## _wp05224

## Source details

**Canonical URL:** [_wp05224](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp05224.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp05224.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp05224.pdf.json)

---

### Introduction and scope
- Focus: macroeconomic factors explaining dynamics of remittances to India (1990s–early 2000s).
- Contextual points:
  - Remittances to developing countries exceed official aid flows and most private capital flows (GDF, 2003).
  - Remittances to India more than quadrupled between 1991 and 2003 and totaled about US$18 billion by 2003.
  - Remittances increased at about 13 percent a year since 1991 and have shown low movement around the trend (characterized as the most stable type of external flows in India).
- Paper scope restriction: does not analyze macroeconomic impact or welfare implications in detail; focuses on determinants and time-series behavior.

### Measurement issues and magnitude
- Data sources:
  - Reserve Bank of India (RBI) databases—Handbook of Indian Economy and RBI Bulletins (RBI data available 1990–2003).
  - IMF Balance of Payments Statistics (BOPS) (IMF data available 1975–2002).
- Definitions and classification notes:
  - “Worker’s remittances” per IMF BOP manual: transfers by migrants employed in new economies and considered residents there (stayed a year or more).
  - “Other transfers”: charitable and religious contributions, gifts, etc.
- Data comparability and discontinuities:
  - Until 1999 IMF “other transfers” roughly matched RBI “gold imports”; “worker’s remittances” matched other RBI components. Divergence appears from 1999 onward.
  - Due to lack of clarity/discontinuity in disaggregated series, analysis focuses on total remittances using IMF data (total transfers of RBI and IMF match well).
- Informal channels and measurement caveats:
  - Reported remittances likely underestimate actual flows because of informal “hawala” transfers.
  - Estimate cited: hawala remittances to India about US$6 billion a year (Reddy (1997)).
  - Changes in costs or regulation of hawala can produce discrete jumps in reported official remittances (notable episodes include early 1990s exchange rate realignment and post-September 11, 2001 clampdowns).
- Magnitude and trend statistics:
  - Remittances more than quadrupled between 1991 and 2003, totaling about US$18 billion in 2003.
  - Remittances equaled about 3 percent of GDP in 2003.
  - India accounted for about 10 percent of total remittances to developing countries and about 25 percent of total remittances to Asian countries in 2002.
  - Remittances accounted for about ½ of receipts on invisibles and 20 percent of total receipts in the current account in 2002.
  - Remittances deemed substantially less volatile than exports of goods and services, NRI deposits, or portfolio flows.

### Literature and conceptual determinants
- Motivations and typical empirical findings from prior literature:
  - Altruistic motives (support family consumption) versus investment motives (investing, repaying loans).
  - Typical findings: remittances motivated more by altruism than investment; remittances are counter-cyclical; remittances used more for consumption than investment; remittances do not respond strongly to relative rates of return at home.
- Determinants considered in this paper’s framework:
  - Income of migrant (earnings abroad), economic conditions in native country, returns on domestic and foreign investments, risk of default proxied by political/geopolitical risk or rating downgrades.
  - Host-country proxies: U.S. nonagricultural employment, LIBOR, oil prices, NASDAQ returns.
  - Home-country variables: industrial growth, drought dummy (agricultural growth negative), BSE returns, agricultural/GDP growth.
  - Risk/policy dummies: rating downgrades, government resignations, geopolitical tensions with Pakistan, Asian crisis dummy, post-September 11, 2001 dummy, festival/wedding-season dummies (Oct–Dec, Apr–Jun).
  - Lagged exchange rate depreciation used due to potential endogeneity between depreciation and remittances.

### Bivariate correlations and time-series properties
- Bivariate associations:
  - HP-filtered remittances show significant correlation primarily with U.S. nonagricultural employment (correlation 0.24; threshold note: coefficient higher than 0.23 is significant at 10 percent).
  - Little or no significant bivariate correlation with most other variables in Table 2.
- Unit root and stationarity findings (examples reported):
  - Remittances (in constant US$): D-F -2.38, P value 0.39.
  - HP filtered Remittances: D-F -4.2, P value .00.
  - Earnings of migrants (in constant US$): D-F -2.69, P value 0.24.
  - US nonagricultural employment (levels): D-F -2.34, P value 0.40.
  - Oil prices (percent change): D-F -4.03, P value 0.00.
  - Nasdaq (percent change): D-F -3.03, P value 0.03.
- Conclusions on integration:
  - Remittances in levels are I(1); migration-related variables also I(1); no cointegration found between remittances and migration variables.
  - Variables measured in first differences, percent changes, or HP deviations were generally I(0).

