## Outsourcing of Services

## Source details

**Canonical URL:** [Outsourcing of Services](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04186.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04186.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04186.pdf.json)

---

### Introduction and context
- Media attention surged in early 2004:
  - 2,634 reports in US newspapers on service outsourcing between January and May 2004.
  - 380 reports in United Kingdom newspapers during the same period.
  - 270 stories in a two-week period from March 1 to March 15, 2004, mentioning outsourcing and either job loss or unemployment in the same story.
- A quarterly FACTIVA index (first quarter 1991 to first quarter 2004) for the United States and the United Kingdom shows an upward trend in media interest in international outsourcing of services.
  - U.S. index local peaks: 1996, 2000, and 2004 (all presidential election years).

### Definitions and measurement of outsourcing
- Definitions used:
  - Outsourcing: procuring service or material inputs by a firm from a source in a foreign country (includes intra-firm and arm’s-length international outsourcing).
  - Offshoring: used synonymously with international outsourcing.
  - Insourcing: outsourcing in the opposite direction (from foreign-located firms to domestic firms).
- Two measures constructed:
  - Economy-wide measure: imports of computing (including computer software designs) and other business services using IMF’s Balance of Payments Statistics Yearbook.
  - Industry-level measure (United Kingdom): outsourcing intensity OSS_i constructed from input/output tables and IMF trade data; analogue OSM_i for material outsourcing.
- Measurement caveats:
  - Likely under-estimate outsourcing: import prices may be lower than domestic purchase prices and quantity data are unavailable.
  - Applying the same import share to all industries is a proxy due to lack of imports by industry.
  - Total input use excludes services produced within the same industry.

### Global and historical magnitudes and trends
- Economy-wide IMF-based shares for computing + other business services:
  - United States: 0.1 percent in 1983; 0.2 percent in 1993; 0.4 percent in 2003.
  - United Kingdom: 0.9 percent in 1983; 0.7 percent in 1993; 1.2 percent in 2003.
- Industry-level outsourcing intensity (output-weighted averages):
  - United Kingdom: service imports share 3.5 percent in 1992 and 5.5 percent in 2001.
  - United States (comparable): 0.4 percent in 1992 and 0.8 percent in 2001.
- Material versus service outsourcing:
  - Material outsourcing intensities: ~27 percent in the United Kingdom and ~12 percent in the United States.
  - Service outsourcing is much lower than material outsourcing but rising faster.

### Country-level outsourcing and insourcing (selected 2002 and comparable data)
- Top importers of business services (2002, Million US Dollars):
  - United States: US$41 billion
  - Germany: US$39 billion
  - Japan: US$25 billion
  - Netherlands: US$21 billion
  - Italy: US$20 billion
  - France: US$19 billion
  - United Kingdom: US$16 billion
  - India: US$11 billion (ranked 11th)
  - China: US$8 billion (ranked 18th)
- Top importers of computer and information services: Germany, United Kingdom, Japan, Netherlands, Spain; United States ranked 6th; China ranked 10th. (No IMF BOP data for India in this category.)
- Outsourcing intensity (imports of business services scaled by GDP, 2002):
  - Examples of very high ratios: Angola, Republic of Congo, Mozambique, Ireland, Vanuatu — imported business services exceeding 10 percent of GDP for some.
  - United States: outsourcing ratio in business services < 0.5 percent of GDP (ranked 117th).
  - United Kingdom: slightly over 1 percent of GDP (ranked 85th).
  - China: ranked 99th (ahead of United States).
  - India: imports of business services as a share of GDP = 2.4 percent.
- Insourcing (exports of business and computing services as proxy for receiving outsourcing, 2002):
  - Top recipients in dollar terms:
    - United States: US$59 billion
    - United Kingdom: US$37 billion
    - Germany: US$28 billion
    - France: US$21 billion
    - Netherlands: US$20 billion
    - India: US$18.6 billion (ranked 6th)
    - China: US$10 billion (ranked 14th)
  - Scaled by GDP, top insource-intensive economies include Vanuatu, Singapore, and Hong Kong SAR, each with exporting services > 10 percent of local GDP.
  - India: exports-as-share-of-GDP = 3.8 percent (noted as somewhat more insourcing-intensive than the United Kingdom by this metric).

