## _wp11148

## Source details

**Canonical URL:** [_wp11148](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11148.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11148.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11148.pdf.json)

---

### I. Research question, scope, and datasets
- Research question: How does employment across sectors adjust in response to reallocative shocks, with a focus on the role of the real exchange rate in China?
- Dataset and period: sectoral and regional data in China from 1980 to 2008; most analysis uses data starting 1994 or later to avoid regime-change driven spurious results.
- Datasets used:
  - Old sectoral employment series ("Employmentdata"): 16 sectors, 1980–2002, ILO-aligned definition.
  - New sectoral employees series ("Employees"): urban unit employees, 1989/1990–2008 (excludes private and self-employed).
  - Regional employment by industry (primary, secondary, tertiary): regional panels 1985/1988–2008 (CEIC; China Statistical Yearbooks).
  - National input-output tables: 1995, 1997, 2000, 2002, 2005.
  - Regional input-output tables: 2002 (30 regions; Chongqing merged with Sichuan); final input-output panel: 28 regions, 16 non-tradable sectors, 6 years (2003–2008).
  - Real effective exchange rate (REER) series from IFS.

### II. Theoretical framework and testable implications
- Mechanisms through which REER affects labor demand:
  - Competitiveness channel: real appreciation reduces exporters’ competitiveness → lower output and employment in export-oriented tradable sectors (price elasticity ηe < 0).
  - Imported-input channel: real appreciation lowers domestic price of imported inputs (s*e) → may reduce marginal costs and increase employment where imported input shares are sizable.
- Net effect on tradable employment ambiguous ex ante; depends on relative strength of competitiveness (negative) and imported-input (positive) channels.
- Non-tradable sectors affected only via imported-input channel in the baseline model → a priori positive employment effect of a real appreciation (absent intermediate-input linkages).
- Cross-sectional heterogeneity:
  - Export share Xs,t affects sectoral price elasticity: ηe s,t = k Xs,t (higher Xs,t → larger |ηe|).
  - Imported-input share (1 − α − β) amplifies imported-input channel.
- Empirical reduced-form relations (log-linearized) imply:
  - Tradable employment/wage equations include coefficients cT3;1, dT3;1 < 0 (competitiveness) and cT3;2, dT3;2 > 0 (imported-input).
  - Non-tradable employment/wage equations include coefficients cN2, dN2 > 0 (imported-input).

### III. Empirical strategy and identification
- Estimation methods:
  - Panel OLS, Instrumental Variable estimation (2SLS), Seemingly Unrelated Regression (SUR), GLS, system GMM (Blundell and Bond (1998)).
  - Dynamic panels with lagged dependent variables to capture gradual employment response.
  - Instruments for aggregate REER: change in Japanese Yen/US dollar nominal exchange rate and US 3-month T-Bill rate.
- Key heterogeneity variables:
  - Sectoral/regional export shares (external exposure).
  - Sectoral/regional imported-input shares.
  - Regional input-output linkage measure: shr_i = p^N_i Q^N_i / (p^N Q^N) (share of non-tradable gross output used as intermediate input in tradable sectors).
- Time-series properties:
  - Most series fail unit-root rejection; no panel cointegration (Westerlund (2007)); many specifications use first differences.

### IV. Core empirical findings — sectoral and aggregate dimensions
- REER-employment associations (exact magnitudes preserved):
  - Table 3 (old Employment, 15 sectors, 1980–2002): estimated REER coefficients range from -0.04 to -0.14.
    - Interpretation in paper: a 10 percent real appreciation lowers employment growth by 0.4 to 1.4 percentage points on average across sectors (except agriculture).
  - Employees dataset (1989/1990–2008): a 10 percent real appreciation leads to about 0.7 percent decrease in employment growth (broad consistency with Table 3).
  - REER × tradable/export interactions:
    - REER X tradable dummy coefficients: -0.095 (0.032) and -0.192 (0.058) reported in Table 4.
    - REER X lagged sectoral export shares: -0.264 (0.091), -0.271 (0.124).
- Symmetry and heterogeneity:
  - Negative employment effects of appreciation tend to be more pronounced than positive effects of depreciation (asymmetry documented).
  - Negative REER effects are observed not only in tradable sectors but also in several non-tradable sectors (transport, utilities, wholesale/retail trade).

### V. Core empirical findings — regional and industry panels
- Broad-industry regional regressions (primary, secondary, tertiary):
  - Primary sector: REER has no significant effect on regional employment.
  - Secondary industry: appreciation has negative effect on employment growth; estimates comparable in magnitude to sectoral results (sometimes significant at 10 percent).
  - Tertiary industry: REER coefficient negative and significant at 5 percent in recent samples; effect slightly more pronounced than for tradable sectors in some specifications.
- SUR and dynamic panel quantified impacts:
  - SUR (Table 8): a 10 percent real appreciation (depreciation) reduces (increases)
    - Industry 2 employment by 1.1 percent.
    - Industry 3 employment by 1.2 to 1.3 percent.
  - Wald test: cannot reject equality of REER coefficients across Industry 2 and Industry 3 (very high p-value >50 percent).
  - Dynamics (Table 9): tertiary sector adjusts mostly within same year; secondary sector adjustment more persistent.
  - Asymmetry (Table 11): negative effect of appreciation drives employment contraction in both tradable and non-tradable industries.
- Instrumental-variable evidence:
  - First-stage correlations: YEN/USD movement correlated 0.6 to 0.7 with Chinese REER.
  - US interest rate correlated with Chinese REER in first stage (except for industry 3).
  - Hansen overidentification tests: cannot reject instrument validity.
  - 2SLS second-stage results are close to OLS; Hausman-Wu tests cannot reject exogeneity of REER.