### Econometric specification and main empirical results
- Regression approaches:
  - Equation (1): regress remittances (levels, constant US$) on linear trend and set of explanatory variables (lagged I(1) variables included).
  - Equation (2): regress HP-filtered (detrended) remittance series on explanatory variables (estimate cyclical component).
  - Estimation methods: OLS with heteroskedasticity-consistent standard errors; AR(1) specification when autocorrelation present.
- Trend (level) regression highlights (Table 5):
  - Strong linear trend: fitted trend y = 0.03 T + 6.75, R^2 = 0.80 (Figure 6).
  - Migration-related variables (number of migrants to the United States; total earnings of Indians in the United States) help explain the long-term trend; migration/earnings variables significant in some specifications (columns I–IV).
  - Oil prices insignificant in trend regressions.
  - Lagged dependent variable example coefficient .67*** in one specification.
  - Sample sizes reported: e.g., 53 observations (R^2 = .84, Adj. R^2 = .83) in one specification; smaller samples (e.g., 29, 28) when migration/earnings variables used.
- Cyclical component (HP-filtered) regression highlights (Table 6):
  - U.S. nonagricultural employment (percent change) positive and significant: coefficient .13** (t-stat 2.04) in column I.
  - Drought dummy positive and sometimes significant: coefficient .082* (t-stat 1.89) in column I.
  - Change in LIBOR positive and marginally significant in some specifications: coefficient .10* (t-stat 1.74).
  - Asian crisis dummy negative and significant: coefficients -.11** (t-stat -2.12) and -.16* (t-stat -1.67) across specifications.
  - Oil price changes not found to significantly affect remittances.
  - No significant effects detected for rating changes, BSE returns, exchange rate changes, or post-September 11, 2001 dummy.
  - Model fit examples: Column I — 53 observations, R^2 .13 (Adj. R^2 .06); Column II — 54 observations, R^2 .16 (Adj. R^2 .07).

### Interpretation of findings
- Drivers of long-term increase:
  - Increase in number of migrants to countries such as the United States, Australia, and Canada; Indians in the United States doubled during the 1990s (Desai, Kapur, and McHale (2001) cited).
  - Migration skewed toward skilled professionals with higher average earnings, consistent with rising measured remittances.
  - Institutional changes (exchange rate devaluations, capital account opening in early 1990s, reductions in import duty on gold) may have shifted flows from informal channels and illegal gold imports into official remittance channels.
- Cyclical behavior:
  - Remittances are higher when economic conditions abroad are benign—consistent with positive association with U.S. employment and other host-country indicators.
  - Some evidence of mild countercyclicality with respect to domestic agricultural shocks (droughts), but this result is weaker and not robust across all specifications.
  - Negative effect during the Asian crisis period possibly due to uncertainty or diversion of funds; interest-rate explanations for diversion to NRI deposits not strongly supported.
- Volatility and stability:
  - Remittances characterized as one of the most stable external flows for India, less volatile than portfolio flows, NRI deposits, and exports.

### Policy implications and recommendations
- Given sustained remittance inflows and their stability, external-sector policies may need to adapt to absorb these flows without creating excessive domestic liquidity or inflationary pressures.
- Suggested policy stance:
  - Allow imports to rise commensurately with incremental annual remittance inflows to alleviate inflationary pressures and aid liquidity management.
  - Such an adjustment may be welfare enhancing overall, since a portion of remittances filters out via higher imports (net impact on current account smaller than gross remittances).

### Limitations and suggestions for further research
- The paper does not study the macroeconomic impact or welfare implications of remittances in depth.
- End-use of remittances (consumption vs. investment) not directly analyzed for India in this study; disaggregated data would be useful to assess contributions to output, employment, and development outcomes.
- Measurement caveats: possible underestimation due to informal channels (hawala) and data discontinuities in disaggregated IMF/RBI series after 1999.

*Source: _wp05224; canonical URL: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp05224.pdf*

### 1. IMF and the RBI Data.................................................................................................

### 1. IMF and the RBI Data.................................................................................................. 5

### 2. Remittances (in Millions of U.S. Dollars) ................................................................... 5
- Section present in the content unit with the exact heading and page reference "5".
- Focus area: remittances measured in millions of U.S. dollars.

