Non-FDI Capital Inflows in Low-Income Countries: Catching the Wave?

Abstract

Low-income countries (LICs) are typically characterized by intermittent and very modest access to private external funding sources. Motivated by recent developments in private flows to these economies, this paper makes two contributions: first, it constructs a new comprehensive dataset on gross private capital flows with special focus on non-FDI flows to LICs. Concentrating on LICs and more specifically on gross non-FDI private flows is intentionally aimed at closing a gap in existing datasets where country coverage of developing economies is limited mainly to emerging markets (EMs). Second, using the new data, it identifies several shifting patterns of gross non-FDI private inflows to LICs. A surprising fact emerges: since the mid-2000s periods of surges in gross non-FDI private inflows to LICs are broadly comparable to those of EMs. Moreover, while gross non-FDI inflows to LICs are on average much lower than those to EMs, we show that gross non-FDI inflows to the top quartile of LICs are comparable to those of the median EM and converging to the top quartile of EMs.

Appendices

Appendix A: Database

1.1 A.1: Description

The database contains the following variables: Direct Investment Abroad, Direct Investment In Country, Portfolio Investment Assets, Portfolio Investment Liabilities, Portfolio Equity Assets, Portfolio Equity Liabilities, Portfolio Debt Assets, Portfolio Debt Liabilities, Financial Derivatives, Other Investment Assets, Other Investment Liabilities, Proxy for Official Other Investment Liabilities, Debt Forgiveness, Private Inflows Excluding Direct Investment (Portfolio Liabilities + Financial Derivatives + Other Investment Liabilities - Proxy for Official Other Investment Liabilities) and Private Outflows Excluding Direct Investment (Portfolio Assets + Other Investment Assets). Nominal GDP and Private Inflows excluding Direct Investment as a percentage of GDP and Private Outflows Excluding Direct Investment as a percentage of GDP are also included.

The database contains data for the period of 1990–2014 including 183 countries (58 Low-Income Developing Countries, 92 Other Developing Countries and 33 Advanced Economies). Advanced economies original BOPS data are included for completeness of the database. More detailed information, including the specific years when WEO data were used for each class of investment and each country, is available in the tab MetaData of the database. Except when indicated, data source for every country and component is the Balance of Payments Statistics (BOPS), IMF’s Balance of Payments and International Investment Position  Manual 6 (BPM6). Years and/or time intervals (for example: 2001 and/or 2008–2010) indicate to the use of the World Economic Outlook (WEO) data series for the specified period. Besides the detailed information on the data used to fill the gaps by individual country and variable, the tab MetaData also contains:

• IFS code: country code.

• Type: 1 for Advanced Economies, 2 for Other Developing Countries and 3 for Low-Income Developing Countries.

• Classification BFOLG and BFOL_G: Indicates correlation (High, Medium, Low) between “Official liabilities” and “Liabilities to official creditors” of WEO and BOPS series (BFOLG and BFOL_G). It does not include advanced economies.

• Notes: country notes.

• Nominal GDP: Gross domestic product, current prices in U.S. dollars (Source WEO; Gross domestic product, current prices in national currency over National currency units per U.S. dollar, period average).

Figure 10 contains a snapshot for the line on the Democratic Republic of Congo.

Figure 10

Snapshot of MetaData

1.2 A.2: The Use of BPM6 and WEO Data

The IMF Statistical Department started publishing data according to the methodology presented in the Sixth Edition of the IMF’s Balance of Payments and International Investment Position Manual (BPM6) in the IMF’s International Financial Statistics (IFS) and in the online Balance of Payments Statistics (BOPS) databases since August 2012. It is not the scope of this annex to explain the new methodology and the differences between the versions; hence we refer the interested reader to the relevant manuals. With respect to the financial account, the main differences refer to gross flows in Direct Investment (although net flows should not suffer major modifications) and changes in the sign of liabilities (negative to positive).