### VI. Transmission channel via intermediate-input linkages — theory and evidence
- Theoretical extension:
  - Non-tradable gross output serves both direct consumption Q^N_c and as intermediate input Q^N_i for tradable production; shr_i measures share of non-tradable output used as intermediate input in tradable sectors.
  - Derived elasticity predicts that larger shr_i amplifies the negative effect of appreciation on non-tradable employment via reduced intermediate demand from tradables.
- Input-output facts (exact figures preserved):
  - National intermediate-input shares to tradables (Table 14; selected years):
    - Banking and insurance: 0.426 (1995), 0.372 (1997), 0.348 (2000), 0.304 (2002), 0.310 (2005).
    - Transportation and Post: 0.399 (1995), 0.423 (1997), 0.357 (2000), 0.381 (2002), 0.384 (2005).
    - Wholesale and retail trades: 0.548 (1995), 0.394 (1997), 0.364 (2000), 0.327 (2002), 0.360 (2005).
    - Utility: 0.706 (1995), 0.730 (1997), 0.738 (2000), 0.590 (2002), 0.607 (2005).
  - Regional intermediate-input shares (Table 15; examples for 2002): Mean shares and ranges reported (e.g., Mean for Banking 0.379, Transportation 0.415, Wholesale/Retail 0.292, Utilities 0.585; Min and Max reported).
- Region-sector regression evidence (2003–2008, 28 regions, 16 non-tradable sectors):
  - Key coefficient REER * shr_ir consistently negative and highly significant across specifications:
    - OLS/GLS contemporaneous: REER * shr_ir ≈ -0.216  (0.036) to -0.228  (0.029).
    - Lagged regressors: REER(t-1) * shr_ir ≈ -0.223  (0.036) to -0.226  (0.031).
    - ln L_Nir;t-1 coefficients near persistence: 0.936  (0.031) contemporaneous; 0.950  (0.029) lagged.
  - System GMM and robustness:
    - GMM estimate of REER term ≈ -0.2 (lagged regressors) and significant at 1 percent.
    - AR(1) estimate (rho_hat) ≈ 0.513 in GLS AR(1); AR(2) p-values reported ~0.367–0.369; Hansen p-values 0.180 (contemporaneous) and 0.279 (lagged).
    - Economic interpretation: for non-tradable sectors with an average intermediate input share to tradables, a 10 percent real appreciation leads to 0.9 percent lower employment growth within 2 years (GMM/lagged regressors interpretation).
  - Interaction with lagged regional export share returns insignificant when added, suggesting input-output interdependence drives non-tradable contraction rather than regional export exposure alone.

### VII. Robustness, magnitudes, dynamics, and auxiliary findings
- Robustness across data and methods: results consistent across old Employment, Employees dataset, regional OLS, SUR, 2SLS, GLS, and system GMM.
- Preserved magnitudes and dynamics:
  - REER coefficients in Table 3: range from -0.04 to -0.14.
  - 10 percent appreciation → 0.4 to 1.4 percentage point change (Table 3 interpretation).
  - Employees dataset: 10 percent appreciation → about 0.7 percent decrease in employment growth.
  - SUR full-sample: Industry 2 = -1.1 percent; Industry 3 = -1.2 to -1.3 percent for a 10 percent appreciation.
  - GLS AR(1) first-order autocorrelation: 0.513.
  - Instrument correlations: 0.6 to 0.7 (YEN/USD with Chinese REER).
  - GMM lagged regressors: REER term ≈ -0.2; implied 10 percent appreciation → 0.9 percent lower employment growth within 2 years for average shr_i.
- Interpretation challenges:
  - Contraction in non-tradable employment after a real appreciation contradicts baseline comparative-static prediction (which predicts expansion of non-tradables).
  - Reallocation frictions alone insufficient to explain observed non-tradable contraction; intermediate-input linkage provides a plausible short-run mechanism.
  - Other channels (learning-by-doing, technology diffusion) may operate over longer horizons and are not ruled out.

### VIII. Policy-relevant implications and recommendations
- Short-run risks of revaluation:
  - A real appreciation can cause employment contraction across both tradable and certain non-tradable sectors, implying broader labor-market adjustment costs than standard focus on manufacturing alone.
  - Short-term adverse employment effects imply limited immediate boost to domestic demand following appreciation.
- Policy packages recommended:
  - Accompany any revaluation that causes real appreciation with macroeconomic policies to support domestic demand.
  - Structural reforms to enhance productivity in non-manufacturing sectors to reduce their dependence on tradable-sector demand and to facilitate labor reallocation.
  - Consider expansion in supply of services—particularly social services and health care—to absorb displaced labor (policy discussion referenced in source).
- Research and data needs:
  - Micro-level (firm-level) data and further study of alternative transmission channels (learning, technology diffusion) are needed to assess medium- and long-term welfare consequences.

*Source: Excerpts from IMF working paper content unit _wp11148 (Introduction, Theory, Data, Empirical strategy, Results, Input-Output analysis, Tables and Figures as provided in the source PDF).*

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

### _wp11148 - References .............................................................................................................

### I. Introduction — research question and scope
- Research question: How does employment across sectors adjust in response to reallocative shocks, with a focus on the role of the real exchange rate in China?
- Dataset and period: sectoral and regional data in China from 1980 to 2008.
- Motivation:
  - China’s rapid growth and role in world trade has intensified international pressure to appreciate the currency to correct global imbalances.
  - Policy discussion has emphasized effects on global imbalances and manufacturing employment in trading partners; less attention to adjustment in the Chinese labor market across all sectors.
- Key empirical focus:
  - Effects of real exchange rate fluctuations on employment and wages across tradable and non-tradable sectors, and across provinces.
  - Use of panel OLS, Instrumental Variable estimation, and seemingly unrelated regression (SUR) techniques to control endogeneity and cross-industry/region correlation.
- Core empirical findings (summary statements from the paper):
  - Movements in the real exchange rate are associated with sizable changes in employment and wages in all sectors.
  - A real appreciation lowers employment growth in tradable sectors (mostly manufacturing) and in some non-tradable sectors (mostly transport, utilities, and wholesale/retail trade).
  - At the provincial level, a real appreciation (depreciation) decreases (increases) employment growth in the secondary industry and tertiary industry; agriculture is an exception.
  - Depending on specification, a 10 percent real appreciation lowers employment growth by 0.4 to 1.4 percentage points across sectors except for agriculture.
  - Negative employment effects of a real appreciation tend to be more pronounced than the positive effect of a real depreciation, and are reinforced by larger regional exposure to exports.
  - Regions and services sectors with a relatively high share of production serving as intermediate input to tradable sectors experience a stronger negative impact of a real appreciation.
- Contribution claims:
  - Extends literature by analyzing non-manufacturing sectors and linking regional input-output data with regional sector-level employment in China.
  - First to combine region-sector input-output structure with employment to study exchange rate impacts in China.