### Who are the largest surplus countries (selected figures)
- Largest surplus countries in combined computing and business services: United Kingdom and United States.
  - The United States has run a surplus in this service category every year since 1980; the United Kingdom likewise.
  - If every country reduced its service outsourcing, the United States and the United Kingdom would be the biggest two losers in net dollars in service trade.
- Patterns by country:
  - Germany: small deficit in business services each year 1980–2001.
  - France: small surplus for much of the sample, switching to a mild deficit at the end.
  - India: imports and exports fairly balanced early; exports surge from 1996 producing a sizable surplus.
  - China: alternating import/export relative sizes; ends with a small surplus.
- Key table-reported figures:
  - Absolute outsourcers (Business Services, Million US Dollars): United States: 40,929; Germany: 39,113; Japan: 24,714; Netherlands: 21,038; Italy: 20,370.
  - Absolute insourcers (Business Services, Million US Dollars): United States: 58,794; United Kingdom: 36,740; Germany: 27,907; France: 20,864; Netherlands: 20,074.
  - Surplus and Deficit Countries, 2002 totals (examples): United Kingdom: 23,628.68; United States: 21,748.30; Hong Kong SAR: 15,663.41; India: 6,813.44; Singapore: 3,826.12.
  - Outsourcing intensity example: United Kingdom engages in about three times as much service outsourcing as a share of GDP (1.2 percent in 2001) as the United States (0.4 percent in 2001).

### Employment and productivity effects — literature synthesis
- Material outsourcing literature:
  - Yeats (2001): 30 percent of OECD exports of machinery and transport equipment were parts and components in 1995 (26 percent in 1978); U.S. share 36 percent in 1978 to 40 percent in 1995; Europe 26 percent to 28 percent; Japan 15 percent to 26 percent.
  - Egger and Egger (2001): international material outsourcing contributed 3.3 percent of real value added per low-skilled worker in the EU from 1993 to 1997 (negative short-run, positive long-run effects).
  - Feenstra and Hanson (1999): outsourcing contributed 17.5 to 40 percent of the increase in the non-production wage share over 1979 to 1990.
- Service outsourcing literature (thinner but growing; selected findings):
  - Mann (2004): international outsourcing of IT hardware lowered IT hardware prices by 10 to 30 percent; IT software outsourcing expected to have similar or larger effects.
  - Amiti and Wei (2004): service outsourcing positively correlated with labor productivity in the United States; material outsourcing insignificant in that study.
  - Gorg and Hanley (2003): positive impact of international outsourcing of services on productivity in Ireland’s electronics industry (1990–1995).
  - Girma and Gorg (2003): positive evidence of service outsourcing on labor productivity and total factor productivity in the United Kingdom (1980–1992) but could not separate domestic from foreign outsourcing.
- Management-consultant and policy-oriented studies:
  - McKinsey Global Institute (2003): projects job losses due to outsourcing from 2003 to 2015; methodology not fully transparent; concludes job losses from outsourcing are relatively trivial compared to typical business-cycle job losses.
  - Brainard and Litan (2004): emphasize distributional effects and provide policy prescriptions.
  - Shultze (2004): indirect evidence of job losses related to service outsourcing, concluding effects are very small.
  - Amiti and Wei (2004): using U.S. data, find a small negative effect of service outsourcing on employment with highly disaggregated data.

### Theoretical context and labor-market considerations
- Standard production-fragmentation models (Jones and Kierzkowski; Dixit and Grossman; Krugman and Venables; Deardorff; Yi; Amiti; Antras; Helpman) typically assume perfect inter-sectoral labor mobility and full employment, predicting no net job losses but changes in factor prices (Samuelson–Stolper effects).
- With short-run frictions (imperfect labor mobility, unions, participation responses), trade and outsourcing can yield net employment effects (Sachs and Shatz (1994); Krugman (1995)).
- McKinsey report: more than 69 percent of workers who lost jobs due to imports in the United States between 1979 and 1999 were re-employed, implying 31 percent were not re-employed.

### Does service outsourcing reduce jobs? — empirical approach and UK case study
- Empirical specification (first-differenced labor demand framework, Hamermesh 1993):
  - ∆ln Lit = α0 + α1 ∆ln wit + α2 ∆ln OSSit + α3 ∆ln OSMit + β ω + γ Dt + δ industry effects + εit
  - ∆ln OSSit = log difference in service outsourcing intensity (inverse proxy for price of imported service inputs).
- Data notes:
  - U.S. companion paper (Amiti and Wei, 2004): negative employment effect detected at 450-sector disaggregation (α2 < 0) but disappears at 96-sector aggregation.
  - United Kingdom case study: 1995–2001; 69 manufacturing industries and 9 service industries (78 sectors).
- UK descriptive evidence:
  - No uniform pattern between sectoral service outsourcing growth and employment growth (examples: “other transport equipment” high employment growth and high outsourcing growth; “preparation and spinning of textile fibers” negative employment growth and high outsourcing growth).
  - Scatter plot of service outsourcing growth vs. employment growth across 78 industries shows no clear one-way relationship.
- UK regression evidence — manufacturing:
  - One-period first-difference specifications: service outsourcing coefficients positive and sometimes statistically significant (e.g., ∆ ln(OSS)_t = 0.079*; ∆ ln(OSS)_t-1 = 0.063*; significance legend: * p < 0.1; ** p < 0.05; *** p < 0.01).
  - Two-period differenced specifications: significance of outsourcing coefficients reduced or eliminated (outsourcing effect weakly significant at 10 percent in one two-period specification and insignificant in others).
  - Wage effect: ∆ ln (wage) consistently negative and significant (examples: -0.742***, -0.683***).
  - Output effect: ∆ ln (real output) consistently positive and significant (examples: 0.444***, 0.461***).
  - Sample sizes and fit examples: N = 345, 276, 256, 276, 207, 192 across columns; R2 ranges from 0.12 to 0.76 in reported columns.
- UK regression evidence — services:
  - Mostly insignificant effects of service outsourcing on service-sector employment in partial-correlation specifications.
  - One two-period differenced specification reports a significant negative effect at the 5 percent level, but this effect is small and not robust to adding industry fixed effects (it disappears with industry-specific effects).
  - Wage and nominal output effects in services: ∆ ln (wage) often negative and significant (examples: -0.939***, -0.975***); ∆ ln (nominal output) often positive and significant (examples: 0.797***, 1.181***).
  - Reported R2 values in some services regressions up to 0.91 in a reported column.
- Main empirical conclusion (United Kingdom 1995–2001):
  - No robust evidence that service outsourcing led to net employment losses in the United Kingdom during 1995–2001 in either manufacturing or services.
  - Positive coefficients on service outsourcing in some one-period specifications are not robust to two-period differencing or to inclusion of industry fixed effects.
  - Aggregate interpretation: jobs displaced by service outsourcing are likely offset by new jobs created in the same sector; detectable negative effects in the United States appear only at very fine (450-sector) disaggregation.

