## _wp1586

## Source details

**Canonical URL:** [_wp1586](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1586.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1586.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1586.pdf.json)

---

### I. Introduction — context and motivation
- International asset flows have expanded rapidly due to financial market development, technology improvements, and liberalization of current and financial accounts.
- Many low-income developing countries (LIDCs) historically relied on foreign aid and concessional loans and experienced limited expansion in private capital inflows.
- Increases in private capital inflows to many LIDCs began in the 1990s, primarily driven by foreign direct investment (FDI).
- Recent evidence suggests non-FDI private inflows may have started to pick up; sovereign bond issuance by some LIDCs raises questions about (re)gaining market access.
- The chapter constructs a newly assembled dataset of gross private flows with extended LIDC coverage (1990–2012) to explore recent patterns and trends, emphasizing non-FDI inflows.

### II. Dataset coverage and purpose
- Coverage:
  - 58 LIDCs and 92 EMs for the period 1990 to 2012.
- Uses:
  - Explore recent patterns and trends of gross private flows in LIDCs, with emphasis on non-FDI inflows.
  - Study episodes of large gross capital inflows (capital inflow surges) in LIDCs relative to EMs.
  - Examine correlations between non-FDI inflows and other balance of payments components, and implications for macroeconomic impact and vulnerabilities.

### III. Definitions and measurement
- LIDC definition:
  - Countries eligible for the IMF’s concessional lending vehicle Poverty Reduction and Growth Trust (PRGT) and with per capita Gross National Income (GNI) less than the PRGT income graduation level for non-small states (USD 2,390).
- Flow definitions:
  - Inflows = changes in non-resident holdings of domestic assets.
  - Outflows = changes in resident holdings of foreign assets.
  - Net flows = inflows − outflows.
- Construction of gross non-FDI private capital inflows (formula preserved from source):
  - Gross non-FDI Private Inflows = ∆Portfolio Debt Liabilities + ∆Portfolio Equity Liabilities + ∆Other Investment Liabilities + ∆Financial Derivatives − ∆Proxy for Official Other Investment Liabilities
- Treatment notes:
  - Debt forgiveness is accounted in capital transfer; counterpart is reduction of liability in the financial account and is included in the Proxy for Official Other Investments and thus excluded from gross non-FDI private inflows.
  - Official portfolio investment proxy typically not included (typically zero flows).
  - BOPS limitation: breakdown by debtor only; proxy approach used for Liabilities to Official Creditors via Official Liabilities from BOPS given strong correlation.

### IV. Data sources, harmonization, and adjustments
- Primary source: Balance of Payments Statistics (BOPS), compiled under IMF Balance of Payments Manual version 5 (BPM5).
  - BOPS originates from national compilers and is validated by the IMF Statistics Department.
  - BOPS coverage gaps exist for some countries and periods, especially among LIDCs.
- Secondary source: World Economic Outlook (WEO) database.
  - WEO gross capital flow data based on estimates and projections from IMF country desks; historical data may be adjusted and ad hoc.
- Auxiliary databases for cross-checking and outlier judgement:
  - Bank for International Settlements (BIS) locational banking statistics.
  - Emerging Portfolio Fund Research (EPFR) Global Database.
  - Dealogic (external bond issuance).
- Harmonization and notable adjustments:
  - WEO practice: whole series sometimes populated by zeros; “n.a." observations replaced by zeros when surrounded by strings of zeros.
  - BOPS distinguishes “n.a.” (nonexistent) from zero flows; WEO does not always preserve that distinction.
  - Specific supplementation examples where WEO was used instead of BOPS: Mauritius (consistency after structural break), Ghana (EPFR evidence 2010), Mauritania (Other Investments Assets 1999-2012), Senegal (BIS data used 2011-12 for Other Investments Assets).
  - 2009 SDR allocations: dataset reflects IMF Statistics Department adoption of BPM6 treatment (recorded as increase in reserves and as other investment liabilities).

### V. Treatment of missing data, outliers, and BPM6 considerations
- Procedure for constructing extended dataset:
  1. For missing BOPS observations, consider available WEO; replacing non-missing BOPS with WEO only in exceptional cases.
  2. Check WEO series for artificial patterns (repeated numbers, linear trends); do not use such series.
  3. Cross-check BOPS and WEO for large discrepancies; investigate spikes via IMF country desks and reports.
  4. Use auxiliary databases (BIS, EPFR, Dealogic) to assess trends and identify outliers.
  5. BOPS coverage up to 2012; where BOPS lacked 2012 data, WEO estimates used when reasonable; otherwise reported as not available.
  6. Final scan: scale flows by GDP and cross-check with WEO for remaining outliers.
- BPM6 vs BPM5:
  - IMF Statistics Department began publishing BPM6-based data since August 2012.
  - Two reasons to opt for BPM5 in this project: (i) BPM6 series shorter, requiring conversion to obtain long series; (ii) WEO database presented according to BPM5 at time of construction facilitating direct comparison.
  - Transition statuses and conversion compatibility categorized as c, M, d; 35 countries already transitioned to BPM6 are listed in source.
  - WEO used to supplement or replace IFS data in specified categories for listed countries (detailed in source).

### VI. Empirical comparisons between BOPS and WEO inflow series (correlations)
- Correlations by liability type (BOPS vs WEO):
  - Portfolio Investment correlations: LIDC=0.523; EM=0.580.
  - Other Investments correlations: LIDC=0.550; EM=0.698.
  - Foreign Direct Investment correlations: LIDC=0.841; EM=0.878.
- Interpretation:
  - Lower correlations for Portfolio and Other Investments indicate sizable heterogeneity and caution when using WEO alone.