### II. Theory — framework and testable implications
- Modeling setup:
  - Tradable sector firm maximizes profits choosing output and inputs; imported foreign inputs denoted Z*; factor costs include wage w, real interest r, and imported input price s*e.
  - Production function exhibits constant returns to scale.
- Mechanisms through which the real exchange rate (e) affects sectoral labor demand:
  - Competitiveness channel: a real appreciation (increase in e) reduces competitiveness for exporters, lowering output and employment in export-oriented tradable sectors (captured by price elasticity ηe < 0).
  - Imported-input channel: a real appreciation lowers the domestic price of imported inputs (s*e), reducing marginal costs and potentially increasing employment where imported input shares are sizable.
- Net effect on tradable employment is ambiguous ex ante and depends on the relative strength of competitiveness (negative) and imported-input (positive) channels.
- Non-tradable sectors:
  - By definition, demand depends only on domestic income; thus the real exchange rate affects non-tradable labor demand only via the imported-input channel, implying a priori positive employment effects of a real appreciation.
- Log-linearized equilibrium relationships used for estimation:
  - Tradable employment and wage equations (schematic form): equations (8) and (9) where coefficients cT3;1, dT3;1 < 0 capture negative competitiveness effect and cT3;2, dT3;2 > 0 capture positive imported-input effect.
  - Non-tradable employment and wage equations (schematic form): equations (10) and (11) where coefficients cN2, dN2 > 0 capture imported-input channel.
- Cross-sectional heterogeneity exploited:
  - Export share Xs,t influences sectoral price elasticity with respect to exchange rates: ηe s,t = k Xs,t (higher export share → larger |ηe|).
  - Imported input share (1 − α − β) amplifies the positive imported-input channel.

### III. Data — sources and key series
- Data sources: China’s National Bureau of Statistics (NBS), provincial/municipal Statistical Yearbooks, Census and Economic Information Center (CEIC), International Financial Statistics (IFS), World Economic Outlook (WEO).
- Real effective exchange rate (REER):
  - Series from IFS, computed as a weighted average of bilateral nominal exchange rates and price index differentials; weights updated periodically and derived from trade shares.
  - Time pattern noted:
    - 1980–1994: persistent real depreciation by more than 100 percent.
    - 1994–2002: relatively mild appreciation and stabilization.
  - Nominal USD/RMB nominal rate accounts for most movement in the REER.
  - Institutional context:
    - 1980–1994: multiple-exchange-rate regime (official, commercial, swap rates); strong depreciating trend across rates; capital and current account transactions largely controlled by the government.
    - 1994: abolition of multiple rate system and consolidation to a nominal rate around 8.2 Yuan/Dollar from 1994 until 2005.
- Empirical samples:
  - Regressions use different sub-samples; although long samples starting from 1980 are examined, most analysis uses data starting 1994 or later to avoid regime-change driven spurious results.
- Sectoral data:
  - Aggregate sectoral employment data from 1980 onward used in descriptive analysis; data quality and disaggregation are limited at the aggregate sectoral level.
- Input-output and regional employment linkage:
  - The paper links regional input-output tables with regional sector-level employment to test transmission channels (highest degree of disaggregation in the analysis).

### IV. Empirical strategy and methods (summary)
- Estimation approaches:
  - Panel OLS and Instrumental Variable estimation to address potential endogeneity of the real exchange rate.
  - Seemingly Unrelated Regression (SUR) to control for correlation across industries and regions.
  - Two-stage least squares (2SLS) and General Method of Moments (GMM) used in disaggregate exercises exploiting regional-sector variation.
- Robustness checks:
  - Inclusion of aggregate, industry-specific, and regional control variables.
  - Stability checks across different sample periods (including pre- and post-1994 regimes).
- Key empirical variables highlighted for heterogeneity tests:
  - Sectoral and regional export shares (measure of external exposure).
  - Sectoral and regional imported-input shares (measure of cost-pass-through via imported inputs).
  - Regional openness and intermediate-input linkages between services and manufacturing.

### V. Policy-relevant implications drawn in the paper
- Exchange rate movements have significant spillovers to sectors without direct international exposure, complicating policy assessment focused only on tradable/manufacturing sectors.
- A revaluation (real appreciation) can reduce employment across both tradable and certain non-tradable sectors, implying broader labor-market adjustment costs than standard comparative-static predictions focusing on relative prices might suggest.
- Regional heterogeneity and input-output linkages matter: regions and service sectors with stronger intermediate links to tradables face amplified adverse employment impacts from appreciation.
- Policy design challenge emphasized: measures to minimize welfare costs of employment adjustments need to account for inter-sectoral production linkages and regional exposure profiles rather than focusing solely on export-oriented manufacturing.

*Source: Excerpts from the paper’s Introduction, Theory, and Data sections (content unit: _wp11148 - References).*

### 2002. This series was discontinued after 2002 as the NBS changed the sector classification as

### _wp11148 - 2002. This series was discontinued after 2002 as the NBS changed the sector classification as

### Sectoral employment data — old classification (Employmentdata)
- Uses the old sectoral employment data series discontinued after 2002.
- Definition of employed persons: includes all persons of 16 years of age and above who are engaged in economic activities and earn remuneration for more than one hour in the reference week (in line with ILO convention).
- Number of sectors: 16 sectors in this old classification system.
- Dataset reference name used here: Employmentdata.
- Advantages:
  - Consistent ILO-aligned employment definition.
  - No structural break at 1990 because the sectoral data used does not include the NBS adjustment based on the 5th national population census.
  - Recommended by Young (2003) for better consistency across years.

### Sectoral real wage data
- Real average wage series by sector: Real Wage Index series corresponding to the old sector classification system (series label CGAHT from the CEIC database).