### 3a. Annual Remittances to India........................................................................................ 7
- Section present in the content unit with the exact heading and page reference "7".
- Scope: annual remittance flows to India.

### 3b. Remittances (In Current, Constant Millions of U.S. Dollars)...................................... 7
- Section present in the content unit with the exact heading and page reference "7".
- Distinguishes current and constant dollar measures of remittances.

### 4. Remittances to Selected Countries............................................................................... 8
- Section present in the content unit with the exact heading and page reference "8".
- Comparative remittance data across selected countries.

### 5. Remittances and Current Account Balance ................................................................. 8
- Section present in the content unit with the exact heading and page reference "8".
- Examines relationship between remittances and the current account balance.

### 6. Trend Growth of Remittances.................................................................................... 14
- Section present in the content unit with the exact heading and page reference "14".
- Addresses trend growth analysis of remittance flows.

### 7. Total Migration to the United States and Remittances .............................................. 14
- Section present in the content unit with the exact heading and page reference "14".
- Links U.S. migration totals with remittance patterns.

### 8. Total Earnings of Indians in the United States and Remittances............................... 14
- Section present in the content unit with the exact heading and page reference "14".
- Considers earnings of Indian migrants in the U.S. and associated remittances.

### Tables (listed in the content unit)
- Table 1. Different Components of the Remittances Data in the RBI and IMF Databases ........ 6
- Table 2. Correlation Coefficients Between Dependent and Independent Variables................ 11
- Table 3. Quarterly Average of Remittances During Specific Periods ..................................... 12
- Table 4. Results of Unit Root Tests ......................................................................................... 13
- Table 5. Regression Results for the Level of Remittances ...................................................... 15
- Table 6. Regression Results for the Cyclical Component of Remittances .............................. 16

### Annex
- Annex I. Data Sources and Definitions..................................................................................... 18

*Source: _wp05224 - 1. IMF and the RBI Data.................................................................................................; canonical URL: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp05224.pdf*

### References..............................................................................................................

### _wp05224 - References.............................................................................................................................. 19

### Introduction and scope
- Focus: macroeconomic factors explaining dynamics of remittances to India (1990s–early 2000s).
- Key contextual points:
  - Remittances to developing countries exceed official aid flows and most private capital flows (GDF, 2003).
  - Remittances to India more than quadrupled between 1991 and 2003 and totaled about US$18 billion by 2003.
  - Remittances increased at about 13 percent a year since 1991 and have shown low movement around the trend (characterized as the most stable type of external flows in India).
- Paper does not analyze macroeconomic impact or welfare implications in detail; it focuses on determinants and time-series behavior.

### Measurement issues and magnitude
- Data sources:
  - Reserve Bank of India (RBI) databases—Handbook of Indian Economy and RBI Bulletins (RBI data available 1990–2003).
  - IMF Balance of Payments Statistics (BOPS) (IMF data available 1975–2002).
- Definitions:
  - “Worker’s remittances” per IMF BOP manual: transfers by migrants employed in new economies and considered residents there (stayed a year or more).
  - “Other transfers”: charitable and religious contributions, gifts, etc.
- Data comparability and breaks:
  - Until 1999 IMF “other transfers” roughly matched RBI “gold imports”; “worker’s remittances” matched other RBI components. A divergence appears from 1999 onward.
  - Due to lack of clarity/discontinuity in disaggregated series, analysis focuses on total remittances using IMF data (total transfers of RBI and IMF match well as per Figure 2).
- Informal channels:
  - Reported remittances likely underestimate actual flows because of informal “hawala” transfers.
  - Estimate cited: hawala remittances to India about US$6 billion a year (Reddy (1997)).
  - Changes in costs or regulation of hawala can produce discrete jumps in reported official remittances (notable episodes include early 1990s exchange rate realignment and post-September 11, 2001 clampdowns).

- Magnitude and trend statistics:
  - Remittances more than quadrupled between 1991 and 2003, totaling about US$18 billion in 2003.
  - Remittances equaled about 3 percent of GDP in 2003.
  - India accounted for about 10 percent of total remittances to developing countries and about 25 percent of total remittances to Asian countries in 2002.
  - Remittances accounted for about ½ of receipts on invisibles and 20 percent of total receipts in the current account in 2002.
  - Remittances deemed substantially less volatile than exports of goods and services, NRI deposits, or portfolio flows.

