between SPR and RES facilitated translation of research to policy; housing in two Research Department divisions may enable greater collaboration with FAD and MCM.
  - Proposed joint RES–FAD project to estimate fiscal multipliers using AI tools for about 24 LICs could create a new database and tools and produce a Staff Guidance Note.
- Value for money and costs:
  - Total project cost: $26 million over 12 years.
  - IMF estimate of cost per working paper: $72,000 per working paper.
- Institutionalization and attention risks:
  - References to LICs in the World Economic Outlook have been increasing; WEO tables include aggregates for a low-income LIDC group.
  - New promotion policy: to be promoted to management level (A15), staff in functional departments must have worked in a low-income country or FCS.
  - Concern: LICs remain below the radar of global financial stability agendas and may lose attention without additional external funding.
- Priority themes for next phase:
  - Immediate priority: address the current economic crisis of low growth and high debt distress by:
    - Applying DIG+ to fiscal consolidation and debt restructuring scenarios.
    - Adapting monetary frameworks under balance-of-payments pressure.
  - Cross-cutting global issues (AI, climate) need work but MRLIC should prioritize LIC-specific crisis responses first.
- Three most important gaps for a next MRLIC phase:
  - (i) Treatment of risk in the DIG+ family of models, including in the LIC-DSA, amid “exceptionally high uncertainty” (Georgieva 2025).
  - (ii) Political economy of fiscal policy—explicit incorporation in models to reduce derailment risk.
  - (iii) Data transparency, particularly on debt—underestimated initial debt levels cause programs to go off-track.
- Conclusion:
  - The themes and gaps identified constitute a rich agenda for a next MRLIC phase.
  - Sustainability of MRLIC achievements at the Fund is uncertain without continued programmatic involvement and additional external funding.

*Source: Final external review of the Macroeconomic Research in Low Income Countries (MRLIC) project (2012–end of calendar 2024), commissioned by DFID/FCDO and the IMF.*

### 1. Introduction 1

### 1. Introduction

### Overview and purpose
- In 2012, the United Kingdom’s Department for International Development (DFID) began a strategic partnership with the International Monetary Fund (IMF) to strengthen research on macroeconomic issues affecting low-income countries (LICs).
- The Macroeconomic Research in Low Income Countries (MRLIC) project aimed to improve macroeconomic policies to increase the pace of economic growth, job creation and poverty reduction in LICs.
- This report is the first comprehensive external review of the project, covering the period from its inception in 2012 to the end of calendar 2024.
- The purpose of the review is to assess overall effectiveness and impact and to extract lessons for future macroeconomic research programming by the Foreign, Commonwealth and Development Office (FCDO) and the IMF.

### Background on UK aid and MRLIC funding environment
- British aid-to-GNI share trends cited:
  - Fell from roughly 0.4 percent to 0.25 percent between 1970 and 1999.
  - Rose during the 2000s to nearly 0.6 percent of GNI when MRLIC was launched in 2012.
  - Reached the OECD target of 0.7 percent in 2013 and remained at that level until 2021.
  - Scaled back to 0.5 percent of GNI during the COVID pandemic, accompanying a reduction in MRLIC’s annual funding during its final 3-year phase.
  - Current plans cited to decline further to 0.3 percent of GNI in 2027.

### Five evaluation dimensions (Terms of Reference)
- Strategic directions
  - Was the research agenda aligned with LIC needs and priorities and with the IMF’s evolving strategic priorities?
- Structure and processes
  - Were MRLIC’s structure, management, governance and oversight effective given IMF internal structures and FCDO oversight?
- Output quality
  - Were research outputs (papers, toolkits, datasets) of intended quality and directly relevant to IMF policy advising and policymaking in LICs?
- Outcomes and impacts
  - To what extent did outputs lead to broader impacts on IMF policy and operations regarding LICs, the profile of LIC issues among IMF senior management, macroeconomic policymaking in LICs, and academic discourse?
- Sustainability
  - How durable are MRLIC accomplishments in research, policy attention and capacity within the IMF and in external partnerships, and how likely is continued influence?

### Theory of change and scope
- The underlying theory of change runs from project funding → enhanced research and capacity building → improved IMF policy engagement → stronger policies and policy institutions in LICs.
- The review focuses on links from funding to research and capacity building and onward to IMF policy engagement; an end-to-end assessment of MRLIC’s impact on growth, job creation and poverty reduction in LICs lies outside the scope of this review.
- Methodology: statistical analysis of project outputs and interviews with researchers, present and former senior IMF and FCDO officials, and policymakers in low-income countries.
- Annual reports documented MRLIC research and training outputs and generated consistent DFID/FCDO ratings of A or above (A+ and A++).

---

### 2. Context: The IMF and low-income countries

### Summary findings
- LICs represent about 40 percent of the IMF’s membership but, because of their small economic size, do not represent systemic risk to the global economy.
- LIC growth and poverty reduction are central to the Fund’s mandate because LICs borrow from the Fund.
- There can be tension between LIC priorities and those of large emerging-market or advanced economies; MRLIC was intended to ease that tension.