### Key takeaways and policy-relevant findings
- International service outsourcing (computing + business services) remains small as a share of GDP in large industrial economies but has been rising:
  - United States: 0.4 percent in 2003.
  - United Kingdom: 1.2 percent in 2003.
- Industry-level measures show higher absolute levels in the United Kingdom (service import share increased from 3.5 percent in 1992 to 5.5 percent in 2001) than in the United States (0.4 percent to 0.8 percent over the same period).
- Material outsourcing is quantitatively much larger than service outsourcing (example: material outsourcing ~27 percent in the United Kingdom vs. service outsourcing much lower).
- Large absolute dollar values of service imports concentrated in large economies, but many smaller economies exhibit much higher outsourcing intensities when scaled by GDP.
- Evidence on employment effects is mixed:
  - Paper’s UK case study finds no employment losses from outsourcing for 1995–2001.
  - Other studies report small or sector-specific impacts; productivity effects of service outsourcing tend to be positive in several studies.
- Policy implication (implicit):
  - Media-driven perceptions of exploding, one-way outsourcing from developed to developing countries are not supported by the presented data.
  - Concerns about massive net job losses from service outsourcing are exaggerated given the evidence in the United Kingdom sample and broader international trade data; policy evaluation should account for sectoral reallocation, aggregation level, and adjustment frictions.

### Data sources and sample construction (key details)
- Trade data: IMF Balance of Payments, International Financial Statistics — imports and exports of computing and information services; and other business services.
- News count: FACTIVA, Dow Jones & Reuters.
- Input/output tables used to match industry classifications; final UK sample: 69 manufacturing industries and 9 service industries.
- Employment sources: Annual Employment Survey (AES) Great Britain, Annual Business Inquiry (ABI) United Kingdom, Census of Employment Northern Ireland (1995 and 1997); harmonization steps described for merging Northern Ireland data.
- Price indices: National Statistics online, United Kingdom; weighted averages constructed using average employment for 1998–2001 as weights.

*Source: _wp04186 (IMF working paper extract).*

### 1.     Outsourcing of   Services........................................................................................

### Outsourcing of Services

### Introduction and context
- Media and political attention on service outsourcing surged in early 2004:
  - There were 2,634 reports in US newspapers on service outsourcing between January and May 2004.
  - The United Kingdom had 380 reports on outsourcing in its newspapers during the same period.
  - In a two-week period from March 1 to March 15, 2004, there were 270 stories that simultaneously mentioned outsourcing and either job loss or unemployment in the same story.
- A quarterly index constructed from FACTIVA (first quarter 1991 to first quarter 2004) for the United States and the United Kingdom shows a clear upward trend in media interest in international outsourcing of services.
  - The United States index exhibits local peaks in 1996, 2000, and 2004 (all presidential election years).

### Definitions and scope
- The paper restricts attention to international outsourcing (concerns about jobs moving from industrial countries to developing countries).
- Focused service categories:
  - Business services.
  - Computing and information services.
- Distinction highlighted between material-input outsourcing and service outsourcing:
  - Material-input outsourcing remains quantitatively larger for a typical industrialized economy.
  - Service outsourcing is increasingly salient because of communication technology improvements (internet) that allow services to cross borders without firms giving up market proximity.