### Sectoral employment data — new classification (Employees series)
- New classification introduced in 2003, roughly in line with ISIC Rev. 3; extended back to data from 1990 for consistency.
- Series name: Employees (referred to here as the Employeesseries).
- Definition and limitation:
  - Labeled ”total number of employees in urban units”.
  - Only includes employees of state-owned units, urban collective-owned units and other ownership units.
  - Excludes workers in private enterprises and self-employed individuals — groups that constitute the most dynamic part of the Chinese labor market in recent past.
- Coverage: Available until 2008, whereas the old Employmentseries is available only until 2002.
- Use in analysis: Employed for robustness checks and to explore systematic differences between state-owned sector adjustment and aggregate economy.

### Measurement of sectoral exposure to real exchange rate fluctuations
- Effective exposure constructed by interacting the aggregate real effective exchange rate (REER) with sector-specific export and import shares.
- Methodology precedent: Used by Campa and Goldberg (2001), Gourinchas (1998).
- Data availability constraints:
  - Sector-specific trade values available starting in 1993 from the China Statistical Yearbook.
  - Combining sector-specific trade values with the old Employmentdata yields too short a sample.
  - No corresponding sector-level output series covering the same period; NBS records real and nominal output only for broadly defined sectors (e.g., industry sector includes mining, manufacturing, production and supply of utilities).
- Adjustment for output-based trade shares:
  - In the sectoral analysis using the Employeesseries, sectors are grouped based on availability of output data so the sample can be extended to 2008 and corresponding trade weights computed.

### Descriptive findings on employment and REER co-movement
- Figure-based observations (based on first/old sector classification):
  - Employment series exhibit very strong trends; growth rates are displayed in the top chart of Figure 3.
  - Strong co-movement among sectors.
  - Negative correlation between sectoral employment growth and REER: employment in all sectors appears to grow stronger in periods of real depreciation and vice versa.
  - After linear detrending (bottom chart of Figure 3), co-movement among sectors and negative relation with REER are even more pronounced for detrended series.
- Observations using the Employees series under the new sector classification (Figure 4):
  - Strong co-movement among sectors persists.
  - Negative correlation of sectoral employment growth with REER also observed for the Employees series (noting this series excludes private units and self-employed individuals).

### Regional data
- Third dataset: broad-industry-level employment data (primary, secondary, tertiary — see Table 1) reported by different regions in China.
- Data sources and coverage:
  - CEIC China Database for period 1993 to 2008.
  - Supplemented by China Statistical Yearbook editions for 1988 to 1992.
- Regional scope and treatment of Chongqing:
  - As of the data context, there are 31 regions in China (including 22 provinces, 5 autonomous regions, and 4 directly administered municipalities).
  - Chongqing became a directly administered municipality in 1997; prior to that it was part of Sichuan province.
  - To maintain consistency with earlier sample, Chongqing data merged with Sichuan after 1996.
  - Resulting regional coverage used: data for 30 regions from 1985 to (end of provided excerpt).

*Source: _wp11148 - 2002. This series was discontinued after 2002 as the NBS changed the sector classification as*

### 2008. For each region and year, we have annual data on total employment in each of the 3

### _wp11148 - 2008. For each region and year, we have annual data on total employment in each of the 3

### Data, scope, and methodology
- Datasets and coverage
  - Regional employment data by industry (primary, secondary, tertiary) and regional exports/imports for years up to 2008.
  - Sectoral datasets: "old Employment" dataset with 15 sectors covering 1980 to 2002; "Employees" dataset with 7 sectors covering 1989 to 2008 (urban unit employees).
  - National input-output tables available for 1995, 1997, 2000, 2002, and 2005.
  - Regional input-output tables available for 2002 covering 30 regions (not including Tibet); Chongqing merged with Sichuan to match employment data.
  - Detailed regional sector-level employment collected from provincial Statistical Yearbooks for 2003 to 2008.
  - Final detailed panel for input-output analysis: 28 regions (Tibet, Hainan, Shandong excluded; Chongqing merged with Sichuan), 16 non-tradable sectors, 6 years (2003–2008).

- Empirical strategy and key estimators
  - Most series failed unit-root rejection by univariate Dickey-Fuller and panel unit-root tests; no evidence of panel cointegration (Westerlund (2007)), so first differences used in many specifications.
  - Panel regressions with lagged dependent variables to capture gradual employment response.
  - Estimation methods: OLS, GLS, SUR (Seemingly Unrelated Regressions), 2SLS (instrumenting REER), system GMM (Blundell and Bond (1998)) for dynamic region-sector panel.
  - Instruments for aggregate REER: change in Japanese Yen/US dollar nominal exchange rate and US 3-month T-Bill rate.
  - Regional-specific REERs constructed by interacting aggregate REER with lagged regional export share X_{r,t-1} and import share M_{r,t-1} (both as percentage of regional GDP).
  - Regional input-output linkage measure: shr_i = p^N_i Q^N_i / (p^N Q^N) = share of non-tradable gross output used as intermediate input in tradable sectors (from input-output tables).

### Main empirical findings — sectoral (aggregate) analysis
- REER coefficients (sectoral regressions)
  - Table 3 (old Employment, 15 sectors, 1980–2002): estimated REER coefficients negative in all columns with values ranging from -0.04 to -0.14.
    - Interpretation: an appreciation (depreciation) of the real exchange rate by 10 percent decreases (increases) employment growth by 0.4 to 1.4 percentage points on average across all sectors.
  - Table 4 (Employees dataset, 7 sectors, 1989–2008):
    - A 10 percent real appreciation (depreciation) leads to about 0.7 percent decrease (increase) in employment growth (broadly consistent with Table 3).
    - For urban unit employees, REER interacted with tradable dummy significantly negative in sub-samples while symmetric REER becomes insignificant (columns 2 and 5).
    - Sector-specific REERs (REER interacted with sector-specific lagged export shares) show sectors with higher export shares experience a stronger negative effect from real appreciation.

- Heterogeneity and symmetry
  - The negative effect of real appreciation is not restricted to tradable sectors only; both tradable and non-tradable sectors show negative responses in several specifications.
  - Tradable-interacted REER sometimes insignificant in pooled specifications (Table 3) but significant when using urban employees dataset and sector-specific REERs (Table 4).