### a. Definition and membership of the LIC group
- The IMF’s low-income category refers to PRGT-eligible countries (Poverty Reduction and Growth Trust), eligible to borrow on concessional terms.
- PRGT eligibility is determined by:
  - Incomes below the IDA operational cutoff used for concessional World Bank resources, or below twice that cutoff if population < 1 million; or
  - Lack of access to global financial markets.
- Countries graduate from PRGT eligibility only if they exceed these criteria persistently and are judged free from serious short-term vulnerabilities (IMF 2012).
- The PRGT-eligible group is nearly identical in practice to IDA eligibility but extends beyond the World Bank income-based low-income classification.
- In 2012, relevant group sizes (by category): 72 (PRGT eligible) > 60 (LIDC) > 35 (World Bank low income) (total IMF membership 188).
- In 2024, corresponding group sizes: 70 (PRGT eligible) > 58 (LIDC) > 26 (World Bank low income) (total IMF membership 191).
- Regional and composition notes:
  - PRGT-eligible group roughly balanced between Sub-Saharan Africa (SSA) and elsewhere; 53 percent located in SSA in 2012.
  - Substantial representation of small island states.
  - Composition has shifted toward SSA and fragile and conflict affected states (FCS) over time.
  - Net group size declined from 72 countries in 2012 to 70 in 2024 as graduations exceeded new entries or re-entries.
- Specific country movements (2012–2024):
  - Graduations: Armenia, Bolivia, Georgia, Guyana, Nigeria and Vietnam.
  - New PRGT-eligible entries: Marshall Islands, Federated States of Micronesia, Syrian Arab Republic and Tuvalu.

### b. LICs and the Fund’s mission
- In emerging-market and advanced countries, the IMF functions as international lender of last resort to limit country-level costs of financial distress and contain systemic risks.
- Most LICs are too small to pose global systemic risks and tend to have chronic balance of payments vulnerability due to structural characteristics and lack of diversification.
- The IMF’s mission in LICs incorporates a development dimension beyond macroeconomic stability and external adjustment, including promotion of policies for economic diversification, inclusive growth and poverty reduction (Boughton, 2009, Section 3; IMF 2024).
- The IMF has developed concessional lending facilities and technical assistance tailored to LIC needs, including responses to commodity price shocks, sectoral and trade reform to foster transformation, and poverty reduction objectives starting in 1999.
- Historical note: Between 1996 and 2005, the Fund participated in cancellation of a substantial portion of multilateral debts of LICs (HIPC and Enhanced HIPC; MDRI in 2005).
- New lending facilities have been introduced during the 2000s to respond to natural disasters and foster environmental resilience.
- Technical assistance and capacity development are particularly suited to strengthening institutional infrastructures for macroeconomic policy in LICs.
- Development trajectory:
  - Strong development performance in many LICs from mid-1990s, followed by slowing after the end of the global commodity boom in 2014 (World Bank study cited).
  - Since the onset of the COVID pandemic in 2020, the slowdown evolved into an outright reversal, with per-capita incomes in fully half of the LIC economies diverging from those of wealthy countries (Chrimes et al. 2024).
  - Fund engagement with LICs intensified since 2020, including use of concessional lending facilities and coordination role in external debt restructurings.

### c. Use of Fund resources by LICs
- LICs (PRGT-eligible) have exclusive access to the Fund’s PRGT lending facilities:
  - Extended Credit Facility (ECF) for medium-term support,
  - Standby Credit Facility (SCF) for short-term needs,
  - Rapid Credit Facility (RCF) for urgent balance of payments needs.
- LICs also have access to concessional funds from the Resilience and Sustainability Trust (RST), initiated in 2022 for low- and middle-income countries under fiscal pressure to address climate change, pandemic preparedness, and other structural challenges.
- Since 2005, LICs had access to the Policy Support Instrument (PSI), a non-lending program considered technical assistance but involving policy conditionality and staff resource commitments beyond normal surveillance. The Fund eliminated the PSI in October 2023; all members remain eligible for the non-lending Policy Coordination Instrument (PCI), which initiated its first program in 2017.
- Empirical usage (LIC-2012 group, 2000–2024):
  - The LIC-2012 group initiated 5.3 new lending programs per member on average over 2000-2024.
  - Non-PRGT-eligible countries initiated 1.9 new programs per member over the same period.
  - Fewer than 10 percent of the PRGT-2012 group made no use at all of IMF lending facilities during 2000-2024, compared with nearly half of non-PRGT-eligible members.
  - Of 27 non-lending conditionality programs initiated by IMF members since 2005, more than 80 percent (22) were undertaken by countries in the PRGT-2012 group.
  - The LIC-2012 group also made intensive use of the Fund’s free-standing technical assistance and capacity development resources.

- Table 1 summary (IMF membership and lending programs by PRGT eligibility in 2012):
  - Total: Number of IMF members in 2012 = 187; Number of IMF lending programs initiated 2000-2024 = 595; Average number of lending programs initiated per IMF member = 3.2; Proportion of members with at least one IMF lending program initiated 2000-2024 (%) = 66.8.
  - Not PRGT-eligible: 115 (61.5%) members; 217 (36.5%) lending programs; 1.9 average programs per member; 51.3 percent with at least one program.
  - PRGT-eligible: 72 (38.5%) members; 378 (63.5%) lending programs; 5.3 average programs per member; 91.7 percent with at least one program.
  - Notes: Table excludes Zimbabwe for lending data; includes South Sudan as PRGT eligible (became eligible in 2012). Annual data on IMF lending programs provided by the IMF cover 68 of the 72 PRGT-eligible IMF members in 2012 and 88 of the 115 non-PRGT-eligible countries in 2012.