### Public opinion, politics, and possible policy responses
- Evidence of declining support for free trade among high-income earners:
  - University of Maryland study: among those with incomes over $100,000, the percentage actively supporting free trade fell from 57 percent in 1999 to 28 percent in January 2004.
- Political actions and pressures noted:
  - In the United States, the Senate passed restrictions on foreign outsourcing for federal contracts in March 2004 (the measure did not become law).
  - Trade and Industry Secretary Patricia Hewitt stated that the United Kingdom will not pass protectionist legislation (Financial Times, 5/3/2004, page 6).
  - Reports of industry lobbying, e.g., Australian software companies lobbying to restrict outsourcing of software designs to India.

### Trends, magnitudes, and stylized facts
- Service outsourcing has been steadily increasing but remains at low absolute levels:
  - In the United States, imports of computing and business services as a share of GDP were 0.4 percent in 2003.
  - Historical shares for the United States:
    - 0.1 percent in 1983.
    - 0.2 percent in 1993.
    - 0.4 percent in 2003.
- Research questions addressed in the paper:
  - Has service outsourcing exploded in recent years?
  - How does service outsourcing compare with material outsourcing?
  - Who are the biggest outsourcers of services?
  - Who are the biggest recipients of service outsourcing (the “insourcers”)?
  - Are there large job losses arising from service outsourcing?

### Effects, concerns, and adjustment costs
- Concerns motivating study:
  - Fear of job losses and resulting political pressure for protectionist policies.
  - Potential short-term adjustment costs due to labor market rigidities (even if long-run benefits accrue).
  - Need to quantify the magnitude of employment effects to inform policy (including possible relocation assistance programs).

### Objectives and approach
- Main objective: distinguish hype from facts about service outsourcing by developing stylized facts and empirical evidence on trends and effects.
- Empirical focus:
  - Business services.
  - Computing and information service trade.
- The paper aims to correct misleading impressions from the news media and to provide evidence on the scale and employment impacts of service outsourcing.

*Source: _wp04186 - 1.     Outsourcing of   Services........................................................................................*

### 0.4 percent in 2003, based on IMF’s balance of payments trade data. A similar picture emerges

### _wp04186 - 0.4 percent in 2003, based on IMF’s balance of payments trade data. A similar picture emerges

### Definitions and measurement of outsourcing
- Outsourcing is defined as the procuring of service or material inputs by a firm from a source in a foreign country; it includes both intra-firm international outsourcing and arm’s-length international outsourcing.
- Offshoring is used synonymously with international outsourcing in this paper’s usage.
- Insourcing is defined here as outsourcing in the opposite direction (from foreign-located firms to domestic firms); e.g., U.S. “insourcing” refers to outsourcing from the rest of the world to the United States.
- Two measures of service outsourcing are constructed:
  - Economy-wide measure: imports of computing (including computer software designs) and other business services using IMF’s Balance of Payments Statistics Yearbook.
  - Industry-level measure (for the United Kingdom): outsourcing intensity OSS_i, calculated using input/output tables and IMF trade data:
    - OSS_i = [imports of service input purchases of service by industry / total nonenergy inputs used by industry production + imports - exports]_j (first bracket from input/output tables) * [economy-wide import share applied to industry use] (second bracket from IMF trade data).
  - Analogue measure OSM_i constructed for material outsourcing.
- Measurement caveats noted:
  - Measures likely under-estimate outsourcing because import prices may be lower than domestic purchase prices and quantity data are unavailable.
  - Applying the same import share to all industries is a proxy due to unavailability of imports by industry.
  - Total input use excludes services produced within the same industry.

### Global and historical patterns — magnitudes and trends
- Economy-wide IMF-based findings for computing + other business services:
  - United States imports of computer and information plus other business services as a share of GDP: 0.1 percent in 1983, 0.2 percent in 1993, and 0.4 percent in 2003.
  - United Kingdom: 0.9 percent in 1983, 0.7 percent in 1993, and 1.2 percent in 2003.
- Industry-level UK outsourcing intensity (output-weighted averages from input/output construction):
  - Service imports share in the United Kingdom: 3.5 percent in 1992 and 5.5 percent in 2001.
  - United States comparable industry-level figures: 0.4 percent in 1992 and 0.8 percent in 2001 (see Amiti and Wei, 2004).
- Material versus service outsourcing:
  - Material outsourcing is substantially higher than service outsourcing.
  - Material outsourcing intensities: around 27 percent in the United Kingdom and 12 percent in the United States.
  - In the United Kingdom material outsourcing peaked in 1996 and trended downward thereafter; in the United States material outsourcing increased steadily but more slowly than service outsourcing.
- Summary statement: Service outsourcing is much lower than material outsourcing, but service outsourcing is increasing at a faster pace.