### Literature and conceptual determinants
- Motivations identified in literature:
  - Altruistic motives (support family consumption) versus investment motives (investing, repaying loans).
- Typical empirical findings from prior literature:
  - Remittances motivated more by altruism than investment.
  - Remittances are counter-cyclical (higher under adverse domestic outcomes).
  - Remittances used more for consumption than investment.
  - Remittances do not respond strongly to relative rates of return at home.
- Determinants considered in this paper’s framework:
  - Income of migrant (earnings abroad), economic conditions in native country, returns on domestic and foreign investments, risk of default proxied by political/geopolitical risk or rating downgrades.
  - Host-country proxies used: U.S. nonagricultural employment, LIBOR, oil prices, NASDAQ returns.
  - Home-country variables: industrial growth, drought dummy (agricultural growth negative), BSE returns, agricultural/GDP growth.
  - Risk/policy dummies: rating downgrades, government resignations, geopolitical tensions with Pakistan, Asian crisis dummy, post-September 11, 2001 dummy, festival/wedding-season dummies (Oct–Dec, Apr–Jun).
  - Lagged exchange rate depreciation used due to potential endogeneity between depreciation and remittances.

### Bivariate correlations and time-series properties
- Bivariate associations:
  - Remittances (HP filtered) show significant correlation primarily with U.S. nonagricultural employment (correlation 0.24; threshold note: coefficient higher than 0.23 is significant at 10 percent).
  - Little or no significant bivariate correlation with most other variables in Table 2.
- Time-series properties:
  - Remittances in levels found to be I(1); migration-related variables also I(1); no cointegration found between remittances and migration variables.
  - Variables measured in first differences, percent changes, or HP deviations were generally I(0).
  - Unit root test examples (D-F statistics and p-values as reported):
    - Remittances (in constant US$): D-F -2.38, P value 0.39.
    - HP filtered Remittances: D-F -4.2, P value .00.
    - Earnings of migrants (in constant US$): D-F -2.69, P value 0.24.
    - US nonagricultural employment (levels): D-F -2.34, P value 0.40.
    - Oil prices (percent change): D-F -4.03, P value 0.00.
    - Nasdaq (percent change): D-F -3.03, P value 0.03.

### Econometric specification and main empirical results
- Regression specifications:
  - Equation (1): regress remittances (levels, constant US$) on linear trend and set of explanatory variables (lagged I(1) variables included to address nonstationarity).
  - Equation (2): regress HP-filtered (detrended) remittance series on explanatory variables (estimate cyclical component).
  - Estimation methods: OLS with heteroskedasticity-consistent standard errors; AR(1) specification when autocorrelation present.
- Trend (level) regressions (Table 5) — main points:
  - Strong linear trend in remittances: Figure 6 displays a fitted trend y = 0.03 T + 6.75, R^2 = 0.80.
  - Migration-related variables (number of migrants to the United States; total earnings of Indians in the United States) help explain the long-term trend in remittances (columns I–IV indicate migration/earnings variables are significant in some specifications).
  - Oil prices insignificant in trend regressions.
  - Lagged dependent variable often included (example coefficient .67*** in one specification).
  - Sample sizes reported: e.g., 53 observations (R^2 = .84, Adj. R^2 = .83) in one specification; smaller samples (e.g., 29, 28) when migration/earnings variables used.

- Cyclical component regressions (HP-filtered, Table 6) — key results:
  - Positive and significant association between remittances (HP filtered) and U.S. nonagricultural employment (percent change): coefficient .13** (t-stat 2.04) in column I.
  - Drought dummy positive and sometimes significant: coefficient .082* (t-stat 1.89) in column I; weaker in other specifications.
  - Change in LIBOR positive and marginally significant in some specifications: coefficient .10* (t-stat 1.74).
  - Asian crisis dummy negative and significant: coefficients -.11** (t-stat -2.12) and -.16* (t-stat -1.67) across specifications.
  - Oil price changes not found to significantly affect remittances.
  - No significant effects detected for rating changes, BSE returns, exchange rate changes, or post-September 11, 2001 dummy.
  - Reported model fit examples: Column I — 53 observations, R^2 .13 (Adj. R^2 .06); Column II — 54 observations, R^2 .16 (Adj. R^2 .07).