- PRGT share in IMF credits and loans outstanding:
  - Because LIC economies are small, the share of PRGT facilities in overall IMF credits is far below their share of IMF programs: PRGT loans averaged 13.5 percent of IMF credits over the period referenced, despite LICs accounting for 63.5 percent of IMF programs in Table 1.
  - PRGT credit outstanding fell sharply in real terms with MDRI debt relief in 2005 and remained below pre-2005 levels as lending to advanced and emerging-market economies surged around the global financial crisis.
  - As of 2011, PRGT loans constituted less than 6 percent of the Fund’s outstanding portfolio.
  - Footnote context: All IMF members have access to non-concessional facilities; the 13.5 percent reported is a lower bound on the share of IMF claims originated to PRGT-eligible countries in total IMF credits and loans outstanding. For PRGT-eligible countries in each year, the aggregate ratio of outstanding non-PRGT credits to total outstanding IMF claims averaged 16.4 percent over 2010-2024; the same ratio calculated as an average across PRGT-eligible countries was 10.1 percent (difference reflects greater relative use of non-concessional borrowing by some larger PRGT economies).

### Historical and institutional context for MRLIC
- MRLIC was introduced against a background of a sharp pivot in IMF attention away from LICs measured by the volume and share of concessional lending.
- Contributors to contemporary debates, including the Fund’s resident historian, urged re-thinking the IMF’s role in LICs in light of the Millennium Development Goals, fiscal space from multilateral debt relief, and prospects of declining official assistance after financial turbulence in advanced countries.
- Quoted perspective: “The only real debate… concerns how best to strengthen the IMF so that it can serve its low-income members most effectively.” (reflection attributed to former Managing Director Michel Camdessus).

*Source: Final external review of the Macroeconomic Research in Low Income Countries (MRLIC) project (2012–end of calendar 2024), commissioned by DFID/FCDO and the IMF.*

### 3. Alignment of priorities, structures and processes

### 3. Alignment of priorities, structures and processes

### Alignment of strategic priorities
- The MRLIC project aligned the priorities of FCDO, the IMF, and LIC governments by focusing initially on two key monetary and fiscal policy issues and later expanding to diversification, natural resources, gender and inequality.
- Interviews and document review reveal an "extraordinary degree of alignment" between FCDO and the IMF, achieved through intensive consultation that "kept IMF priorities in the lead."
- Alignment with LIC priorities was indirect by design—driven by the Fund’s engagement with LICs—yet interviews with LIC officials showed "very strong alignment" on macroeconomic management where LICs were:
  - navigating fiscal policy reforms,
  - addressing infrastructure deficits,
  - managing limits on non-concessional borrowing,
  - modernizing monetary policy frameworks.
- The project was particularly well aligned where local policy needs were poorly served by external academic incentives, and where country priorities sometimes moved ahead of long-standing IMF practices.

### Project genesis and stakeholder roles
- DFID first approached IMF researchers around 2010 about external support for research on financial inclusion; the IMF initially did not pursue an innovative external funding arrangement.
- The project emerged later when DFID invited the project’s founding co-directors to develop a proposal, intentionally aligning the project with the Fund’s own LIC research agenda.
- Within the Fund, MRLIC operated as a cross-departmental collaboration between the Research Department (RES) and the Strategy, Policy and Review Department (SPR), led by the two co-directors who already collaborated on multiple LIC-related research papers.
- The project deliberately navigated the tension between producing new empirically-grounded ideas and maintaining institutional policy discipline, thereby elevating the profile of LIC research and influencing Fund strategic thinking.

### Structures, processes, and accountability
- The thematic and activity structure of each of the four 3-year phases was based on an MRLIC research proposal agreed by both institutions.
- The FCDO–IMF collaboration expanded resources for LIC-related research while adding external scrutiny through:
  - annual quantitative output targets, and
  - detailed performance reviews,
  leveraging the FCDO’s evidence-based internal project cycle.
- This external scrutiny helped mobilize internal support within the IMF for LIC research, analogous to how IMF conditionality can help authorities manage domestic opposition when reform intentions align.

### Comparative advantage and institutional impediments
- Confidential external assessments of the Research Department in 2009 and 2013, together with interviews, underscored:
  - the IMF’s comparative advantage in producing macroeconomic research on low-income countries, and
  - the internal factors that marginalize LIC-related research within the Fund.
- The same factors that make LIC research difficult externally—poor aggregate data, limited systemic relevance, and weak domestic research systems—also draw the Fund’s internal research priorities toward emerging-market and advanced countries.

### Thematic focus and evolution of MRLIC research
- Four areas of consistent focus central to the Fund’s engagement with LICs were central to MRLIC:
  - public investment and debt sustainability,
  - fiscal management of primary commodity revenues,
  - monetary policy frameworks including exchange-rate management,
  - structural determinants of diversification and growth.
- Early incorporation of inequality research built on Fund work underway before MRLIC and responded to the Fund-wide impetus provided by Managing Director Christine Lagarde (appointed in 2011).
- Gender inequality research was incorporated as gender issues were elevated across Fund work.
- Issues related to natural disasters and climate resilience were gradually incorporated across the MRLIC agenda in line with evolving LIC priorities.

### Mechanisms amplifying MRLIC impact within the Fund
- Close thematic alignment with Fund institutional priorities unlocked broader Fund resources for MRLIC researchers and strengthened synergies with other Fund research.
- Ongoing co-authorship between MRLIC researchers and other Fund staff, including area departments, facilitated application of MRLIC findings in country work.
- Dissemination channels within the Fund included:
  - presentations by Visiting Scholars,
  - conversion of MRLIC contractuals to Fund staff,
  - movements of MRLIC researchers into other parts of the Fund.