### Main empirical findings — regional analysis
- Baseline regional industry regressions (three broad sectors: primary, secondary, tertiary)
  - Primary sector: REER has no significant effect on regional employment.
  - Secondary industry (tradable/manufacturing/mining): appreciated REER has negative effect on employment growth; estimate less significant (10 percent level) and comparable in magnitude to aggregate sectoral results.
  - Tertiary industry (non-tradable/services): REER coefficient negative and significant at 5 percent in recent sample; negative effect slightly more pronounced than for tradable sectors.

- SUR and dynamic patterns
  - SUR results (Table 8): REER negative and significant at 1 percent for Industry 2 and Industry 3.
    - Quantified impacts: a 10 percent real appreciation (depreciation) reduces (increases)
      - Industry 2 employment by 1.1 percent.
      - Industry 3 employment by 1.2 to 1.3 percent.
  - Wald test: cannot reject equality of REER coefficients across Industry 2 and Industry 3 (very high p-value >50 percent).
  - Table 9 (dynamics): most adjustment in tertiary sector occurs in same year; secondary sector adjustment more persistent.
  - Table 11 (asymmetry): the negative effect of appreciation matters most for employment contraction in both tradable and non-tradable industries (both whole sample and 1994–2008).

- Instrumental variables and exogeneity tests
  - First-stage correlations: stronger US dollar against JPY associated with real effective appreciation; correlation between 0.6 to 0.7.
  - Increase in US interest rate associated with strengthening Chinese REER (except for industry 3).
  - Hansen test: cannot reject validity of instruments.
  - 2SLS second-stage results close to OLS; Hausman-Wu test cannot reject exogeneity of REER.

### Transmission channels — input-output and intermediate-input mechanism
- Theoretical extension
  - Non-tradable gross output used for direct consumption Q^N_c at price p^N_c and as intermediate input for tradable production Q^N_i at price p^N_i.
  - Derived elasticity (equation (14)): ∂L_N/L_N ∂e/e = p^N α [ η^e_i shr_i  - (1-αβ) (∂Q_N/∂Z'_N)^{-1} ]
    - shr_i = p^N_i Q^N_i / (p^N Q^N) is share of non-tradable output used as intermediate input in tradable sectors.
    - η^e_i < 0 is elasticity of intermediate input demand w.r.t. REER.
    - Prediction: larger shr_i amplifies negative effect of appreciation on non-tradable employment via intermediate-input channel.

- Input-output evidence and empirical test
  - Aggregate/region-level input-output facts:
    - Large share (30 to 70 percent for some services) of gross output in banking, transport, wholesale/retail trade and utilities is used in tradable production.
    - Regional input-output tables show intermediate usage shares in these service sectors on average in range 30 to 60 percent.
    - Positive correlation between tradable sector share in regional gross output and regional intermediate usage of wholesale/retail trade in tradable production (Figure 7).
  - Regression results using region-sector panel (2003–2008, 28 regions, 16 non-tradable sectors):
    - Baseline equation ln L_irt = γ_0 + μ_r + γ_1 ln Y_rt + γ_2 ln H_rt + γ_3 Z_t + γ_4 ln e_t * shr_ir + ε_irt.
    - Column 1 (no interaction): appreciation leads to lower non-tradable employment (significant at 10 percent).
    - Column 2 onwards: REER interacted with region-sector shr_ir strongly significant; non-interacted REER becomes insignificant.
    - Including time dummies (column 3) does not alter coefficient of interest.
    - GLS with region-sector random component (column 4) and GLS allowing AR(1) (column 5) confirm results; AR(1) estimate ≈ 0.513.
    - Interacting REER with lagged regional export share (as in SUR) yields insignificant coefficient (column 6), suggesting export-interaction effect is driven by input-output interdependence.
    - System GMM (column 7) in first differences with lagged employment and population treated as predetermined: negative REER × shr_ir effect remains significant at 5 percent; Arellano-Bond and Hansen tests support specification.
    - Table 13 (lagged regressors): REER through intermediate-input channel even stronger; GMM estimate of REER term ≈ -0.2 and significant at 1 percent.
      - Economic interpretation: for non-tradable sectors with an average intermediate input share to tradables, a 10 percent real appreciation leads to 0.9 percent lower employment growth within 2 years.

### Robustness, magnitudes, and interpretation
- Robustness across datasets and methods: results consistent across sectoral datasets (old Employment, Employees), regional OLS, SUR, IV (2SLS), GLS, and system GMM.
- Magnitudes to preserve exactly:
  - REER coefficients in Table 3: range from -0.04 to -0.14.
  - 10 percent appreciation → 0.4 to 1.4 percentage points change (Table 3 interpretation).
  - Employees dataset: 10 percent appreciation → about 0.7 percent decrease in employment growth.
  - SUR full-sample: Industry 2 employment change for 10 percent appreciation = -1.1 percent; Industry 3 = -1.2 to -1.3 percent.
  - First-order autocorrelation detected: 0.513 (GLS AR(1) estimate).
  - Instrument correlations: 0.6 to 0.7 between Japanese Yen/USD movement and Chinese REER.
  - GMM/lagged regressors: REER term ≈ -0.2; 10 percent appreciation → 0.9 percent lower employment growth within 2 years for average shr_i.

- Interpretation challenges
  - Finding that non-tradable employment contracts after a real appreciation contradicts standard theory (which predicts expansion of non-tradables).
  - Reallocation frictions alone insufficient to explain contraction in non-tradable employment.
  - Intermediate-input linkage provides a plausible short-run mechanism: contraction in tradable sectors reduces demand for services used as intermediate inputs, leading to employment losses in services.
  - Other channels (learning spillovers, technology diffusion) may operate over longer horizons and are not ruled out.

### Policy implications and recommendations
- Short-run risks of revaluation
  - A revaluation that induces a real appreciation can cause employment contraction across both tradable and non-tradable sectors.
  - Short-term adverse effects on employment imply limited immediate boost to domestic demand following appreciation.