### VII. Key empirical findings and stylized facts
- Aggregate and distributional findings:
  - Top quartile of non-FDI inflows as a share of GDP to LIDCs are comparable to the EM median inflow.
  - After the global financial crisis, size of top quartile inflows in LIDCs converged to EM top quartile inflows.
  - More LIDCs began to experience capital inflow surges from 2007 onwards; this was three years after surge frequency in EMs picked up during the third wave.
  - Non-FDI inflows to LIDCs positively correlated with some other balance of payments components similarly to EMs, but display a higher negative correlation with current account balances in LIDCs—suggesting a potentially more direct macroeconomic impact in LIDCs.
  - Importance of analyzing gross inflows (rather than net flows) for LIDCs because gross inflows better capture changes in market access and can reflect distinct financial stability implications.
- Composition and trends (benchmark LIDC subgroup):
  - Largest component: change in Other Investment Liabilities.
  - Financial derivatives: very small component.
  - Portfolio flows (Debt + Equity) increased notably in later years; portfolio debt shows sharp increase in 2012; equity inflows increased steadily during 2006-2012.
  - Other Investments breakdown shifted: until mid-2000’s most flows classified as Other Liabilities of Other Sectors; from 2005 onwards bank loans, trade credit, currency and deposits became largest components.
  - BIS comparison: close correspondence between BOPS Other Investments flows and BIS Loans and Deposits despite coverage and valuation/timing differences.
- Heterogeneity across LIDC subgroups:
  - Small States: 29 small states in dataset (population below 1.5 million); note that using this definition there are 33 developing small states total; 4 lacked capital flows data.
  - Fragile States: 30 fragile states included (World Bank definition: CPIA ≤ 3.2 or UN/regional peace-keeping mission in prior 3 years).
  - Benchmark sample: non-small and non-fragile economies includes 32 LIDCs and 63 EMs; used as baseline.
  - Flows as percent of GDP considerably larger in small states.
  - Flows to fragile LIDCs more volatile than benchmark LIDC sample.
  - In benchmark sample, gross FDI inflows trend upwards since the 1990’s; gross non-FDI private inflows remain relatively flat until 2007.
  - Some LIDCs had non-FDI private inflows in the 1990’s (Kenya, Nigeria, Zambia); many others experienced increases only more recently.

### VIII. Correlations with other variables (Table 1 summary; variables in percent of GDP)
- Correlation coefficients (excluding small and fragile economies; significance as reported):
  - Non-FDI Private Inflows vs FDI Inflows: EMs -0.0250; LIDCs 0.154 ∗∗∗
  - Non-FDI Private Inflows vs Non-FDI Outflows: EMs 0.341 ∗∗∗; LIDCs 0.489 ∗∗∗
  - Non-FDI Private Inflows vs Current Account Balance: EMs -0.0355; LIDCs -0.322 ∗∗∗
  - Non-FDI Private Inflows vs Trade Balance: EMs -0.0939 ∗∗∗; LIDCs -0.156 ∗∗∗
  - Non-FDI Private Inflows vs ∆Reserves: EMs 0.150 ∗∗∗; LIDCs 0.142 ∗∗∗
  - Non-FDI Private Inflows vs Errors and Omissions: EMs 0.0825 ∗∗∗; LIDCs 0.000729
  - Non-FDI Private Inflows vs Net Non-FDI Flows: EMs 0.159 ∗∗∗; LIDCs 0.427 ∗∗∗
- Sample size note: N 14497 (EMs) and 36 (LIDCs) as shown in Table 1 note.
- Interpretations:
  - Gross non-FDI private inflows associated with weaker current account balance in LIDCs (statistically significant), not significant for EMs.
  - Gross non-FDI private inflows negatively related to trade balance in both groups (significant).
  - Positive correlation between gross non-FDI private inflows and gross FDI inflows in LIDCs (complementarity); not present for EMs.
  - Gross non-FDI private inflows associated with gross non-FDI outflows, especially in LIDCs— inflows not completely offset by outflows.
  - Some association with errors and omissions for EMs but not for LIDCs.

### IX. Episodes of surges in gross non-FDI private inflows
- Surge definition (two criteria):
  - (i) gross non-FDI private inflows as percent of GDP are in top quartile of the country’s own time series;
  - (ii) gross non-FDI private inflows as percent of GDP are in top quartile of the cross-country sample.
- Sensitivity and thresholds:
  - Top 25th percentile global threshold = 3.7 percent of GDP.
  - Top 30th percentile global threshold = 3.1 percent of GDP.
  - Top 20th percentile global threshold = 4.4 percent of GDP.
- Number of surges detected (95 developing countries excluding small and fragile economies, 2034 observations):
  - Top 25th percentile: total 296 surges; 77 surges in LIDCs (about 26 percent of total surges).
  - Top 30th percentile: 384 surges.
  - Top 20th percentile: 219 surges.
- Waves of surges:
  - Wave 1: 1990-1994 (U.S. monetary loosening; Brady Plan debt restructuring).
  - Wave 2: 1996-1997 (pre-Asian and Russian crises).
  - Wave 3: 2004-2008 (pre-global financial crisis build-up); LIDCs picked up this wave with a lag (not until 2007 did surges approach levels seen in other developing countries).
  - Wave 4: 2010-2012 (post-global financial crisis resurgence).
- Frequency and distribution:
  - Surges less frequent in LIDCs than other developing economies, but proportion of surges in LIDCs relative to overall surges is stable at around 26 percent across cutoff choices.
  - 77 surges identified in LIDCs; surges spread over 25 of 32 non-small and non-fragile LIDCs.
  - About half of surges occurred in Sub-Saharan Africa.
  - Many surges coincide with periods of large debt reduction programs (HIPC, MDRI); about half of surges occurred in HIPCs.