*Source: Final report of the MRLIC project evaluative chapter titled "3. Alignment of priorities, structures and processes."*

### 4. Research Outputs

### 4. Research Outputs

### Summary and overall assessment
- The MRLIC project generated a substantial volume of macroeconomic research on low-income countries (LICs) in a context where the macro content of development journals is declining and the LIC content of macroeconomic journals is low.
- The project appears to have arrested a decline in the share of IMF Working Papers on LICs.
- The quality of the research output, measured by citation indices, is high.
- Dissemination through capacity development as well as social media has been substantial.

### The IMF in the Macro/LIC research ecosystem
- Comparison of thematic content (abstract text analysis) shows the IMF Working Paper series has:
  - Extremely high LIC content when construed as a macroeconomics outlet.
  - Extremely high macroeconomics content when construed as a development-economics outlet.
- Trends in academic journals and macro journals:
  - Between the early 2000s and 2012, the macro content of leading development journals fell by nearly a third.
  - The share of LIC content in leading macro journals remained below 5 percent in 2012.
- Constraints limiting LIC representation in academic research:
  - Data availability and reliability are lower in LICs.
  - Overburdened educational systems and weak public institutions constrain local research capacity.
  - Das et al. (2013) report a research-wealth elasticity of 0.32 between a country’s per-capita GDP and economic research output on that country.
- MRLIC reinforcement of IMF positioning:
  - The LIC content of MRLIC-origin Working Papers was more than 200 percent (50 percentage points) higher than for Working Papers originating elsewhere at the Fund.
  - The macro content of MRLIC papers was nearly 13 percentage points higher than non-MRLIC IMF Working Papers.

### Research outputs — volume and categories (aggregated fiscal-year averages and totals)
- Source: MRLIC Annual Report 2024 and List of Outputs: Year 1 to 12 (Excel) on the MRLIC website.
- Notes: IMF fiscal years run from May 1 of the previous calendar year to April 30.
- Averages per fiscal year and totals by period (2013-16, 2017-20, 2021-24):
  - Working papers: averages per fiscal year — 14; 18.5; 15.5. Total, all years — 192.
  - Published papers: averages per fiscal year — 6.3; 13; 9. Total, all years — 113.
  - Books: averages per fiscal year — 0; 2; 1. Total, all years — 3.
  - Toolkits: averages per fiscal year — 3; 7; 6. Total, all years — 16.
  - Courses (the first was in FY2017): averages per fiscal year — 0; 6.7; 8. Total, all years — 52.
  - Instances of uptake by country authorities and IMF teams: averages per fiscal year — 14.3; 26.0; 15.5. Total, all years — 223.
  - High-level policy conferences attended by senior IMF staff: averages per fiscal year — 3.3; 6.8; 4. Total, all years — 56.
  - Research reflected in IMF policy papers, such as Board papers, Staff Discussion Notes, and memos to management: averages per fiscal year — 3.3; 7; 5.3. Total, all years — 62.
  - Thematic areas of IMF policy influenced and made LIC-specific: averages per fiscal year — 1.3; 1.8; 2.3. Total, all years — 21.
- Trajectory:
  - Peak conventional research productivity occurred in 2017-20 for several categories.
  - Every category shows higher output in 2021-24 than in 2013-16, indicating ongoing coherence and strong management.

### Citations and social media attention
- Citations (Lens.org; counts as of May 5, 2025):
  - MRLIC papers published in regularized form were cited 5,969 times cumulatively by other published papers.
  - Self-citations by IMF authors excluded: 529.
  - 86 percent of MRLIC papers have been cited at least once.
  - 37 percent of MRLIC papers were cited more frequently than the average of papers published over the same time period in ‘applied economics’ and ‘econometrics’, according to Field Citation Ratio (FCR) calculations by Dimensions.ai.
  - Eleven papers achieved FCR rankings 10 times the average.
  - Only 35 percent of MRLIC publications have reached the age of eight years; Galvez and Galiani (2017) note social-science research typically takes approximately eight years to reach maximum citations.
- Policy-institution citations (Overton.io):
  - MRLIC research was cited in 2,786 publications mainly working papers issued by policy institutions.
  - Nine of the citing central banks are located in LICs.
  - Among citing organizations: 71.5% are international governmental organizations, 13.1% are government agencies, and 15.4% are think tanks, legislative bodies, and non-governmental organizations.
- Social media mentions (Altmetric Explorer):
  - 1,568 social media mentions of MRLIC papers.
  - 130 separate papers were mentioned, primarily on the X platform.
  - Mentions span 78 countries including 16 LICs.
  - Overall sentiment of mentions is positive, with 58 percent classified as positive or strongly positive.
  - Example of press impact: Furceri and others (2019) paper on macroeconomic consequences of tariffs featured in 63 recent news stories from 53 media outlets including CNN, MSN, Yahoo!, and CNBC.

### FCDO funding and research value for money
- Project budget allocation (MRLIC Annual Report 2024):
  - 70.1 percent of the project budget was devoted to producing research papers and capacity-development tools; remainder to country applications and indirect costs.
- Estimated research-staffing cost per IMF Working Paper:
  - $72,417 per IMF Working Paper (research-staffing cost).
  - This figure is less than half of the typical range for comparable research projects.
- Allocation-free back-of-the-envelope calculation:
  - Including the remaining 29.9 percent of the project budget, the overall budgetary cost per IMF Working Paper was roughly half the total cost per paper incurred by similar donor-financed research arrangements in a comparable development institution.
- Project contribution to RESDM spending:
  - Discussions with IMF officials suggest MRLIC accounted for between 20 and 35 percent of the spending of the Development Macroeconomics Division (RESDM) in recent years.