- Policy packages to accompany appreciation
  - Macroeconomic policy measures to support domestic demand should accompany any revaluation that causes real appreciation.
  - Structural reforms to enhance productivity in non-manufacturing sectors to reduce their dependence on tradable-sector demand and to facilitate labor reallocation.
  - Consider expansion in supply of services—particularly social services and health care—to absorb displaced labor (referenced policy discussion in IMF (2010) and Blanchard and Giavazzi (2005) in source text).

- Research and data needs
  - Further analysis of welfare consequences requires deeper understanding of transmission channels across sectors.
  - Need for additional micro-level data (e.g., firm-level data) and investigation of alternative channels (learning, technology diffusion) to fully assess medium-term and long-term impacts.

*Source: _wp11148 - 2008. For each region and year, we have annual data on total employment in each of the 3 (IMF working paper content provided).*

### REFERENCES

### _wp11148 - REFERENCES

### References
- Aizenman, J., and J. Lee, 2008, “The real exchange rate, mercantilism, and the learning-by-doing externality,”NBER Working Paper, No. 13853.
- Bayoumi, T., J. Lee, and S. Jayanthi, 2005, “New Rates from New Weights,”IMF Working Paper, No. 05/99.
- Blanchard, O., and F. Giavazzi, 2005, “Rebalancing Growth in China: A Three-Handed Approach,”MIT Working Paper Series, No. 05/32.
- Blundell, R., and S. Bond, 1998, “Initial Conditions and Moment Conditions in Dynamic Panel Data Models,”Journal of Econometrics, Vol. 87, pp. 115-143.
- Brandt, L., and T.G. Rawski, 2008,China’s Great Economic Transformation(Cambridge University Press).
- Branson, S., and J. Love, 1988, “United States Manufacturing and the Real Exchange Rate,” Misalignment of Exchange Rates: Effects on Trade and Industry.R. Marston (edt.), University of Chicago Press.
- Burgess, S., and M. Knetter, 1998, “An International Comparison of Employment Adjustment to Exchange Rate Fluctuations,”Review of International Economics, Vol. 4, No. 2, pp. 121-138.
- Campa, J.M., and L.S. Goldberg, 2001, “Employment versus Wage Adjustment and the US Dollar,”The Review of Economics and Statistics, Vol. 83, No. 3, pp. 477-489.
- CICC, 2010, “RMB Outlook and industry implications,”Macroeconomy Report, (March). China International Capital Corporation Ltd.: Beijing.
- Frankel, J., and S.J. Wei, 2007, “Assessing China’s Exchange Rate Regime,”Economic Policy, (April).
- Goldberg, L., J. Tracy, and S. Aaronson, 1999, “Exchange Rates and Employment Instability: Evidence from Matched CPS Data,”American Economic Review, Vol. 89, No. 2, pp. 204-210.
- Goldberg, L.S., and J. Tracy, 2000, “Exchange Rates and Local Labor Markets,” International Trade and Wages.Robert Feenstra (edt.), Chicago University Press: Chicago.
- Gourinchas, P.O., 1998, “Exchange Rates and Jobs: What Do We Learn from Job Flows?,” NBER Macroeconomics Annual, Vol. 13, pp. 153-222.
- Hua, P., 2007, “Real exchange rate and manufacturing employment in China,”China Economic Review, Vol. 18, pp. 335-353.
- IMF, 2010, “People’s Republic of China, Selected Issues,”Asia Pacific Department. International Monetary Fund: Washington, DC.
- McKinnon,R., and G. Schnabl, 2008, “China’s Exchange Rate Impasse and the Weak U.S. Dollar,”CESifo Working Paper, No. 2386.
- Moser, C., D. Urban, and B. Weder di Mauro, 2010, “International competitiveness, job creation and job destruction - An establishment-level study of German job flows,” Journal of International Economics, Vol. 80, pp. 302-317.
- Nucci, F., and A.F. Pozzolo, 2010, “The exchange rate, employment and hours: What firm-level data say,”Journal of International Economics, Vol. 82, pp. 112-123.
- Rajan, R., and A. Subramanian, 2008, “Aid and Manufacturing Growth,”Journal of Development Economics.forthcoming.
- Revenga, A. L., 1992, “Exporting Jobs? The impact of Import Competition on Employment and Wages in U.S. Manufacturing,”The Quarterly Journal of Economics, Vol. 107, No. 1, pp. 255-284.
- Westerlund, J., 2007, “Testing for Error Correction in Panel Data,”Oxford Bulletin of Economics and Statistics, Vol. 69, No. 6, pp. 709-748.
- Young, A., 2003, “Gold into Base Metals: Productivity growth in PRC during the Reform period,”Journal of Politial Economy, Vol. 111, No. 6, pp. 1220-1261.

### Appendix: Data
- Industry and sector classifications in China
  - China classifies its economy into three industries: primary industry, secondary industry, and tertiary industry. Under each industry, the economy is separated into different sectors.
  - In 2003, China began to use the new sector classification standards, which contain 19 2-bit code sectors and differ from the old sector classification standards, which contain 16 2-bit code sectors.
  - Our sector-level analysis uses old sector classification covering period from 1980 to

*Source: _wp11148 - REFERENCES*

### 2002. In order to have longer series of data, we group some sectors for broad sector-level

### _wp11148 - 2002. In order to have longer series of data, we group some sectors for broad sector-level

### Data sources, definitions, and construction
- Period covered for broad sector-level analysis: 1980 to 2008.
- Definition of tradable sectors: agriculture, manufacture, and mining.
- Employment and employees:
  - Employment data from NBS and CEIC (unit: million person).
  - Total employed persons at year-end by sector: available from 1980 to 2002.
  - Employment by region and industry: available from 1985 to 2008.
  - Employees (urban units, excluding private enterprises and self-employed): available from 1988 to 2008 (NBS).
- Wages:
  - Real wage indices by sector (old classification) from 1980 to 2002 from CEIC (1978 = 100).
  - CEIC provides regional real wage indices from 1993 to 2008.
  - For broad sector analysis, real wage constructed as weighted average of nominal wage by sector from NBS deflated by CPI; employment numbers used as weights.
- GDP:
  - Regional GDP indices and broad sectoral GDP indices from 1980 to 2008 (CEIC and NBS).
  - GDP at current prices by sector from 1980 to 2002 from NBS (unit: billion RMB).
- Working age population:
  - Demographics from UN Population Database; five-year data interpolated to annual.
  - Working age population defined as population aged 15 to 64.
- Exchange rate:
  - Real effective exchange rate (REER) from IFS, constructed as weighted average of exchange rates with trading partners, adjusted for relative consumer price changes.
- Interest rates:
  - Lending rate from IFS deflated by CPI to obtain (ex-post) real interest rates.
  - US interest rate: annual average of the 3-month T-Bill rate.
- Oil price:
  - Average oil prices from WEO (published on October 2009); relative oil price computed by deflating with China’s CPI.
- World imports:
  - World imports in billions of USD from WEO (published on October 2009).