### X. Selected country-level surge list (PPP-adjusted per capita GDP as reported)
- Selected entries (country — PPP-adjusted per capita GDP — surge years reported):
  - Mongolia — 5297.71 — 1998, 2008, 2010-2012
  - Bolivia — 5041.31 — 1992, 1994
  - Honduras — 4743.52 — 1997
  - Nicaragua — 4351.67 — 1997, 2006-2008, 2011
  - Vietnam — 3787.82 — 2007, 2009-2010, 2012
  - Moldova — 3381.16 — 1994, 1997, 2007-2008
  - Ghana — 3315.99 — 2007
  - Lao People's Dem. Rep — 2846.44 — 2010
  - Papua New Guinea — 2736.07 — 1991, 2003, 2005
  - Nigeria — 2697.08 — 1991-1992, 1994-1995, 1997
  - Cambodia — 2395.21 — 2000, 2011-2012
  - Kyrgyz Republic — 2349.93 — 1996, 1998, 2007, 2009, 2012
  - Cameroon — 2337.73 — 2001
  - Lesotho — 2105.51 — 2012
  - Mauritania — 2097.90 — 2008, 2010-2011
  - Senegal — 1908.44 — 2006-2008, 2012
  - Kenya — 1736.93 — 1994, 1998-1999, 2011-2012
  - Zambia — 1683.12 — 1990-1991, 2000-2001, 2007
  - Benin — 1553.80 — 1991, 1994, 1999, 2007, 2009
  - Burkina Faso — 1488.77 — 1994, 2007
  - Mali — 1101.74 — 2000
  - Mozambique — 1027.04 — 1999-2000, 2002, 2012
  - Madagascar — 956.99 — 1996, 2007-2010
  - Malawi — 848.44 — 1994
  - Niger — 813.05 — 2010
- Note: GDP data sourced from WEO database as reported in source.

### XI. Comparisons with other datasets and coverage improvements
- Comparison dataset: Financial Flows Analytics (FFA) by Bluedorn et al. (2013) based on IMF BOP (BPM6) — LIDC coverage limited.
- Coverage improvements in extended database:
  - Aggregate measure of non-FDI private inflows: increased coverage is 19 percent for the LIDC group.
  - Range of additional coverage across components for LIDCs:
    - Official Other Investment Flows: 8.4 percent
    - Other Investments Liabilities: 13.4 percent
  - Additional coverage for entire sample of LIDCs:
    - Other investments: 9.1 percent
    - Portfolio assets: 8.1 percent
  - Coverage increases for direct investment components:
    - Direct investment into the country: 16.1 percent
    - Direct investment abroad: 1.7 percent
- Methodological note: comparison used conservative approach by comparing only non-zero values in each individual series to avoid counting missing variables mistakenly replaced by zeros; aggregates computed whenever at least one subcomponent available.

### XII. Major outliers, caveats, and country-specific notes
- Financial centers:
  - Bahrain and Bahamas: treat inflows in conjunction with outflows due to financial center dynamics.
- Event-related outliers:
  - 1991 Kuwait: GDP shrank due to Gulf War leading to large capital flow over GDP.
- Country supplementation examples:
  - Countries where WEO supplemented BOPS or were used to replace IFS in specific series include Chad, Comoros, Equatorial Guinea, Eritrea, Gabon, Madagascar, Malaysia, Mauritania, Niger, Saudi Arabia, Turkmenistan, Uzbekistan, and others listed in appendices.
- BIS comparison methodology:
  - BIS table 7A “Loans and deposits” used for cross-checking other investments; BIS flows calculated by difference in stocks adjusted by exchange-rate change and therefore not identical to BoP flow data.

### XIII. Key implications and concluding observations
- Objective achieved: enhanced coverage for developing economies, focusing on LIDCs and gross non-FDI private flows to fill gaps in datasets that emphasize EMs.
- Main empirical conclusions:
  - Elevated activity in gross non-FDI private capital inflows evident for many “frontier" LIDCs, indicating increased market access since early 2000’s.
  - Although average gross non-FDI private inflows (as percent of GDP) to LIDCs are much lower than EMs, LIDC top quartile inflow as share of GDP is comparable to EM median; post-crisis LIDC top quartile converged to EM top quartile.
  - LIDCs participated less in first two waves of global inflow surges but were materially affected by the third wave (2004-2008) with a lag and by the fourth wave (2010-2012).
- Research and policy considerations:
  - Classification issues, series breaks, and differences across WEO, BOPS, BIS, and other sources require careful handling; the dataset documents sources and outliers.
  - The dataset enables researchers and policymakers to better include LIDCs in analyses of capital flow dynamics and to study balance-sheet implications and market access changes through gross flow measures.

*Source: Extended LIDC database and related documentation as presented in the provided chapter content.*

### 10.    Other Private Investments Inflows and BIS Loans (USD MM) .  .  .  .  .  .  .  .  .  .  .  .    32

### 10.    Other Private Investments Inflows and BIS Loans (USD MM)

### I. Introduction — context and motivation
- The international flow of assets has expanded rapidly over past decades driven by financial market development, technology improvements, and liberalization of current and financial accounts.  
- Many low-income developing countries (LIDCs) have experienced limited expansion in private capital inflows and historically relied on foreign aid and concessional loans.  
- Many LIDCs began to see increases in private capital inflows only in the 1990s, primarily driven by foreign direct investment (FDI).  
- Recent speculation and fragmented evidence suggest non-FDI private inflows may have started to pick up, and sovereign bond issuance by some LIDCs raises questions about (re)gaining market access.  
- The paper constructs a newly assembled dataset of gross private flows with extended LIDC coverage to explore recent patterns and trends, with a particular focus on non-FDI inflows.