The original version of our database was based on the methodology presented in the Fifth Edition of the IMF’s Balance of Payments and International Investment Position Manual (BPM5). The main reason for the use of BPM5 at the time was that the World Economic Outlook (WEO) database was still presented according to BPM5 methodology and thus comparison with BOPS BPM5, a resource heavily used in this project, was direct. Additionally, the series published according to the new methodology were usually shorter than the ones in BPM5 implying that some kind of conversion would be necessary to obtain longer series.

In order to make the process of updating the dataset easier in the future, an effort was made to convert the dataset to BPM6 and to extend coverage from 2012 to 2014. Two events helped support the conversion process of our database: WEO publication on a BPM6 basis started in October 2014; and in early 2016, the IMF Statistical Department released the entire historical series of Balance of Payments Statistics on a BPM6 basis by using a standardized conversion tool.³¹ Currently, as shown in Figure 11, a large share of countries have already migrated to the BPM6 methodology, and in those cases, publication of BPM5 updated series have been discontinued.³² For countries where data are still collected using BPM5 standard, the series are converted and published by the IMF Statistical Department in BPM6 facilitating future updates of the database. It is important to note that most series present breaking points between the data originally collected using BPM6 standard and data collected using BPM5 standard and converted by the IMF Statistical Department.

Figure 11

Source: IMF BOPS. The sample includes 33 advanced markets (AM), 79 emerging markets (EM) and 70 low-income countries (LIC). Migration date as of December 2015.

Migration of Countries to BPM6 (percent of each group)

In what follows, we describe the methodology used to convert our database from BPM5 to BPM6. First, we identified when data were collected according to BPM6 in each country series. This was done by comparing the values of the BPM5 and BPM6 series backwards from 2012 year by year until 1990. The splicing point was identified as the first point where there was no significant difference (to the second digit to the right of the decimal mark) between BPM5 and BPM6 series. In the vast majority of cases, the splicing point corresponded to the year reported by the IMF Statical Department (listed in footnote 32) as the date up to when countries revised their series according to the new methodology. In that specific year, the new BPM6 series was connected to the BPM5 series. If no WEO data were used for the specific series, then the converted BPM5 series was used for the period prior to the splicing point. In most cases, series in the BPM5 standard and the converted BPM6 series are the same. Possible reasons for this are: information that could differentiate both series was not collected under the BPM5 standard and the differences between standards are not relevant at the level of aggregation of the database, specially because of the focus in the financial account.

For the cases where WEO data were used before the splicing point, the values were updated with WEO BPM6. In most cases, the WEO data were exactly the same as the one collected under BPM5. When WEO data were originally used after the new splicing point, if BOPS BPM6 data were no longer missing, then BOPS BPM6 series were left intact (i.e., no replacement by WEO data). In the case that gaps persisted in the BOPS BPM6, the same techniques described in Section II were implemented with the use of WEO BPM6.³³ The years of 2013 and 2014 were included in the database using BOPS BPM6. If BOPS BPM6 data were missing then WEO BPM6 was considered to fill the gaps (using the same techniques described in Section II). If WEO data were available for 2013–2014, but it was incompatible with BOPS, the gaps would remain. With the database converted to BPM6, BOPS BPM6 data can be used to update the extended LIC database. The cases when WEO BPM6 data were used for 2013–2014 are documented in the MetaData of the database and could be replaced by BOPS BPM6 data if available in the future.