### Country-level outsourcing and insourcing rankings and values (2002 data unless stated)
- Top importers (value of imports of business services, 2002):
  - United States: US$41 billion
  - Germany: US$39 billion
  - Japan: US$25 billion
  - Netherlands: US$21 billion
  - Italy: US$20 billion
  - France: US$19 billion
  - United Kingdom: US$16 billion
  - India: US$11 billion (ranked 11th)
  - China: US$8 billion (ranked 18th)
- Top importers of computer and information services (quantitatively smaller category):
  - Top five: Germany, United Kingdom, Japan, Netherlands, and Spain; United States is 6th; China ranks 10th. (No IMF Balance of Payments Yearbook data for India in this category.)
- Outsourcing intensity (imports of business services scaled by GDP, 2002):
  - Countries with highest ratios (examples): Angola, Republic of Congo, Mozambique, Ireland, Vanuatu — ratio of imported business services to GDP exceeding 10 percent for some.
  - United States: outsourcing ratio in business services less than 0.5 percent of GDP (ranked 117th in the world).
  - United Kingdom: slightly over 1 percent of GDP (ranked 85th in the world).
  - China: ranked 99th in the world (ahead of the United States).
  - India: imports of business services as a share of GDP = 2.4 percent (larger than United States and United Kingdom by this metric).
- Insourcing (exports of business and computing services as proxy for receiving outsourcing):
  - Top recipients in dollar terms (2002):
    - United States: US$59 billion
    - United Kingdom: US$37 billion
    - Germany: US$28 billion
    - France: US$21 billion
    - Netherlands: US$20 billion
    - India: US$18.6 billion (ranked 6th)
    - China: US$10 billion (ranked 14th)
  - When scaled by GDP, top insource-intensive economies include Vanuatu, Singapore, and Hong Kong SAR, each with exporting services as a share of local GDP exceeding 10 percent.
  - By the exports-as-share-of-GDP metric, India (3.8 percent of GDP) is somewhat more insourcing-intensive than the United Kingdom (3.8 percent vs. [United Kingdom value truncated in source text]; the source notes India is somewhat more insourcing-intensive than the United Kingdom).

### Employment and productivity effects — literature synthesis and paper’s case study
- Empirical literature:
  - Material outsourcing: many studies find a steady increase in international outsourcing of material inputs across OECD countries (examples cited include Feenstra and Hanson (1996), Campa and Goldberg (1997), Hummels, Ishii and Yi (2001), Yeats (2001), Hanson, Mataloni and Slaughter (2004), and Borga and Zeile (2004)).
  - Yeats (2001): estimates that 30 percent of OECD exports of machinery and transport equipment comprised parts and components in 1995, and 26 percent in 1978.
    - United States share increased from 36 percent in 1978 to 40 percent in 1995.
    - Europe increased from 26 percent to 28 percent; Japan from 15 percent to 26 percent.
    - Apparent consumption of transport and machinery components (1995): EU highest at 16 percent, United States 11 percent, Japan 8 percent.
  - Effects on productivity and skill premium:
    - Egger and Egger (2001): negative short-run effect of international material outsourcing on productivity of low-skilled workers, positive long-run effect; international outsourcing of materials contributed to 3.3 percent of real value added per low-skilled worker in the EU from 1993 to 1997.
    - Feenstra and Hanson (1999): outsourcing contributed between 17.5 to 40 percent of the increase in the non-production wage share over 1979 to 1990.
  - Service outsourcing: thinner literature but growing evidence of positive productivity effects in some studies:
    - Mann (2004): globalization and international outsourcing of IT hardware led to a fall of 10 to 30 percent in prices of IT hardware, increasing productivity; IT software outsourcing expected to have similar or larger effects if more price elastic.
    - Amiti and Wei (2004): service outsourcing positively correlated with labor productivity in the United States; material outsourcing insignificant in that study.
    - Gorg and Hanley (2003): positive impact of international outsourcing of services on productivity in Ireland’s electronics industry (1990–1995); tangible inputs not significant.
    - Girma and Gorg (2003): positive evidence of service outsourcing on labor productivity and total factor productivity in the United Kingdom (1980–1992) but could not separate domestic from foreign outsourcing.
  - Employment effects from management-consultant studies:
    - McKinsey Global Institute (2003): predictions on job loss due to outsourcing from 2003 to 2015; methodology not fully transparent; concludes job losses from outsourcing are relatively trivial compared to typical business-cycle job losses.
    - Brainard and Litan (2004): emphasize distributional effects, low-paid jobs replaced with higher-paid jobs, provide policy prescriptions for the United States.
    - Shultze (2004): offers indirect evidence of job losses related to service outsourcing, concluding effects are very small.
  - Amiti and Wei (2004): using U.S. data, find a small negative effect of service outsourcing on employment with highly disaggregated data.
- The paper’s UK case study (Section V):
  - Analyzes effects of service and material outsourcing on employment in the United Kingdom for 1995–2001.
  - Main finding: no evidence that outsourcing led to employment losses in the United Kingdom during 1995 to 2001 in either the manufacturing or the services sectors.