### Interpretation of findings
- Structural explanation for long-term buoyancy:
  - Increase in number of migrants to countries such as the United States, Australia, and Canada; in particular Indians in the United States doubled during the 1990s (Desai, Kapur, and McHale (2001) cited).
  - Migration skewed toward skilled professionals with higher average earnings, consistent with rising measured remittances.
  - Institutional changes: exchange rate devaluations and capital account opening in early 1990s, and reductions in import duty on gold, may have shifted flows from informal channels and illegal gold imports into official remittance channels.
- Cyclical behavior:
  - Remittances are higher when economic conditions abroad (source/host countries) are benign — consistent with positive association with U.S. employment and other host-country indicators.
  - Some evidence of mild countercyclicality with respect to domestic agricultural shocks (drought periods), but this result is weaker and not robust across all specifications.
  - Negative effect during Asian crisis period possibly due to uncertainty or diversion of funds into RIB/IMD bonds (Resurgent Indian Bond issuance contemporaneous), but interest-rate explanations for diversion to NRI deposits not strongly supported by the data.
- Volatility and stability:
  - Remittances characterized as one of the most stable external flows for India, less volatile than portfolio flows, NRI deposits, and exports.

### Policy-relevant implications and recommendations (as discussed)
- Given sustained remittance inflows and their stability, external-sector policies may need to adapt to absorb these flows without creating excessive domestic liquidity or inflationary pressures.
- Suggested policy stance:
  - Allow imports to rise commensurately with incremental annual remittance inflows to alleviate inflationary pressures and aid liquidity management.
  - Such an adjustment may be welfare enhancing overall, especially since a portion of remittances filters out via higher imports (net impact on current account smaller than gross remittances).

### Limitations and suggestions for further research
- The paper does not study the macroeconomic impact or welfare implications of remittances in depth.
- End-use of remittances (consumption vs. investment) not directly analyzed for India in this study; disaggregated data would be useful to assess contributions to output, employment, and development outcomes.
- Measurement caveats: possible underestimation due to informal channels (hawala) and data discontinuities in disaggregated IMF/RBI series after 1999.

*Source: IMF working paper content (References section and accompanying pages as supplied). *

### REFERENCES

### _wp05224 - REFERENCES

### Working papers and IMF publications
- Chami, R., C. Fullenkamp, and S. Jahjah, 2003, “Are Immigrant Remittance Flows a Source of Capital for Development?,” IMF Working Paper 03/189 (Washington: International Monetary Fund).
- Gordon, James P., and Poonam Gupta, 2003, “Portfolio Flows to India: Do Domestic Fundamentals Matter?” IMF Working Paper 02/42 (Washington: International Monetary Fund).
- International Monetary Fund, Balance of Payments Manual, Fifth Edition (Washington).

### Journal articles and academic studies
- Desai, M.A., D. Kapur, and J. McHale, 2001, “The fiscal Impact of the Brain Drain: Indian Emigration to the U.S.” Weekly Political Economy Discussion Paper (Massachusetts: Harvard University).
- El-Qorchi, M. 2002, “Hawala,” Finance and Development, Vo. 39, No. 4 (December) (Washington: International Monetary Fund).
- ———, 2004, “Non-Resident Deposits in India: In Search of Return?,” Economic and Political Weekly, Vol. 39, No. 37, pp. 4165 (September 19).
- Lucas, R.E.B., and O. Stark, 1985, “Motivations to Remit: Evidence from Botswana,” Journal of Political Economy, Vol. 93, No. 5, pp. 901–18.
- Straubhaar, T., 1986, “The Determinants of Worker’s Remittances: The Case of Turkey,” Weltwirtschaftsliches Archive, Vol. 122, No. 4, pp. 728–40.

### Reports, chapters, addresses, and conference papers
- Global Development Finance, 2003, “Worker’s Remittances: An Important and Stable Source of External Development Finance,” Chapter 7, pp. 157–75 (Washington: World Bank).
- Jadhav, Narendra, 2003, “Maximizing Developmental Benefits of Migrant Remittances: The Indian Experience,” paper presented at the joint conference of DFID-World Bank, London.
- Y.V. Reddy, 1997, “Capital Flight: Myths and Realities,” address at Center for Economic and Social Studies, Hyderabad, June 21.
- Reserve Bank of India, 2003, “Current Account Dynamics in an Open Economy,” Report on Currency and Finance, Chapter 5, pp 139–41.

*Source: _wp05224 - REFERENCES*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp05224.pdf_