### II. Dataset coverage and purpose
- The dataset covers 58 LIDCs and 92 EMs for the period 1990 to 2012.  
- The dataset is used to:  
  - Explore recent patterns and trends of gross private flows in LIDCs with emphasis on non-FDI inflows.  
  - Study episodes of large gross capital inflows (capital inflow surges) in LIDCs relative to EMs.  
  - Examine correlations between non-FDI inflows and other balance of payments components, and implications for macroeconomic impact and vulnerabilities.

### III. Key empirical findings reported
- The top quartile of non-FDI inflows as a share of GDP to LIDCs are comparable to the EM median inflow.  
- After the global financial crisis, the size of top quartile inflows in LIDCs converged to EM top quartile inflows.  
- More LIDCs began to experience capital inflow surges from 2007 onwards; this was three years after surge frequency in EMs picked up during the third wave of inflows.  
- Non-FDI inflows to LIDCs are positively correlated with some other balance of payments components similarly to EMs, but display a higher negative correlation with current account balances in LIDCs—suggesting a potentially more direct macroeconomic impact in LIDCs.  
- The paper emphasizes the importance of analyzing gross inflows (rather than net flows) for LIDCs because changes in gross inflows better capture changes in market access and can reflect distinct financial stability implications.

### IV. Definitions and methodological framing
- LIDCs are defined as countries designated eligible for the IMF’s concessional lending vehicle Poverty Reduction and Growth Trust (PRGT) and with per capita Gross National Income (GNI) less than the PRGT income graduation level for non-small states (USD 2,390).  
- Inflows refer to changes in non-resident holdings of domestic assets; outflows to changes in resident holdings of foreign assets; net flows = inflows − outflows.  
- The study concentrates on gross inflows to better capture market access and potential balance-sheet implications of various asset flows.

### V. Data sources and comparative assessment
- Primary data source: Balance of Payments Statistics (BOPS), compiled following IMF Balance of Payments Manual version 5 (BPM5).  
  - BOPS originates from official national compilers and is validated by the IMF Statistics Department.  
  - Some gaps exist in BOPS coverage for some countries and periods, especially among LIDCs.  
- Secondary source used to fill gaps and cross-check: World Economic Outlook (WEO) database.  
  - WEO gross capital flow data are based on estimates and projections from IMF country desks; historical data may be adjusted by country desks and can be ad hoc.  
- Auxiliary databases used for cross-checking and judging outliers and trends: Bank for International Settlements (BIS) locational banking statistics, Emerging Portfolio Fund Research (EPFR) Global Database, and Dealogic (external bond issuance).

### VI. Empirical comparison between BOPS and WEO inflow series (reported correlations)
- Correlations between BOPS and WEO inflow series by liability type:
  - Portfolio Investment correlations: LIDC=0.523; EM=0.580.
  - Other Investments correlations: LIDC=0.550; EM=0.698.
  - Foreign Direct Investment correlations: LIDC=0.841; EM=0.878.
- The lower correlations for Portfolio and Other Investments across datasets indicate sizable heterogeneity and raise caution when using WEO alone.

### VII. Treatment of missing data and outliers — procedural steps
- Procedure followed when constructing the extended dataset:
  1. For missing observations in BOPS, available WEO data are considered. Replacing non-missing BOPS with WEO is only done in exceptional cases (discussed in the paper).  
  2. WEO series are checked for artificial patterns (e.g., repeated numbers or linear trends); series with such patterns are not used. WEO is used to fill BOPS gaps only if there is consistent overlap between BOPS and WEO for the country/time period.  
  3. BOPS and WEO are cross-checked for large discrepancies. Large spikes are investigated to determine if they are genuine events, methodological differences, or data noise; IMF country desks and country reports are consulted when appropriate.  
  4. Auxiliary databases (BIS locational banking statistics, EPFR, Dealogic) are used to judge overall trends and identify outliers.  
  5. As of dataset construction, available BOPS covered up to 2012; where BOPS lacked 2012 data, WEO estimates were used when reasonable; otherwise data were reported as not available.  
  6. Final scan of the extended dataset involved scaling flows by GDP and cross-checking with WEO for remaining outliers.

### VIII. Data notes on auxiliary sources
- BIS locational banking statistics: quarterly data on international financial claims and liabilities of bank offices resident in reporting countries; compiled on principles consistent with balance of payments.  
- EPFR: weekly frequency data on institutional investors using vehicles such as mutual funds to allocate resources; provides equity and bond flow information relevant for cross-checking portfolio flows.  
- Dealogic: used for external bond issuance evidence.

*Source: WEO and BOPS.*

### 7.  The data source used for each series and time period for every country are tracked and

### 7.  The data source used for each series and time period for every country are tracked and comments on outliers/events are incorporated in the database.

### Data sources, harmonization, and notable adjustments
- WEO and BOPS are the main databases used; harmonized concepts between databases are taken into account (example: derivatives treated differently across datasets).
- WEO practice: whole series of assets are sometimes populated by zeros; “n.a." observations are replaced by zeros when surrounded by strings of zeros.
- BOPS distinguishes between nonexistent data (“n.a.”) and zero flows; WEO does not always preserve that distinction.
- Only a few cases where WEO data are used instead of BOPS when BOPS is available. Notable examples:
  - Mauritius: WEO used for consistency due to a structural break from a new survey (2010 onwards) capturing offshore company capital flows that appeared in BOPS for 2010-2012.
  - Ghana: inflows captured in EPFR in 2010 are taken into account.
  - Mauritania: WEO used for Other Investments Assets 1999-2012.
  - Senegal: BIS data used for 2011-12 for Other Investments Assets.
- 2009 SDR allocations: under BPM6 recorded both as an increase in reserves (asset side) and as other investment liabilities (liabilities side). BOPS followed BPM5 historically, but IMF Statistics Department adopted BPM6 treatment and this dataset reflects that.
- Dealogic: described as an electronic platform used by investment banks worldwide providing primary market equity, bond and loans information (contextual mention in source).