### Theoretical context and labor-market considerations
- Many theoretical models of production fragmentation and location (Jones and Kierzkowski; Dixit and Grossman; Krugman and Venables; Deardorff; Yi; Amiti; Antras; Helpman) focus on where production stages are located and on input trade; most assume perfect inter-sectoral labor mobility and full employment, thus predicting no net job losses but changes in factor prices (Samuelson–Stolper effects).
- When short-run frictions exist (imperfect inter-sectoral labor mobility, unions, participation responses), trade and outsourcing can produce net employment effects (Sachs and Shatz (1994); Krugman (1995) examples).
- McKinsey report note: more than 69 percent of workers who lost jobs due to imports in the United States between 1979 and 1999 were re-employed (U.S. Bureau of Labor Statistics data), implying 31 percent were not re-employed and indicating possible labor market rigidities.

### Key takeaways
- International service outsourcing (computing + business services) remains small as a share of GDP in large industrial economies but has been rising (United States: 0.4 percent in 2003; United Kingdom: 1.2 percent in 2003).
- Industry-level measures show higher absolute levels in the United Kingdom (service import share increased from 3.5 percent in 1992 to 5.5 percent in 2001) than in the United States (0.4 percent to 0.8 percent over the same period).
- Material outsourcing is quantitatively much larger than service outsourcing (e.g., material outsourcing ~27 percent in the United Kingdom vs. service outsourcing much lower).
- Large absolute dollar values of service imports are concentrated in large economies, but when scaled by GDP many smaller economies exhibit much higher outsourcing intensities.
- Evidence on employment effects is mixed; the paper’s UK case study finds no employment losses from outsourcing for 1995–2001, while other studies report small or sector-specific impacts; productivity effects of service outsourcing tend to be positive in several studies.

*Source: _wp04186 - 0.4 percent in 2003, based on IMF’s balance of payments trade data.*

### 2.4 percent); and China is somewhat ahead of the United States (0.8 percent of GDP vs.

### 2.4 percent); and China is somewhat ahead of the United States (0.8 percent of GDP vs.

### Key statistic
- "2.4 percent); and China is somewhat ahead of the United States (0.8 percent of GDP vs."

*Source: IMF PDF chapter/section _wp04186 - 2.4 percent); and China is somewhat ahead of the United States (0.8 percent of GDP vs.*

### 0.6 percent).

### _wp04186 - 0.6 percent).

### Who Are the Biggest Surplus Countries?
- Largest surplus countries in combined computing and business services are the United Kingdom and the United States.
- The United States has been running a surplus in this service category every year since 1980; the United Kingdom has likewise been running a surplus.
- If every country reduced its overall service outsourcing, the United States and the United Kingdom would be the biggest two losers in terms of net dollars lost in service trade; the U.S. current account deficit would become bigger, not smaller.
- Patterns vary across other industrialized countries:
  - Germany ran a small deficit in business services every year between 1980 and 2001.
  - France ran a small surplus for much of the sample, switching to a mild deficit at the end.
  - India: imports and exports of business services were fairly balanced early in the sample; starting from 1996, exports surged and now produce a reasonably large surplus.
  - China: relative size of imports and exports alternates; ends the sample with a small surplus.
- For computer and information services, patterns broadly similar; Ireland is identified as the largest surplus country in computing.

Key country figures (selected, as reported in tables):
- Table 2: Absolute outsourcers (Business Services, Million US Dollars) — United States: 40,929; Germany: 39,113; Japan: 24,714; Netherlands: 21,038; Italy: 20,370.
- Table 4: Absolute insourcers (Business Services, Million US Dollars) — United States: 58,794; United Kingdom: 36,740; Germany: 27,907; France: 20,864; Netherlands: 20,074.
- Table 6 (Surplus and Deficit Countries, 2002, totals): United Kingdom: 23,628.68; United States: 21,748.30; Hong Kong SAR: 15,663.41; India: 6,813.44; Singapore: 3,826.12.
- Outsourcing intensity by country (examples): United Kingdom engages in about three times as much service outsourcing as a share of its GDP (1.2 percent in 2001) as the United States (0.4 percent in 2001).

### Does Service Outsourcing Reduce Jobs? — Theory and Empirical Approach
- Theoretical predictions are ambiguous:
  - Outsourcing can cause short-run net employment losses if labor mobility and market rigidities prevent reallocation.
  - Outsourcing can generate job growth via firm efficiency gains, expanded production in comparative-advantage activities, lower prices, increased demand, and job creation.
- Empirical specification used (labor demand framework, Hamermesh 1993) — estimating equation (first-differenced):
  - ∆ln Lit = α0 + α1 ∆ln wit + α2 ∆ln OSSit + α3 ∆ln OSMit + β ω + γ Dt + δ industry effects + εit
  - ∆ln OSSit is log difference in service outsourcing intensity (used as inverse proxy for price of imported service inputs).
- Data notes:
  - U.S. companion paper (Amiti and Wei, 2004): negative employment effect detectable at 450-sector disaggregation (α2 < 0), but disappears at 96-sector aggregation.
  - United Kingdom case study: data for 1995 to 2001; includes 69 manufacturing industries and 9 service industries (78 sectors total).