### Quantifying MRLIC impacts on IMF Working Papers
- Role of IMF Working Papers:
  - Constitutes the largest single destination for IMF staff research and serves upstream roles for models, toolkits, policy advising, capacity building, and lending.
- Volume and share:
  - MRLIC produced Working Papers at a pace of over 1 a month, constituting nearly 5 percent of the total flow of Working Papers produced at the Fund since 2012.
  - MRLIC contributions correspond to 13.3 Working Papers contributed per year by the project.
- Counterfactual considerations and trends:
  - Table 2a shows the IMF Working Papers series totals and proportions:
    - 2000-2011 All: 3,175 published; Proportion with LIC content (%) = 264.6; Proportion with Macro content (%) = 25.4; Total Annual average = 61.4.
    - 2012-2024 All: 3,640 published; Proportion with LIC content (%) = 280; Proportion with Macro content (%) = 25.3; Total Annual average = 59.5.
    - MRLIC project: 173 published; Proportion with LIC content (%) = 13.3; Proportion with Macro content (%) = 73.0; Total Annual average = 71.8.
    - Non-MRLIC: 3,467 published; Proportion with LIC content (%) = 266.7; Proportion with Macro content (%) = 23.0; Total Annual average = 58.9.
  - Apparent paradox:
    - The LIC content of the overall series shows no change across the two periods (2000-2011 vs 2012-2024) based on means alone, suggesting MRLIC additions might have been offset by reductions elsewhere.
  - Trend analysis and counterfactual:
    - A sharp increase in attention to LICs occurred between 2000 and 2004, followed by a precipitous decline starting in 2005.
    - Separate linear trends estimated for 2005-11 and 2012-24 suggest a plausible counterfactual constructed around 2010 or 2011 would predict ongoing LIC content substantially below the pre-2012 average.
    - Against such a counterfactual, the MRLIC project had a strongly additional impact on attention devoted to LIC concerns in the IMF Working Papers series.

### Capacity building
- Institutional context:
  - The Institute for Capacity Development (ICD) was created in 2012 with the mandate of defining the Fund’s capacity development strategy, merging the IMF Institute with related technical assistance functions.
  - The IMF’s commitment to capacity building in LICs was long standing prior to MRLIC.
- MRLIC contributions to ICD:
  - MRLIC contributed an array of ICD courses starting in 2017, building on papers and toolkits developed earlier in the project.
  - Incorporation of MRLIC research into the Fund’s capacity building curriculum was facilitated by the move of RESDM Division Chief and project co-founder Andrew Berg to the ICD in late 2014.
- Uptake and influence:
  - Instances of uptake by country authorities and IMF teams totaled 223 across the project period.
  - Research is reflected in IMF policy papers, Board papers, Staff Discussion Notes, and memos to management (total instances across years = 62).
  - Thematic areas of IMF policy were influenced and made LIC-specific in 21 instances across the project period.

*Source: Devarajan-O'Connell final report on the MRLIC project — Section 4: Research Outputs.*

### 5.  How MRLIC research affected IMF policies and practices

### 5. How MRLIC research affected IMF policies and practices

### Summary of uptake and channels
- The MRLIC research saw substantial takeup in IMF policy advice and program conditions, especially on:
  - alternative monetary policy frameworks (“inflation targeting lite”),
  - debt-investment-growth strategies,
  - diversification,
  - natural resource management.
- Uptake was facilitated by policy papers (building on working papers), MRLIC researchers’ participation in missions, additional MRLIC researchers joining the Fund staff, and an extensive capacity development program.
- The FCDO grant increased the quantity and quality of IMF research on low-income countries during the 2012-2024 period.
- MRLIC publications contributed to macroeconomic research in major development and macroeconomic journals and were widely cited and referenced in social media.

### Quantitative evidence of country-level uptake
- The project’s records list 122 IMF country reports (Article IV reports and project documents) over the past twelve years that explicitly incorporate MRLIC research.
- At least three IMF policy documents distilled lessons from the research into guidance for IMF teams: two IMF Policy Papers (2014 and 2015) and Gurara et al. (2019).
- Uptake coverage by country group:
  - 46 separate countries recorded at least one uptake,
  - 29 countries in Sub-Saharan Africa (SSA), including 10 fragile states,
  - 9 non-SSA small island states,
  - 8 other PRGT-eligible countries (including Myanmar and Yemen, which are also fragile states).
- Uptake intensity by group:
  - SSA countries average 3.1 uptakes each,
  - small island states average 2.1 uptakes each,
  - SSA average is nearly twice the average for the ‘other PRGT-eligible’ subset.
- Timing and representation:
  - From FY2017, MRLIC research frameworks and toolkits featured in Institute for Capacity Development (ICD) courses at headquarters and regional centers.
- Public diffusion indicators cited from external literature:
  - 37 percent of World Bank policy reports have never been downloaded,
  - 87 percent were never cited (Doemeland and Trevino, 2014).

### How working papers influenced policy
- Working papers rarely translate directly into single policy prescriptions; influence typically arises when a body of work is distilled into policy or synthesis documents.
- MRLIC working papers provided the basis for IMF policy papers (2014, 2015) and staff guidance, including the 2019 DIG applications paper (Gurara et al. 2019).
- The combination of working papers and policy papers adapting the Forecasting and Policy Analysis System (FPAS) for LICs contributed to an emerging consensus on inflation targeting in LICs.
  - Result noted: most LICs with inflation-targeting avoided high inflation during the 2022 global surge.
- Working papers also indirectly influenced policy by informing central-bank researchers in LICs, who then requested IMF assistance to explore and adopt new frameworks; example: a country that adopted a full-fledged inflation-targeting regime and maintained low inflation post-pandemic.
- Identified dissemination gap: limited direct familiarity of some senior African policymakers with MRLIC studies of their own countries and limited overlap in bibliographical references between MRLIC researchers and leading African policy researchers.