### Chinese industry and sector classification (summary)
- Old and new classifications map primary, secondary, tertiary industries to sectoral labels (examples preserved from source):
  - Primary Industry — Agriculture: Agriculture, Forestry, Animal Husbandry, and Fishery.
  - Secondary Industry — Mining; Manufacturing; Production and Supply of Electricity, Gas, and Water; Construction.
  - Tertiary Industry — Transport, Storage and Post; Wholesale and Retail Trades and Catering Services; Financial Intermediation (Banking); Real Estate; Services to Households and Other Services; Health Care, Sports and Social Welfare; Information Transmission, Computer Services, and Software; Education, Culture and Arts, Radio, Film and Television; Scientific Research, Technical Services and Geologic Prospecting; Public Management and Social Organizations; Culture, Sports and Entertainment.

### Regional summary statistics (selected)
- National averages (sample period mostly 1985 to 2008):
  - Avg. employment growth (National): -0.001
  - Avg. real output growth (National): 0.031
  - Avg. Ind.1: 0.049
  - Avg. Ind.2: 0.040
  - Avg. Ind.3: 0.111
  - Avg. Ind.2 (alternate column): 0.097
  - Avg. export share (National): 0.242
- Example provinces (averages taken over available sample period; export shares from 1993 to 2008):
  - Guangdong: Avg. employment growth -0.006; Avg. real output growth 0.051; Ind.1 0.044; Ind.2 0.053; Ind.3 0.157; Ind.2 (alt) 0.134; Export share 0.822
  - Jiangsu: Avg. employment growth -0.036; Avg. real output growth -0.010; Ind.1 0.036; Ind.2 0.025; Ind.3 0.100; Ind.2 (alt) 0.119; Export share 0.531
  - Shanghai: Avg. employment growth -0.036; Avg. real output growth -0.010; Ind.1 0.036; Ind.2 0.025; Ind.3 0.100; Ind.2 (alt) 0.119; Export share 0.531
  - Note: Figures for Sichuan include Chongqing.

### Key regression results — cross-sector and regional panels (selected coefficients and statistics preserved exactly as reported)

- Table 3. Cross-sector panel regression for employment growth: 1980 - 2002, 16 sectors
  - Sample sizes: N = 352 (columns 1–2); N = 128 (columns 3–6). No. sectors = 16.
  - China’s real GDP coefficients (columns reproduced): 0.357 (0.127), 0.357 (0.127), 1.719 (0.000), 1.719 (0.000), 1.564 (0.074), 2.396 (0.313).
  - Working age population coefficients: 2.159 (0.440), 2.158 (0.438), -3.962 (0.000), -3.962 (0.000), -3.690 (0.330), -5.961 (1.395).
  - Relative oil price coefficients: -0.013 (0.013), -0.013 (0.013), -0.031 (0.000), -0.031 (0.000), -0.032 (0.002), -0.033 (0.010).
  - Real interest rate coefficients: 0.039 (0.069), 0.039 (0.070), 0.077 (0.000), 0.077 (0.000), 0.057 (0.013), 0.180 (0.055).
  - Note: Standard errors in parentheses; significance indicators shown in original table. Columns (4), (5) and (6) include a linear trend and no population growth due to lack of (census) data.

- Table 4. Cross-sector panel regression for employment growth: 1990 - 2008, 8 sectors
  - N = 144 (columns 1–3); N = 112 (columns 4–6). No. sectors = 8.
  - REER interactions:
    - REER X tradable dummy coefficients: -0.095 (0.032) and -0.192 (0.058) reported (significance indicated in table).
    - REER X lagged sectoral export shares: -0.264 (0.091), -0.271 (0.124).
  - Lagged sectoral employment coefficients: 0.276 (0.073), 0.225 (0.070), 0.207 (0.068), 0.124 (0.041), 0.120 (0.055), 0.117 (0.047).
  - China’s real GDP: 0.501 (0.249), 0.530 (0.243), 0.575 (0.236), 1.456 (0.143), 1.371 (0.189), 1.460 (0.175).
  - Real interest rate: -0.412 (0.310), -0.432 (0.303), -0.507 (0.294), -1.181 (0.132), -1.079 (0.172), -1.059 (0.164).

- Table 5. Regional panel regression by industry OLS: 1988-2008 (examples)
  - N = 592 or 548 across columns; adj. R2 reported: 0.210, 0.386, 0.326, 0.211, 0.420, 0.340.
  - Emp Ind i (total employment in industry i): 0.406 (0.085), 0.007 (0.119), -0.702 (0.308), 0.415 (0.096), 0.022 (0.118), -0.599 (0.315).
  - Regional real GDP: -0.015 (0.020), 0.125 (0.051), 0.026 (0.077), 0.003 (0.023), 0.118 (0.063), -0.038 (0.085).
  - World demand coefficient examples: 0.033 (0.061), 0.295 (0.060), 0.031 (0.070), 0.315 (0.087).
  - Real interest rate: 0.039 (0.073), 0.117 (0.075), -0.758 (0.121).