### Statistical Results (United Kingdom, 1995–2001)
- Descriptive checks:
  - No uniform pattern between sectoral service outsourcing growth and employment growth (Figures and Tables 7A/7B): examples include “other transport equipment” with second highest employment growth and one of the highest service outsourcing growth; “preparation and spinning of textile fibers” experienced negative employment growth and high outsourcing growth.
  - Scatter plot (Figure 5) of service outsourcing growth vs. employment growth across 78 industries shows no clear one-way relationship.
- Regression evidence — manufacturing industries (Tables 8A and 8B):
  - One-period first-difference specifications (Table 8A, columns with one-period differences): service outsourcing coefficients positive and sometimes statistically significant (e.g., ∆ ln(OSS)_t = 0.079*, ∆ ln(OSS)_t-1 = 0.063* in column (1); significance legend: * p < 0.1; ** p < 0.05; *** p < 0.01).
  - Two-period differenced specifications reduce or eliminate significance of outsourcing coefficients (Table 8A, columns (4)–(6): outsourcing effect weakly significant at 10 percent in column (4) and insignificant in other two-period specifications).
  - Sensitivity tests (Table 8B): outsourcing coefficient positive and significant in some one-period specifications (e.g., column (2) shows significance at **), but becomes insignificant or barely significant in two-period differenced specifications and with industry fixed effects.
  - Wage effect: ∆ ln (wage) consistently negative and significant (e.g., Table 8A: -0.742***, -0.683***), as theorized.
  - Output effect: ∆ ln (real output) consistently positive and significant (e.g., Table 8A: 0.444***, 0.461***).
  - Sample sizes and goodness of fit (examples): Table 8A N = 345, 276, 256, 276, 207, 192 across columns; R2 ranges from 0.12 to 0.76 in reported columns.
- Services industries (Tables 9A and 9B):
  - Partial-correlation specifications show mostly insignificant effects of service outsourcing on service-sector employment.
  - A single specification (Table 9B, two-period differenced data, column (3)) reports a significant negative effect at the 5 percent level, but this effect is small and not robust to adding industry fixed effects — it disappears when industry-specific effects are included.
  - Wage and nominal output effects in services: ∆ ln (wage) negative and significant in many specifications (e.g., Table 9A: -0.939***, -0.975***); ∆ ln (nominal output) positive and significant in several specifications (e.g., Table 9A: 0.797***, 1.181***).
  - Reported R2 values in services regressions are high in some specifications (e.g., Table 9A: R2 up to 0.91 in column (3)).

Main empirical conclusion:
- Across specifications for the United Kingdom, there is no robust evidence that service outsourcing leads to net job losses at the sectoral level. Positive coefficients on service outsourcing in several one-period specifications are not robust to two-period differencing or to inclusion of industry fixed effects, and no negative coefficients robustly appear in manufacturing regressions.
- Aggregate interpretation: jobs displaced by service outsourcing are likely to be offset by new jobs created in the same sector (consistent with findings that U.S. negative effects appear only at very fine (450-sector) disaggregation but not at broader (96-sector) aggregation).

### Concluding Remarks — Synthesis and Policy-Relevant Findings
- Media-driven perceptions of exploding, one-way outsourcing from developed to developing countries are not supported by the data presented:
  - Many developed countries are not more outsourcing-intensive (adjusted for economic size) than many developing countries.
  - Several major developed economies—most notably the United States and the United Kingdom—are net exporters (insourcers) in business services and computer and information services.
  - United States and United Kingdom are reported as the largest and second largest surpluses in services trade in recent years.
- Labor-market implications:
  - The risk that service outsourcing will dramatically reduce aggregate job growth in advanced economies is greatly exaggerated by the data in this study.
  - Evidence indicates that outsourcing can coexist with sectoral job creation, and at typical levels of aggregation the aggregate employment effect is negligible or offsetting.
- Policy implication (implicit in analysis):
  - Concerns about job losses from service outsourcing should be evaluated carefully with attention to sectoral reallocation, aggregation level, and adjustment frictions; blanket fears of massive net employment losses are not empirically supported in the United Kingdom sample or in broader international trade data presented.

*Source: IMF working paper content provided in the supplied PDF extract.*

### Appendix 1: Data Sources

### Appendix 1: Data Sources

### Trade Data
- Source: IMF Balance of Payments, International Financial Statistics
- Variables: imports and exports of computing and information services; and other business services

### News count
- Source: FACTIVA, Dow Jones & Reuters, www.factiva.com

### Input/output tables
- Source: (not explicitly listed; used for matching industry classifications)
- Note: Industries were aggregated to align information from all sources. Employment data from ABI were first aggregated into SIC92 3-digit level to match AES categories. A second stage of aggregation was applied where there was a multiple-to-multiple correspondence between the I/O tables codes and the SIC92 3-digit codes. Industries not of interest (e.g., agriculture and mining) or with incomplete information were dropped. Final sample: 69 manufacturing industries and 9 service industries.