### Participation of MRLIC researchers in country teams and missions
- The largest operational uptake occurred when MRLIC researchers joined country teams and missions, providing country-specific applications of FPAS and the DIG family of models.
- Roles played by researchers on missions:
  - provide real-time applications of models,
  - conduct training courses and seminars for country teams and authorities,
  - act as interlocutors between IMF operational staff and country counterparts.
- Illustrative country impacts:
  - Ghana: DIG analysis showed that with gradual increases in user fees and investment efficiency, a more ambitious investment program could be sustained with modest debt when LIC-DSA signaled limited scope for non-concessional borrowing.
  - A country with serious macro imbalances: DIGNAR model (DIG + natural resources) opened discussions on structural transformation and policies beyond stabilization.
  - Senegal: model application designed a program focusing on the energy sector, modernizing the sector while reducing unproductive subsidies and making fiscal adjustment more palatable.
- The presence of MRLIC researchers allowed policy dialogue to broaden beyond stabilization to include investment prioritization and distributional implications (e.g., financial sector reform accounting for inequality).

### MRLIC contractuals, visiting scholars, and IMF staff impacts
- MRLIC enabled LIC divisions in Research and SPR to hire visiting scholars and contractual staff, largely from universities, to pioneer analytical frameworks and translate research into policy.
- Employment transitions:
  - Rough estimates indicate about one-fifth of researchers brought in under MRLIC joined the regular IMF staff.
- Institutional capacity benefits:
  - Brought economists with expertise beyond standard Fund tools (e.g., heterogeneous-agent models incorporating inequality and gender).
  - Staff retained MRLIC-derived skills as they took on senior institutional roles (examples provided of career progression and institutional diffusion of MRLIC tools).
- Impact multiplier: visiting/contractual researchers developed models, tested them with data, engaged country teams, and those who remained at the Fund carried methods into broader work across country groups.

### Capacity building and training contributions
- MRLIC devoted significant resources to developing macro models and toolkits and to formal and informal training activities that supported uptake.
- Curriculum integration:
  - FPAS and DIG+ material are represented in ICD curricula, Africa Training Institute in Mauritius, and regional CD centers including AFRITAC-East in Tanzania.
- Gaps and opportunities:
  - Less activity observed at AFRITAC-West 2 (Accra).
  - Potential collaborations recommended with African research and training institutions such as the AERC and MEFMI to build research and policy-analytic capacity.
  - Suggestion to promote the demand side for Fund CD services in LICs to avoid countries remaining in a low-capacity, low-growth trap due to denial of capacity shortfalls.

### Identified mechanisms of impact (synthesis)
- Four primary channels through which MRLIC research affected IMF policies and practice:
  - a. Working papers that were synthesized into policy papers and staff guidance (e.g., 2014, 2015 policy papers; Gurara et al. 2019).
  - b. Participation of MRLIC researchers in country teams and missions providing applied model work and training.
  - c. Hiring of MRLIC contractuals and visiting scholars, with about one-fifth joining IMF staff, embedding expertise institutionally.
  - d. Capacity building via ICD and regional centers embedding FPAS and DIG+ in training programs.
- Constraints to fuller impact:
  - Working papers can be perceived as generic or detached from country realities; some operational managers resist adopting novel approaches without synthesized policy guidance.
  - Limited direct dissemination to and engagement with African research institutions and some senior African policymakers.

*Source: MRLIC project final report, IMF.*

### 6.  How specific MRLIC projects affected IMF policy and country policies

### 6. How specific MRLIC projects affected IMF policy and country policies

### Overall summary and purpose
- The MRLIC research catalyzed new approaches to monetary and fiscal policy in low-income countries (LICs), notably by:
  - Spurring a shift to inflation targeting lite and flexible exchange rate regimes in some LICs through central bank demand for MRLIC outputs and Fund assistance.
  - Fostering more effective dialogue between the Fund and country authorities on non-concessional borrowing for infrastructure via the debt-investment-growth (DIG) family of models.
- The MRLIC project aimed to help LICs accelerate growth while maintaining debt sustainability, an objective challenged by subsequent low growth and widespread debt distress.