A summary of the cases (specific years are reported in the MetaData of the database) where WEO data are used to supplement or replace BOPS data are (not including substitution of “n.a” for zeros): Foreign Direct Investment Abroad: Bhutan, Central African Republic, Chad, Gambia, Madagascar, Mauritania, Nepal, Nigeria, Qatar, Serbia, Turkmenistan, Vietnam and Zimbabwe. Direct Investment In the Country: Bhutan, Central African Republic, Chad, Comoros, Dem.Rep. of Congo, Equatorial Guinea, Eritrea, Gabon, Mauritania, Qatar, Serbia, Turkmenistan, Vietnam and Zimbabwe. Portfolio Assets: Belize, Bhutan, Cape Verde, Central African Rep., Chad, Comoros, Rep. of Congo, Gambia, Haiti, Qatar, São Tomé e Príncipe, Turkmenistan, UAE and Yemen. Portfolio Liabilities: Angola, Chad, Mongolia, Mozambique, Qatar, São Tomé e Príncipe, Seychelles, Turkmenistan, UAE and Uzbekistan. Other Investments Assets: Central African Republic, Chad, Equatorial Guinea, Gabon, Malaysia, Mauritania, Qatar, Senegal, Turkmenistan, UAE, Uzbekistan and Zambia. Other Investments Liabilities: Chad, Comoros, Democratic Republic of Congo, Equatorial Guinea, Eritrea, Gabon, Georgia, Malaysia, Mauritania, Qatar, São Tomé e Príncipe, Senegal, Turkmenistan, UAE, Uzbekistan and Zambia. OtherGov: Brunei Darussalam, Chad, Comoros, Equatorial Guinea, Eritrea, Gabon, Malaysia, Mauritania, Saudi Arabia, Turkmenistan, Uzbekistan.

Appendix B: Comparison with BIS Data

See Figures 12 and 13.

Figure 12

Other Private Investments Inflows and BIS Loans (U.S. dollars in millions)

Notes: From the extended LIC database, Other Investments private inflows are Other Investments subtracted of the proxy for Other Investments official flows. In the case of outflows, we use Other Investments assets, since official outflows are expected to be minimal. From BIS database, we use Table 7A of the Locational Banking Statistics, consistent with the balance of payments statistics. This table refers to Loans and deposits, which comprise interbank deposits and loans and advances, besides trade credit. This is closely related to Other Investments of the BoP. The counterpart is all sectors, which includes banks, non-bank financial institutions and non-financial sectors. Flows of assets and liabilities are calculated by difference in stocks and adjusted by exchange-rate change. BIS uses the currency breakdown to calculate the exchange-rate value-adjusted changes. The result is not identical to actual flow data (because the actual transactions are distributed over the whole year during which exchange rates may fluctuate, and because other valuation effects may impact stocks). Please note that axes are not always aligned in zero.

Figure 13

Other Private Investments Inflows and BIS Loans (U.S. dollars in millions)

Notes: From the extended LIC database, Other Investments private inflows are Other Investments subtracted of the proxy for Other Investments official flows. In the case of outflows, we use Other Investments assets, since official outflows are expected to be minimal. From BIS database, we use Table 7A of the Locational Banking Statistics, consistent with the balance of payments statistics. This table refers to .Loans and deposits, which comprise interbank deposits and loans and advances, besides trade credit. This is closely related to Other Investments of the BoP. The counterpart is all sectors, which includes banks, non-bank financial institutions and non-financial sectors. Flows of assets and liabilities are calculated by difference in stocks and adjusted by exchange-rate change. BIS uses the currency breakdown to calculate the exchange-rate value-adjusted changes. The result is not identical to actual flow data (because the actual transactions are distributed over the whole year during which exchange rates may fluctuate, and because other valuation effects may impact stocks). Please note that axes are not always aligned in zero.

Appendix C: Comparison with FFA Dataset

We compare our dataset with the Financial Flows Analytics (FFA) dataset constructed by Bluedorn and others (2013). In general, data point coverage of our database is higher in almost every category of the financial account of the balance of payments for LICs. In terms of the aggregate measure of non-FDI private inflows, the increased coverage is 19 percent for the LIC group. For the components of this measure, the range of additional data for LICs varies from 8.4 percent (Official Other Investment Flows) to 13.4 percent (Other Investments Liabilities). In terms of the Other Investments and portfolio assets, the additional coverage is 9.1 and 8.1 percent, respectively, for the entire sample of LICs. In terms of direct investment into the country and direct investment abroad, coverage of our database is 16.1 percent and 1.7 percent greater for the group of LICs.³⁴