### Construction of gross non-FDI private capital inflows (measurement)
- Purpose: compute measure of gross non-FDI private capital inflows, excluding official flows directed to private domestic agents.
- Distinction:
  - Creditor official flow vs debtor official flow: foreign private investor buying official domestic assets = private inflow; foreign official inflows directed to private domestic agent = excluded.
- WEO components relevant:
  - Official Liabilities (including use of Fund credit) — codes: BFOLG; BFOL_O and BFOL_B refer to complementary groups (Banks and Other Sectors).
  - Liabilities to Official Creditors — BFOL_G in WEO.
  - WEO provides BFPL_G for portfolio investment by official creditors (rare).
- BOPS limitation: breakdown is by debtor only; can reproduce Official Liabilities but not Liabilities to Official Creditors.
- Proxy approach: Official Liabilities from BOPS (Other Investments Liabilities for general government and monetary authorities broken into Trade Credit, Loans, Currency and Deposits, Other Liabilities) is used as a proxy for Liabilities to Official Creditors given strong correlation.
- Formula used:
  - Gross non-FDI Private Inflows = ∆Portfolio Debt Liabilities + ∆Portfolio Equity Liabilities + ∆Other Investment Liabilities + ∆Financial Derivatives − ∆Proxy for Official Other Investment Liabilities
- Treatment of debt forgiveness:
  - Debt forgiveness accounted in capital transfer; counterpart is reduction of liability in financial account. Outflows from debt forgiveness are included in the Proxy for Official Other Investments and thus excluded from gross non-FDI private inflows.
- Official portfolio investment proxy not included (typically zero flows).
- Comparison and validation:
  - Other Investments compared to BIS Locational Banking Statistics category Loans and Deposits (BIS Table 7A); flows derived from stock differences adjusted by exchange-rate changes.
  - Official Liabilities from WEO compared to a synthetic series reproducing Official Liabilities from BOPS; observations supplemented with WEO when series were sufficiently close.
  - Countries where supplementation occurred include: Chad, Comoros, Equatorial Guinea, Eritrea, Gabon, Madagascar, Malaysia, Mauritania, Niger, Saudi Arabia, Turkmenistan, Uzbekistan.
  - Malaysia exception: series complemented for 2010-2012 using coefficients estimated from regressions of the relationship between BOPS and WEO in previous years.
- Major outliers and caveats:
  - Bahrain and Bahamas: financial centers—analyze inflows in conjunction with outflows.
  - 1991 Kuwait: GDP shrank due to Gulf War leading to large capital flow over GDP.
- Comparison dataset: Financial Flows Analytics (FFA) by Bluedorn et al. (2013) based on IMF BOP (BPM6) provides aggregate financial account components but LIDC coverage limited; this paper’s dataset aims to significantly improve LIDC coverage.

### Stylized facts: country heterogeneity
- LIDC sample distinctions:
  - Small States: 29 small states in dataset (population below 1.5 million). Note: using this definition there are 33 developing small states total; 4 lacked capital flows data.
  - Fragile States: 30 fragile states included (World Bank definition: CPIA ≤ 3.2 or UN/regional peace-keeping mission in prior 3 years).
  - Benchmark sample: non-small and non-fragile economies includes 32 LIDCs and 63 EMs; used as baseline for trend analysis.
- Observations:
  - Flows as percent of GDP considerably larger in small states.
  - Flows to fragile LIDCs more volatile than benchmark LIDC sample.
  - In benchmark sample, gross FDI inflows trend upwards since the 1990’s; gross non-FDI private inflows remain relatively flat until 2007.
  - Some LIDCs had non-FDI private inflows in the 1990’s (Kenya, Nigeria, Zambia); many others experienced increases only more recently.

### Stylized facts: composition and trends of non-FDI private inflows
- Composition (benchmark LIDC subgroup, percent of GDP averages shown in Figure 2 and Figure 5 in source):
  - Largest component: change in Other Investment Liabilities.
  - Financial derivatives: very small component.
  - Portfolio flows (Debt + Equity) increased notably in later years; portfolio debt shows sharp increase in 2012; equity inflows increased steadily during 2006-2012.
- Other Investments breakdown (BOPS classification; Figure 6 in source):
  - Components include trade credit, loans (banks and other sectors), currency and deposits (banks and other sectors), and other liabilities by debtor type.
  - Until mid-2000’s most flows classified as Other Liabilities of Other Sectors; from 2005 onwards bank loans, trade credit, currency and deposits became largest components.
  - Change may reflect improved classification and diversification of flows (financial and trade integration).
- BIS comparison: close correspondence between BOPS Other Investments flows and BIS Loans and Deposits despite coverage differences and valuation/timing differences.