### Total compensation
- Variable: Output in current values
- Source: National Statistics online, United Kingdom, (www.statistics.gov.uk)

### Employment
- Sources and coverage:
  - Annual Employment survey (AES) Great Britain, SIC92 3 digit, 1995-98
  - Annual Business Inquiry (ABI) United Kingdom, SIC92 5 digit 1998-2001
  - Census of Employment, Northern Ireland, 1995 and 1997
- Data harmonization steps:
  - Northern Ireland employment data were added to Great Britain to obtain United Kingdom figures for 1995-1997.
  - For 1996, Northern Ireland employment was taken as a simple average of 1995 and 1997 employment.
  - For industries lacking Northern Ireland correspondence, the overlapping year (1998) served as a bridge: employment of Great Britain industries assumed to be constant ratios of those of United Kingdom.

### Price Indices
- Source: National Statistics online, United Kingdom, (www.statistics.gov.uk)
- Coverage: Manufacturing industries only, SIC92.
- Construction method:
  - Price indices available at different SIC 92 disaggregation levels that do not correspond one-to-one with sample industries.
  - A weighted average of these PPI was constructed using average employment for the United Kingdom for the period 1998 to 2001 as weights (the only years available at the appropriate level of disaggregation) to obtain price indices at 3 digit SIC level.

### Notes on sample construction
- Final sample composition after aggregation and exclusion:
  - Total Manufacturing Industries = 69
  - Total Service Industries = 9
- Specific grouping: Shading indicates industries grouped to match I/O classifications. Specifically, industries 14–17, 18–19, 25–26, 36–40, 100–102, and 109–111 are treated as 6 groups of industries.

### Appendix II. List of Industries in the U.K. Sample

Manufacturing Industries (indexed as in source)
- 8 Meat processing
- 9 Fish and fruit processing
- 10 Oils and fats
- 11 Dairy products
- 12 Grain milling and starch
- 13 Animal feed
- 14 Bread, biscuits etc
- 15 Sugar
- 16 Confectionery
- 17 Other food products
- 18 Alcoholic beverages
- 19 Soft drinks and mineral waters
- 20 Tobacco products
- 21 Textile fibres
- 22 Textile weaving
- 23 Textile finishing
- 24 Made-up textiles
- 25 Carpets and rugs
- 26 Other textiles
- 27 Knitted goods
- 28 Wearing apparel and fur products
- 29 Leather goods
- 30 Footwear
- 31 Wood and wood products
- 32 Pulp, paper and paperboard
- 33 Paper and paperboard products
- 34 Printing and publishing
- 36 Industrial gases and dyes
- 37 Inorganic chemicals
- 38 Organic chemicals
- 39 Fertilisers
- 40 Plastics & synthetic resins etc
- 41 Pesticides
- 42 Paints, varnishes, printing ink etc
- 43 Pharmaceuticals
- 44 Soap and toilet preparations
- 45 Other chemical products
- 46 Man-made fibres
- 47 Rubber products
- 48 Plastic products
- 49 Glass and glass products
- 50 Ceramic goods
- 51 Structural clay products
- 52 Cement, lime and plaster
- 53 Articles of concrete, stone etc
- 54 Iron and steel
- 55 Non-ferrous metals
- 56 Metal castings
- 57 Structural metal products
- 58 Metal boilers and radiators
- 59 Metal forging, pressing, etc
- 60 Cutlery, tools etc
- 61 Other metal products
- 62 Mechanical power equipment
- 63 General purpose machinery
- 64 Agricultural machinery
- 65 Machine tools
- 66 Special purpose machinery
- 67 Weapons and ammunition
- 68 Domestic appliances nec
- 69 Office machinery & computers
- 70 Electric motors and generators etc
- 71 Insulated wire and cable
- 72 Electrical equipment nec
- 73 Electronic components
- 74 Transmitters for TV, radio and phone
- 75 Receivers for TV and radio
- 76 Medical and precision instruments
- 77 Motor vehicles
- 78 Shipbuilding and repair
- 79 Other transport equipment
- 80 Aircraft and spacecraft
- 81 Furniture
- 82 Jewellery and related products
- 83 Sports goods and toys
- 84 Miscellaneous manufacturing nec & recycling
- 85 Electricity production and distribution
- 88 Construction

Service Industries (indexed as in source)
- 108 Research and development
- 99 Telecommunications
- 109 Legal activities
- 100 Banking and finance
- 110 Accountancy services
- 101 Insurance and pension funds
- 111 Market research, management consultancy
- 102 Auxiliary financial services
- 112 Architectural activities and technical consultancy
- 106 Renting of machinery etc
- 113 Advertising
- 107 Computer services
- 114 Other business services

- Total Manufacturing Industries = 69
- Total Service Industries = 9

*Source: Appendix 1: Data Sources (as provided in the content unit).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04186.pdf_