### Monetary policy frameworks (FPAS and variants)
- Key outcomes:
  - The Forecasting and Policy Analysis System (FPAS) and modernized monetary policy frameworks took root in a number of LICs, including Uganda, Tanzania, and Ghana.
  - All three countries have adopted, or are in the process of adopting, inflation targeting and more flexible exchange rate regimes.
  - MRLIC produced several working papers on the subject, two Board papers (2014 and 2015), and supported missions and workshops that aided adoption of “IT-lite”.
  - Flexible exchange rate regimes performed better than attempts to keep exchange rates stable (IMF 2021). Inflation targeting regimes fared a little better than reserve-money targeting counterparts, though sample sizes are small.
  - A systematic study (Morazumi et al. 2020) found that inflation targeting did not reduce inflation in LICs by as much as it did in emerging market economies.
  - A more recent dataset of inflation targeting countries has only one low-income country, Kenya.
- Positive spillovers and capacity building:
  - IT-lite adoption enabled policy coordination (e.g., productive discussions on monetary policy integration in the East African Community among Kenya, Tanzania, and Uganda).
  - AFRITAC provides capacity-building on demand and is supporting work in Malawi; AFRITAC’s support on communications strategies was noted as especially useful.
  - Collaboration with African capacity-building organizations (MEFMI and AERC) was identified as mutually beneficial.
- Implementation challenges and institutional issues:
  - The Fund’s technical assistance expertise and program conditionality were historically built around reserve-money programming, creating friction as some LIC central banks were “more ready” for IT-lite than the Fund.
  - Resistance and institutional turf issues arose with the Monetary and Capital Markets (MCM) Department, compounded by intellectual disagreement over reliance on a short-run policy interest rate when financial markets were poorly developed.
  - With support from the front office of the African Department, MRLIC researchers helped move both MCM and countries forward; MCM’s emphasis on institutions, frameworks, and capabilities complemented the shift.
- Recommendations and outstanding concerns from central bankers:
  - Greater attention to monetary-fiscal interactions is needed given current debt-related pressures for fiscal consolidation and struggles with fiscal dominance (e.g., Ghana) or fiscal backsliding (as may be happening in Kenya).
  - Deeper analysis of multiple-anchor regimes is necessary because many LIC central banks weight the nominal exchange rate heavily; models that ignore exchange rate objectives are poorly adapted to such realities.
  - Central bankers expressed acute concern about understanding and managing implications of rapid growth of artificial intelligence (AI) for central bank research, operations, and financial stability.

### Fiscal policy: Debt-Investment-Growth (DIG) models
- Purpose and contributions:
  - The DIG family (starting with Buffie et al. (2012)) provided a consistent analytical framework to evaluate non-concessional borrowing for public investment in LICs, showing that non-concessional foreign borrowing can be superior to domestic borrowing when financing productive, efficiently implemented public investment and when the country maintains its inflation anchor.
  - The DIG models complemented the LIC-DSF by capturing the link between borrowing and growth through public investment; DIG collapses detailed debt profiles into a single interest rate while LIC-DSF retains creditor-specific details.
- Empirical application and uptake:
  - MRLIC researchers, collaborating with Fund country teams, applied the DIG and DIGNAR models to over 65 countries.
  - Applications in Ghana and Senegal illustrated feasible paths to increased borrowing, higher growth, and sustainable debt, opening richer dialogue on non-concessional borrowing flexibility.
  - LIC-DSF reviews incorporated spreadsheet-based tools drawing on DIG insights, including an investment-growth model with public investment inefficiencies and a model of fiscal adjustment and growth.
- Limits, critiques, and institutional adoption issues:
  - The DIG models were not systematically adopted by the Fiscal Affairs Department (FAD); reasons include FAD’s broader remit and a perception summarized by a former senior FAD manager: “The main message [of the DIG models] is that ‘If you choose productive investments and implement them effectively, then you can borrow non-concessionally, accelerate growth and maintain sustainable debt.’ We already knew that.”
  - DIG models emphasize public investment efficiency and rate of return to capital but do not address how to ensure that investments financed by borrowing are productive and implemented efficiently.
- Key data and risk shortcomings identified:
  - Data issues:
    - Hidden or misreported debt revealed after model application undermined trajectories (example: Mozambique’s about $2 billion (12% of GDP) secretly borrowed and misappropriated).
    - Senegal’s debt-to-GDP ratio was revised from 74 percent to 111 percent following audits that revealed hidden debts.
    - MRLIC could provide a global public good by developing methods to increase data transparency and accuracy and by integrating DIG+ with data on public capital stocks.
    - Buffie et al. (2012) assumed public capital stock in steady state in the baseline; if public capital stock is actually deteriorating due to inadequate recurrent spending, the calibration may over-estimate marginal product of public capital and productivity of public investment.
  - Risk and uncertainty:
    - DIG models use perfect-foresight assumptions and do not allow for unanticipated shocks; MRLIC addressed uncertainty by introducing a commodity price shock a few years into an optimal trajectory and then assuming full anticipation thereafter.
    - This deterministic approach limits DIG’s usefulness in increasingly volatile global environments (commodity prices, interest rates, inflation, growth) since 2015.
    - The Fund’s World Economic Outlook (WEO) produced alternative scenarios, but these do not appear to have been incorporated in DIG applications or LIC-DSAs beyond historical one-standard-deviation approaches.
    - MRLIC could undertake a retrospective assessment of LIC-DSF and DIG models’ performance to revise approaches to risk.
  - Optimism bias:
    - There is a possibility DIG models provided “too much” flexibility contributing to optimism bias in program design; several DIG-applied countries deviated from modeled trajectories.
    - Evidence on fiscal multipliers is mixed: some Fund research (Olivier Blanchard et al.) suggested larger multipliers in advanced economies, while other papers found multipliers in LICs were considerably smaller; Warner (2014) cautioned against public investment booms in LICs.
    - A thorough empirical retrospective of DIG-model applications and their relation to the emergence of debt distress is a high priority.
- Forward-looking uses:
  - DIG models could be run in reverse to derive insights for fiscal consolidation and debt restructuring, making them potentially relevant to current consolidation needs.
  - MRLIC may wish to take up debt restructuring and fiscal consolidation as a theme given their centrality to the Fund’s mandate.