### Stylized facts: comparing LIDCs and EMs
- Share of gross non-FDI private inflows to total gross private inflows:
  - Developing countries sample mean (1990-2012): varies between 30 and 40 percent.
  - LIDC mean for entire period: closer to 20 percent, with an increasing trend lifting it from below average in early 1990’s to above average in later years.
  - EMs show decreasing trend, influenced by 2008 crisis.
- Correlations (Table 1 in source) — variables measured in percent of GDP; country sample excludes small and fragile economies; significance levels as reported:
  - Correlation between Non-FDI Private Inflows and:
    - FDI Inflows: EMs -0.0250; LIDCs 0.154 ∗∗∗
    - Non-FDI Outflows: EMs 0.341 ∗∗∗; LIDCs 0.489 ∗∗∗
    - Current Account Balance: EMs -0.0355; LIDCs -0.322 ∗∗∗
    - Trade Balance: EMs -0.0939 ∗∗∗; LIDCs -0.156 ∗∗∗
    - ∆Reserves: EMs 0.150 ∗∗∗; LIDCs 0.142 ∗∗∗
    - Errors and Omissions: EMs 0.0825 ∗∗∗; LIDCs 0.000729
    - Net Non-FDI Flows: EMs 0.159 ∗∗∗; LIDCs 0.427 ∗∗∗
  - Sample size: N 14497 (EMs) and 36 (LIDCs) as shown in Table 1 note.
- Interpretation of correlations:
  - Gross non-FDI private inflows associated with weaker current account balance in LIDCs (statistically significant), not significant for EMs.
  - Gross non-FDI private inflows negatively related to trade balance in both groups (significant).
  - Gross non-FDI private inflows correlated with gross FDI inflows in LIDCs (positive and significant), suggesting complementarity among flow types in LIDCs; not present for EMs.
  - Positive association between gross non-FDI private inflows and foreign reserve accumulation is similar across groups.
  - Gross non-FDI private inflows associated with gross non-FDI outflows, especially in LIDCs — inflows not completely offset by outflows.
  - Some association with errors and omissions for EMs but not for LIDCs.

### Episodes of surges in gross non-FDI private inflows
- Surge definition (two criteria):
  - (i) gross non-FDI private inflows as percent of GDP are in top quartile of the country’s own time series;
  - (ii) gross non-FDI private inflows as percent of GDP are in top quartile of the cross-country sample.
- Sensitivity to cutoff: top 25th percentile used in main discussion; alternatives produce different counts:
  - Top 25th percentile global threshold = 3.7 percent of GDP.
  - Top 30th percentile global threshold = 3.1 percent of GDP.
  - Top 20th percentile global threshold = 4.4 percent of GDP.
  - Number of surges detected (95 developing countries excluding small and fragile economies, 2034 observations):
    - Top 25th percentile: total 296 surges; 77 surges in LIDCs (about 26 percent of total surges).
    - Top 30th percentile: 384 surges.
    - Top 20th percentile: 219 surges.
- Waves of surges (Figure 9 in source; shaded area denotes episodes with share of surges > 12 percent):
  - Wave 1: 1990-1994 (U.S. monetary loosening; Brady Plan debt restructuring).
  - Wave 2: 1996-1997 (pre-Asian and Russian crises).
  - Wave 3: 2004-2008 (pre-global financial crisis build-up); LIDCs picked up this wave with a lag (not until 2007 did surges approach levels seen in other developing countries).
  - Wave 4: 2010-2012 (post-global financial crisis resurgence).
- Frequency and distribution:
  - Surges less frequent in LIDCs than other developing economies, but proportion of surges in LIDCs relative to overall surges is stable at around 26 percent across cutoff choices.
  - 77 surges identified in LIDCs; surges spread over 25 of 32 non-small and non-fragile LIDCs (Table 2 in source).
  - Regional and program associations:
    - About half of surges occurred in Sub-Saharan Africa.
    - Many surges coincide with periods of large debt reduction programs (HIPC, MDRI); about half of surges occurred in HIPCs.
- Percentile comparisons (Figure 8 in source):
  - Top quartile inflows to LIDCs as share of GDP comparable to EM median inflows.
  - After the crisis, LIDC top quartile inflow converged to EM top quartile inflow.
  - For top-quartile LIDCs, rise in inflows occurred later relative to EM top-quartile rise, and LIDC top quartile rebounded to pre-crisis highs by latest sample years.

### Country-level surge list (selected from Table 2)
- Countries with surges and PPP-adjusted per capita GDP values as reported:
  - Mongolia — 5297.71 — 1998, 2008, 2010-2012
  - Bolivia — 5041.31 — 1992, 1994
  - Honduras — 4743.52 — 1997
  - Nicaragua — 4351.67 — 1997, 2006-2008, 2011
  - Vietnam — 3787.82 — 2007, 2009-2010, 2012
  - Moldova — 3381.16 — 1994, 1997, 2007-2008
  - Ghana — 3315.99 — 2007
  - Lao People's Dem. Rep — 2846.44 — 2010
  - Papua New Guinea — 2736.07 — 1991, 2003, 2005
  - Nigeria — 2697.08 — 1991-1992, 1994-1995, 1997
  - Cambodia — 2395.21 — 2000, 2011-2012
  - Kyrgyz Republic — 2349.93 — 1996, 1998, 2007, 2009, 2012
  - Cameroon — 2337.73 — 2001
  - Lesotho — 2105.51 — 2012
  - Mauritania — 2097.90 — 2008, 2010-2011
  - Senegal — 1908.44 — 2006-2008, 2012
  - Kenya — 1736.93 — 1994, 1998-1999, 2011-2012
  - Zambia — 1683.12 — 1990-1991, 2000-2001, 2007
  - Benin — 1553.80 — 1991, 1994, 1999, 2007, 2009
  - Burkina Faso — 1488.77 — 1994, 2007
  - Mali — 1101.74 — 2000
  - Mozambique — 1027.04 — 1999-2000, 2002, 2012
  - Madagascar — 956.99 — 1996, 2007-2010
  - Malawi — 848.44 — 1994
  - Niger — 813.05 — 2010
- Note: GDP data from WEO database as reported in source.

### Key implications and concluding observations
- Dataset objective: enhance coverage for developing economies, focusing on low-income developing countries (58 LIDCs) and gross non-FDI private flows to fill gaps in existing datasets that emphasize EMs.
- Main empirical findings:
  - Elevated activity in gross non-FDI private capital inflows evident for many “frontier" LIDCs, indicating increased market access since early 2000’s.
  - Although average gross non-FDI private inflows (as percent of GDP) to LIDCs are much lower than EMs, LIDC top quartile inflow as share of GDP is comparable to EM median; post-crisis LIDC top quartile converged to EM top quartile.
  - LIDCs participated less in first two waves of global inflow surges but were materially affected by the third wave (2004-2008) with a lag and by the fourth wave (2010-2012).
- Data caveats and research potential:
  - Classification issues, series breaks, and differences across WEO, BOPS, BIS, and other sources require careful handling; the dataset documents sources and outliers.
  - The dataset aims to enable researchers and policymakers to better include LIDCs in analyses of capital flow dynamics.