We provide some country examples in Figure 14 for the aggregate measure of non-FDI private inflows. Regarding the extended data, the range of cases goes from non available data from FFA (Uzbekistan) to only one observation missing for the last year (Cameroon). In some of the cases, there are large gaps in the middle of the series and mild discrepancies (Zambia). In the case of Madagascar and Mauritania, the data missing from FFA contains interesting patterns. The same is true for Bhutan, Comoros and Eritrea.

Figure 14

Source: Extended LIC database and financial flows analytics.

Comparison between Extended LICs Database and FFA—Non-FDI Private Inflows (percent of GDP)

Appendix D: Comparison with Alfaro and others (2014) Dataset

We compare our dataset with the dataset from Alfaro and others (2014). Because they use a variety of measures of capital flows as well as different samples, an explanation of each one is given below. The measures of capital flows are:

• Net capital flows (CA/GDP): 1980–2007 average of the annual current account balance with the sign reversed in current US dollars, normalized by GDP in current US dollars.

• Net FDI + portfolio flows/GDP: The 1980–2007 average of the annual flows of foreign liabilities minus annual flows of foreign assets in current US dollars, normalized by GDP in current US dollars. Annual flows are computed as the difference between FDI plus portfolio equity investment liability and asset flows in current US dollars from the IMF (under source “IMF”) or as annual changes in stocks of FDI plus portfolio equity investment liabilities minus annual changes in assets in current US dollars, adjusted for valuation effects normalized by nominal GDP in US dollars (under source “LM” for Lane and Milesi-Ferretti (2007), which includes valuation effects).

• Net FDI + portfolio + private debt flows/GDP: it adds the annual changes in stocks of total debt from private creditors (private nonguaranteed debt and public and publicly guaranteed debt from private creditors) from World Bank to net FDI + portfolio flows/GDP computed from the IMF or LM data.

The main samples are:

• 1. Raw World: 22 advanced OECD countries and all non-OECD countries where data on their current account balances and GDP per capita are available for more than 30% of the sample over 1980–2007.

• 2. Raw Developing: Excludes 22 advanced OECD countries from Raw World.

• 3. Developing: Raw Developing sample minus the outlier countries based on formal econometric outlier tests in terms of capital flows and growth.

• 4. Benchmark: Raw Developing sample minus the outlier countries based on formal econometric outlier tests in terms of the components of capital flows (equity and public and publicly guaranteed debt) and growth.

As detailed above, the samples from Alfaro and others (2014) are based on the availability of data as well as exclusion of outliers. The comparison with our database is not direct, since the measures of interest and data sources in both papers are distinct. The closest definition used by Alfaro and others (2014) to the measure of interest in our paper is Net FDI + portfolio + private debt flows/GDP. Table 4 reports the correlation of this measure with our measure of gross non-FDI private inflows. Table 5 reports the correlation of this measure with a constructed measure from our database of net private inflows Net FDI + portfolio + other private flows/GDP. Correlations are not strikingly high, reflecting in part the different categories included in our measure of other flows, and by the fact highlighted by Alfaro and others (2014) (technical appendix), that for developing countries there are discrepancies between the loan flows reported in the IMFs BOP statistics and the changes in the external debt stocks as reported by the WBs GDF database. We also compared if each observations is available in each dataset for all LICs and years between 1990 and 2014. Our database provides, respectively, 17.3, 17.3, 22.1 and 30% more observations for LICs with comparison to the samples in the order of the list above (from Raw World to Benchmark), although their database has many more observations for earlier periods as their coverage starts in 1980.