### Diversification, inequality, gender, and climate change
- Diversification:
  - MRLIC sponsored work on diversification in LICs and developed a database on diversification in low-income countries.
  - The database is a global public good enabling research and policy work; maintaining and refreshing it regularly is critical.
- Inequality and gender:
  - DIG models were extended with heterogeneous agents to address inequality and gender, enabling anticipation of winners and losers from reforms and allowing design of compensating measures to keep programs on track.
  - Views on the future of inequality and gender work at the Fund are mixed; IMF gender mainstreaming strategy (IMF 2022) states the IMF will help members address gender disparities when macro-critical.
  - The review’s view aligns with mainstreaming: gender and inequality are central to the macroeconomics of investment and growth and thus to the Fund’s core mandate.
- Climate change:
  - Extensions of the DIG model to address climate change adaptation and mitigation (DIGNAD) are important, especially for small island economies subject to massive weather shocks.
  - MRLIC is supporting applications of the model as part of the Fund’s Resilience and Sustainability Trust.
  - Questions remain whether the DIGNAD platform—derived from a framework focused on financing infrastructure—is appropriate for managing climate risks, particularly given DIG’s limited scope for incorporating uncertainty and the extreme uncertainty of climate effects.

### Recommendations and possible future directions
- Strengthen data transparency and integration:
  - Develop methods to increase data transparency and accuracy on debt-related statistics.
  - Integrate DIG+ with public capital stock data to avoid mis-calibrated investment productivity assessments.
- Incorporate risk and uncertainty:
  - Undertake a retrospective assessment of LIC-DSF and DIG models’ performance to revise approaches to risk and scenario analysis.
  - Explore stochastic or scenario-based extensions that balance nonlinear DIG features with operational feasibility.
- Address political-economy and implementation realism:
  - Incorporate political risk and heterogeneity explicitly in model applications to better anticipate deviations from optimal trajectories.
  - Combine DIG insights with public investment management research to address how to ensure investment productivity and implementation efficiency.
- Broaden analytical focus to current priorities:
  - Apply DIG family tools to analyze fiscal consolidation and debt restructuring (run models “in reverse”).
  - Continue and expand work on diversification, inequality, gender, and climate change, maintaining databases and model extensions that are policy-relevant to LICs.

*Source: MRLIC final report (section 6), International Monetary Fund.*

### 7.  Going forward

### Going forward

### Overall assessment and impact
- The MRLIC project produced 192 working papers, exceeding expectations for a project of its type and scale.
- MRLIC research was used in 122 IMF country reports and several Board papers, demonstrating impact on IMF programs and policy.
- Signature themes—more flexible monetary policy frameworks and dynamically consistent debt-investment-growth strategies—are now routinely applied in most, if not all, low-income countries.
- Recent project extensions incorporate diversification, gender, inequality and climate change to address macro-critical issues.

### Institutional benefits and capacity building
- The project enhanced the profile of LIC research at the IMF by providing rigorous and empirically tractable analytical frameworks and datasets (e.g., diversification, inflation targeting), reducing barriers created by limited data.
- Active dissemination through training courses, toolkits, and dialogue with LIC authorities supported a potential virtuous cycle of growing demand for and supply of LIC research.
- The project enabled the Fund to hire contractual staff (some later joining regular staff) who engaged with country teams and authorities in disseminating the research.
- Strategic leadership by founding co-directors Andrew Berg and Catherine Pattillo, and institutional mechanisms to safeguard continuity beyond inaugural leadership, supported sustained, policy-relevant research.
- Collaboration between SPR and RES facilitated translation of research to policy; the project’s housing in two Research Department divisions may enable greater collaboration with Fiscal Affairs (FAD) and Monetary and Capital Markets (MCM).
- A joint RES–FAD project to estimate fiscal multipliers using AI tools for about 24 LICs would create a new database and tools for country teams and could produce a Staff Guidance Note in the future.

### Value for money
- Total project cost: $26 million over 12 years.
- IMF estimate of cost per working paper: $72,000 per working paper.
- The scale and cost structure allowed qualitative improvements in the Fund’s work in low-income countries.

### Institutionalization and attention to LICs
- References to LICs in the Fund’s flagship World Economic Outlook have been increasing over the past two decades; for the past decade the WEO data tables have included aggregates for a low-income LIDC group comprising most PRGT-eligible countries.
- A recently-instituted promotion policy requires that, to be promoted to management level (A15), staff in functional departments must have worked in a low-income country or FCS.
- Despite these developments, LICs remain largely below the radar of global financial stability and compete with emerging-market and advanced-country agendas for senior-management attention and research resources.
- Concern: progress on LIC research may not be sustained without additional funding from external partners.

### Priority themes and recommended research directions
- Immediate priority for LICs: address the current economic crisis characterized by low growth and high rates of debt distress.
  - Apply DIG+ work to fiscal consolidation and debt restructuring scenarios that typically follow debt distress.
  - Adapt monetary frameworks to examine inflation targeting and exchange-rate management under balance-of-payments pressure.
- Cross-cutting global issues (e.g., artificial intelligence and climate change) require more work but are not LIC-specific; MRLIC should prioritize LIC-specific crisis responses first.

### Gaps identified for a next MRLIC phase
- Three most important areas to address:
  - (i) Treatment of risk in the DIG+ family of models, including in the LIC-DSA, particularly as the world enters an era of “exceptionally high uncertainty” (Georgieva 2025).
  - (ii) Political economy of fiscal policy—explicit incorporation into models to reduce the risk that proposed fiscal trajectories are derailed by political considerations.
  - (iii) Data transparency, particularly on debt—too many programs go off-track because initial levels of debt were underestimated.

### Conclusion
- Taken together, the themes and identified gaps constitute a rich agenda for a next phase of a highly successful MRLIC project.
- Sustainability of the project’s achievements at the Fund is uncertain without continued programmatic involvement and additional external funding.

---


_Source: https://www.imf.org/-/media/files/topics/lics/macro-research-for-development/related-links/devarajan-oconnell-final-report-on-mrlic-project-final.pdf_