*Source: Extended LIDC database and related documentation as presented in the provided chapter content.*

### REFERENCES

### _wp1586 - REFERENCES

### References cited
- Alfaro, L., Kalemli-Ozcan, S., and Volosovych, V. (2008). “Why Doesn’t Capital Flow from Rich to Poor Countries? An Empirical Investigation.” The Review of Economics and Statistics, 90(2), 347-368.
- Alfaro, L., Kalemli-Ozcan, S., and Volosovych, V. (2014). “Sovereigns, Upstream Capital Flows and Global Imbalances.” Journal of European Economic Association, 12(5), 1240-1284.
- Araujo, J., David, A., van Hombeeck, C., and Papageorgiou, C. (2015). “Joining the Club? Procyclicality of Private Capital Inflows in Low Income Developing Countries.” IMF Working Paper (forthcoming).
- Bluedorn, J., Duttagupta, R., Guajardo, J. and Topalova, P. (2013). “Capital Flows are Fickle: Anytime, Anywhere". IMF Working Paper 13/183.
- Broner, F., Didier, T., Erce, A., and Schmukler, S. (2013). “Gross Capital Flows: Dynamics and Crises.” Journal of Monetary Economics, 60(1), 113-133.
- Calvo, G.A. (1998). “Capital Flows and Capital-Market Crises: The Simple Economics of Sudden Stops.” Journal of Applied Economics, 1(1), 35-54.
- Calvo, G., and Mendoza, E. (2000). “Rational Contagion and the Globalization of Securities Markets.” Journal of International Economics, 51(1), 79-113.
- Coase, R.H. (1937). “The Nature of the Firm.” Economica, 4(16), 386-405.
- Dharmapala, D., and Hines Jr, J.R. (2009). “Which Countries Become Tax Havens?” Journal of Public Economics, 93(9), 1058-1068.
- Dorsey, T., Tadesse, H., Singh, S., and Brixiova, Z. (2008). “The Landscape of Capital Flows to Low-Income Countries.” IMF Working Paper 08/51.
- Forbes, K.J. and Warnock, F.E. (2012). “Capital Flow Waves: Surges, Stops, Flight, and Retrenchmen.” Journal of International Economics, 88, 235-251.
- Gelos, G., Sahay, R. and Sandleris, G. (2004). “Sovereign Borrowing by Developing Countries: What Determines Market Access?” IMF Working Paper 04/221.
- Ghosh, R.A., Qureshi, M., Zalduendo, J., and Kim, J. (2014). “Surges.” Journal of International Economics, 92(2), 266-285.
- Gourinchas, P.O., and Rey, H. (2013). “External Adjustment, Global Imbalances and Valuation Effects, NBER Working Paper w19240.
- Hausmann, R., and Fernández-Arias, E. (2001). “Foreign Direct Investment: Good Cholesterol?" In Foreign Direct Investment Versus Other Flows to Latin America, edited by Jorge Braga de Macedo and Enrique V. Iglesias. Paris: Inter-American Development Bank and OECD.
- International Monetary Fund (2013). Macroeconomic Issues in Small States and Implications for Fund Engagement. Washington, DC: Author.
- Lane, P.R. (2015). “International Financial Flows in Low-Income Countries.” Pacific Economic Review, 20: 1, 49-72.
- Milesi-Ferretti, G.M., and Lane, P.R. (2010). “Cross-Border Investment in Small International Financial Centers.” IMF Working Paper 10/38.
- Milesi-Ferretti, G.M., and Tille, C. (2011). “The Great Retrenchment: International Capital Flows During the Global Financial Crisis.” Economic Policy, 26(66), 289-346.
- Obstfeld, M. (2012). “Financial Flows, Financial Crises, and Global Imbalances.” Journal of International Money and Finance, 31(3), 469-480.
- Presbitero, A. F., Ghura, D., Adedeji, O. and Njie, L., (2015). “Sovereign Bond Issuances by Developing Countries: Determinants of Market Access and Bond Spreads.” IMF Working Paper (forthcoming).
- Qureshi, M. and Sugawara, N., (2015), “Managing Capital Flows: Lessons from Emerging Markets for Frontier Economies.” Unpublished.
- Reinhart, C.M., and Reinhart, V.R. (2008). “Capital Flow Bonanzas: An Encompassing View of the Past and Present.” NBER Working Paper w14321.
- Williamson, O.E. (1985). The Economic Institutions of Capitalism. New York, NY: Simon and Schuster.
- World Bank (2011). World Development Report 2011 on Conflict, Security and Development. Washington, DC: Author.