Table 4 Correlation between Non-FDI Private Inflows from Extended LIC Database and AKV Private Flows

Table 5 Correlation between net FDI $+$ non-FDI Private Flows from Extended LIC Database and AKV Private Flows

Appendix E: Sample of Countries

Country	LIC	Small$^\mathrm{a}$	Fragile$^\mathrm{b}$	Country	LIC	Small$^\mathrm{a}$	Fragile$^\mathrm{b}$
Afghanistan, I.R. of	x		x	Dominican Republic
Albania				Ecuador
Algeria				Egypt
Angola			x	El Salvador
Antigua and Barbuda		x		Equatorial Guinea
Argentina				Eritrea	x		x
Armenia				Estonia
Azerbaijan, Rep. of				Ethiopia	x
Bahamas, The		x		Fiji		x
Bahrain, Kingdom of				Gabon
Bangladesh	x			Gambia, The	x
Barbados		x		Georgia			x
Belarus				Ghana	x
Belize		x		Grenada		x
Benin	x			Guatemala
Bhutan	x	x		Guinea	x		x
Bolivia	x			Guinea-Bissau	x		x
Bosnia & Herzegovina			x	Guyana		x
Botswana				Haiti	x		x
Brazil				Honduras	x
Brunei Darussalam				Hungary
Bulgaria				India
Burkina Faso	x			Indonesia
Burundi	x		x	Iran, I.R. of
Cambodia	x			Iraq			x
Cameroon	x			Jamaica
Cape Verde		x		Jordan
Central African Rep.	x		x	Kazakhstan
Chad	x		x	Kenya	x
Chile				Kiribati	x	x	x
China,P.R.: Mainland				Kosovo, Republic of			x
Colombia				Kuwait
Comoros	x	x	x	Kyrgyz Republic	x
Congo, Dem. Rep. of	x		x	Lao People’s Dem.Rep	x
Congo, Republic of	x		x	Latvia
Costa Rica				Lebanon
Cote d’lvoire	x		x	Lesotho	x
Croatia				Liberia	x		x
Djibouti	x	x		Libya
Dominica		x		Lithuania
Macedonia, FYR				Senegal	x
Madagascar	x			Serbia, Republic of
Malawi	x			Seychelles		x
Malaysia				Sierra Leone	x		x
Maldives		x		Solomon Islands	x	x	x
Mali	x			South Africa
Mauritania	x			Sri Lanka
Mauritius			x	St. Kitts and Nevis		x
Mexico				St. Lucia		x
Moldova	x			St. Vincent & Grens.		x
Mongolia	x			Sudan	x		x
Montenegro		x		Suriname		x
Morocco				Swaziland		x
Mozambique	x			Syrian Arab Republic
Myanmar	x		x	Tajikistan	x		x
Namibia				Tanzania	x
Nepal	x		x	Thailand
Nicaragua	x			Timor-Leste		x	x
Niger	x			Togo	x		x
Nigeria	x			Tonga		x
Oman				Trinidad and Tobago		x
Pakistan				Tunisia
Panama				Turkey
Papua New Guinea	x			Turkmenistan
Paraguay				Uganda	x
Peru				Ukraine
Philippines				United Arab Emirates
Poland				Uruguay
Qatar				Uzbekistan	x
Romania				Vanuatu		x
Russian Federation				Venezuela, Rep. Bol.
Rwanda	x			Vietnam	x
Samoa		x		Yemen, Republic of	x		x
Sao Tome & Principe	x	x	x	Zambia	x
Saudi Arabia				Zimbabwe	x		x

1. ^aData were not avaiable for Marshall Islands, Micronesia, Palau and Tuvalu.
2. ^bMarshall Islands, Micronesia, Somalia, Tuvalu are not included in the database. West Bank and Gaza and Western Sahara are not part of the Fund.

Appendix F: Net Flows to Selected LICs

See Figures 15 and 16.

Figure 15

Source: Extended LIC database.

Net Non-FDI private Flows to Selected LICs (percent of GDP)

Figure 16

Source: Extended LIC database.

Net Non-FDI Private Flows to Selected LICs (percent of GDP)