### Appendix A — The use of BPM6 and WEO data
- IMF Statistics Department began publishing BPM6-based data in IFS and online BOPS since August 2012.
- Main BPM6 differences relevant to the financial account:
  - Gross flows in Direct Investment are affected (net flows should not suffer major modifications).
  - Changes in the sign of liabilities: in BPM6 increases of assets and liabilities are shown with positive signs and decreases with negative signs.
- Two reasons for opting for BPM5 in this project:
  - (i) The series published according to the new methodology usually are shorter than the ones in BPM5 and, hence, a conversion methodology would be necessary to obtain longer series.
  - (ii) At the time of construction of this database, the World Economic Outlook (WEO) database was presented according to BPM5 methodology and thus comparison with BPM5 BOPS was direct.
- Caveat: countries that already transitioned to BPM6 stopped publishing BPM5 data, creating missing observations especially for the years of 2011 and 2012.
- Compatibility exercise and classification of observations after conversion attempts:
  - c: Series are compatible with minor modifications. The differences may extend to more years than 2011 and 2012 but the general dynamics of the flows are not compromised. Those series can be used to fill missing values in the original database.
  - M: Major differences between versions. BPM6 cannot be used to fill gaps for those countries and series; gaps are maintained.
  - d: Series that have minor differences but were not necessary to convert because the original database is complete for that country and category.
- Countries that already made the transition (35 countries): Angola, Armenia, Australia, Azerbaijan, Bangladesh, Belarus, Belize, Benin, Bermuda, Bosnia and Herzegovina, Burundi, Canada, Chile, China, P.R.: Hong Kong, Fiji, Finland, Georgia, India, Jamaica, Kosovo, Kuwait, Malawi, Malaysia, Nicaragua, Philippines, Russian Federation, Samoa, Saudi Arabia, Seychelles, Singapore, Solomon Islands, Thailand, Ukraine, United States, Vanuatu.
- WEO data are used to supplement or replace IFS data in specified categories for listed countries. Cases where WEO supplements/replaces IFS (not including substitution of “n.a” for zeros):
  - Foreign Direct Investment Abroad: Bhutan, Central African Republic, Chad, Kuwait, Madagascar, Mauritania, Nepal, Nigeria, Qatar, Serbia, Suriname, UAE, Uzbekistan, Zimbabwe.
  - Direct Investment In the Country: Bhutan, Central African Republic, Chad, Comoros, Equatorial Guinea, Ecuador, Eritrea, Guyana, Mauritania, Qatar, Serbia, Suriname, Turkmenistan, UAE, Uzbekistan, Vietnam.
  - Portfolio Assets: Dem. Rep. of Congo, Cape Verde, Central African Republic, Chad, The Gambia, Qatar, Thailand.
  - Portfolio Liabilities: Chad, Ghana, Malaysia, Niger, Qatar, Thailand, Turkmenistan, Uzbekistan, Zimbabwe.
  - Other Investments Assets: Central African Republic, Chad, Equatorial Guinea, Gabon, Madagascar, Malaysia, Mauritania, Qatar, Senegal, Turkmenistan, UAE, Zambia.
  - Other Investments Liabilities: Burkina Faso, Chad, Comoros, Democratic Republic of Congo, Equatorial Guinea, Eritrea, Gabon, Georgia, Madagascar, Malaysia, Mauritania, Qatar, São Tomé e Príncipe, Seychelles, Turkmenistan, UAE, Zambia.

### Appendix B — Comparison with BIS data (notes and methodology)
- Figures compare "Other Investments Private Inflows" from the extended LIC database with BIS "Loans and deposits" (Locational Banking Statistics, table 7A).
- From extended LIC database: other investments private inflows = other investments minus proxy for other investments official flows. For outflows, other investments assets are used (official outflows expected minimal).
- From BIS database: table 7A refers to “Loans and deposits”, comprising interbank deposits and loans and advances, and trade credit — closely related to “other investments” of the BoP. Counterpart is “all sectors” (banks, non-bank financial institutions, non-financial sectors).
- BIS flows of assets and liabilities are calculated by difference in stocks and adjusted by exchange-rate change using the currency breakdown. Result is not identical to actual flow data due to timing of transactions and valuation effects.
- Figures illustrate panels for Cambodia, Kenya, Madagascar, Mongolia, Nicaragua, Nigeria, Senegal, Vietnam, Zambia (notes preserved from source).

### Appendix C — Comparison with FFA dataset (key comparative statistics)
- Overall data point coverage of the extended database is higher in almost every category of the financial account for LIDCs.
- Aggregate measure of non-FDI private inflows: increased coverage is 19 percent for the LIDC group.
- Examples of country-level comparisons depicted in Figure 12 for non-FDI private inflows (Percent of GDP): Bhutan, Cameroon, Comoros, Congo DR, Eritrea, Madagascar, Mauritania, Uzbekistan, Zambia (graphical comparisons shown in source).
- Range of additional coverage across components for LIDCs:
  - Official Other Investment Flows: 8.4 percent
  - Other Investments Liabilities: 13.4 percent
- Additional coverage for entire sample of LIDCs:
  - Other investments: 9.1 percent
  - Portfolio assets: 8.1 percent
- Coverage increases for direct investment components:
  - Direct investment into the country: 16.1 percent
  - Direct investment abroad: 1.7 percent
- Methodological note: comparison used a conservative approach by comparing only non-zero values in each individual series to avoid counting missing variables mistakenly replaced by zeros. For aggregate measures, aggregates computed whenever at least one subcomponent available.

### Appendix D — Sample of countries (classification markers)
- The source lists countries with markers for LIC, Small, Fragile. Notes:
  - a: Data were not available for Marshall Islands, Micronesia, Palau and Tuvalu.
  - b: Marshall Islands, Micronesia, Somalia, Tuvalu are not included in the database. West Bank and Gaza and Western Sahara are not part of the Fund.
- Country listing and markers are presented in the source (multiple pages listing countries and the markers “x” as applicable).

### Appendix E — Net flows to selected LIDCs (figure notes)
- Figures 13 and 14 present Net Non-FDI Private Flows to selected LIDCs (Percent of GDP) for years 2003–2012.
- Countries shown (with Gross Inflows, Gross Outflows, Net Inflows series): Cambodia, Kenya, Madagascar, Mongolia, Nicaragua, Nigeria, Senegal, Vietnam, Zambia.
- Source for these figures: Extended LIDC database.

*Source: _wp1586 - REFERENCES (IMF PDF content provided).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1586.pdf_