- Table 7. Regional panel regression by industry using 2 SLS: 1994-2008
  - Columns (1)-(3): N = 410, 410, 390; adj. R2 = 0.212, 0.444, 0.397.
  - REER coefficients: 0.003 (0.030), -0.096* (0.055), -0.147** (0.066).
  - Emp Ind i: 0.068 (0.226), 0.207 (0.193), 0.893*** (0.310).
  - Regional real GDP: -0.078* (0.041), 0.301*** (0.081), 0.078 (0.083).
  - First stage reported coefficients for instruments:
    - YEN/USD: 0.638 (0.043), 0.708 (0.042), 0.567 (0.042).
    - US Interest rate: 0.030 (0.003), 0.036 (0.003), -0.008 (0.004).
  - Overid. (p-val): 0.585, 0.452, 0.186. Wu-Hausman (p-val): 0.362, 0.886, 0.574.

- Tables 8–11. SUR and dynamics (selected highlights)
  - SUR results for employment growth rates (I–IV) report multiple specifications for periods 1988-2008 and 1994-2008 with Reg Emp Ind 1/2/3.
  - Examples of REER effects:
    - Table 8 (1988-2008): REER coefficients: -0.038* (0.021), -0.111*** (0.028), -0.150*** (0.036).
    - Table 8 (1994-2008): REER coefficients: 0.019 (0.021), -0.105*** (0.040), -0.133*** (0.036).
    - Table 9 (1994-2008) includes REER lags: REER (t-1) coefficients: 0.048* (0.025), -0.070 (0.050), -0.019 (0.049); REER (t-2): 0.063*** (0.024), -0.144*** (0.043), 0.025 (0.041).
    - Table 11 (1988-2008) REER X appreciation: 0.298** (0.131), -0.433** (0.170), -0.466** (0.217).
  - Real interest rate often shows large coefficients with significance in several specifications (examples):
    - Table 8 (1994-2008): Real interest rate for Reg Emp Ind 1: -0.117*** (0.042), Reg Emp Ind 2: 0.184** (0.082), Reg Emp Ind 3: -0.208* (0.122).
    - Table 10 and 11 report additional significant interest rate interactions and asymmetric REER effects (appreciation vs depreciation).

- Tables 12–13. Results with regional I-O tables and sectoral employment (contemporaneous and lagged regressors)
  - Dependent variable: ln L_Nir;t (log sector-region employment).
  - Key coefficient: REER * shr_ir (sector-region-specific intermediate input to tradables as share of gross output) reported consistently:
    - OLS1-GLS3 (contemporaneous): REER * shr_ir = -0.216  (0.036), -0.216  (0.036), -0.224  (0.031), -0.214  (0.067), -0.228  (0.029), -0.156  (0.063).
    - Lagged regressors (REER(t-1) * shr_ir): -0.223  (0.036), -0.223  (0.036), -0.220  (0.032), -0.216  (0.067), -0.226  (0.031), -0.193  (0.066).
  - ln L_Nir;t-1 coefficients: 0.936  (0.031) in contemporaneous; 0.950  (0.029) in lagged specification.
  - GMM specification statistics:
    - AR(2): 0.369 (contemporaneous), 0.367 (lagged).
    - rho_hat (ˆρ) = 0.513 (contemporaneous), ˆρ = 0.464 (lagged).
    - Hansen p-values: 0.180 (contemporaneous), 0.279 (lagged).

### Intermediate input usage (input-output evidence)
- Table 14. National Intermediate Input Usage in Tradable Sectors (selected years)
  - Banking and insurance: 0.426 (1995), 0.372 (1997), 0.348 (2000), 0.304 (2002), 0.310 (2005).
  - Construction: 0.004 (1995), 0.009 (1997), 0.008 (2000), 0.005 (2002), 0.007 (2005).
  - Transportation and Post: 0.399 (1995), 0.423 (1997), 0.357 (2000), 0.381 (2002), 0.384 (2005).
  - Wholesale and retail trades: 0.548 (1995), 0.394 (1997), 0.364 (2000), 0.327 (2002), 0.360 (2005).
  - Utility: 0.706 (1995), 0.730 (1997), 0.738 (2000), 0.590 (2002), 0.607 (2005).
  - Note: Each cell equals the share of gross output value of each input sector that serves as intermediate input for the tradable sectors (NBS Input-Output Tables, authors’ calculations).

- Table 15. Intermediate input usage in tradables by region in 2002 (selected rows and summary)
  - Examples (Banking, Transport, Trade, Utility, % of Tradables):
    - Anhui: 0.262, 0.423, 0.331, 0.669, 0.587
    - Beijing: 0.087, 0.192, 0.081, 0.726, 0.366
    - Guangdong: 0.228, 0.375, 0.276, 0.438, 0.607
    - Jiangsu: 0.387, 0.511, 0.512, 0.413, 0.678
    - Zhejiang: 0.500, 0.535, 0.422, 0.662, 0.675
  - Mean: 0.379, 0.415, 0.292, 0.585
  - Min: 0.087, 0.192, 0.041, 0.346
  - Max: 0.972, 0.628, 0.528, 0.932
  - Note: Figures for Sichuan include Chongqing.

- Table 16. Regional distribution of nontradable intermediate input shares in tradable sectors (summary statistics)
  - Banking and insurance: Mean 0.304, Std. Dev. 0.098, Min. 0.087, Max. 0.51
  - Construction: Mean 0.004, Std. Dev. 0.006, Min. 0.00, Max. 0.027
  - IT Services: Mean 0.139, Std. Dev. 0.116, Min. 0.025, Max. 0.629
  - Utilities: Mean 0.481, Std. Dev. 0.105, Min. 0.289, Max. 0.688
  - Transportation & post: Mean 0.395, Std. Dev. 0.129, Min. 0.192, Max. 0.595
  - Wholesale/Retail trade, hotel/catering: Mean 0.29, Std. Dev. 0.134, Min. 0.041, Max. 0.528

### Substantive analytical findings (as evidenced in regressions and I-O analysis)
- REER effects on employment growth are heterogeneous across sectors and regions:
  - Cross-sector and SUR specifications report negative and statistically significant REER effects for certain industries (examples include REER coefficients of -0.111*** and -0.150*** in Table 8).
  - Interaction terms reveal that REER impacts depend on tradability and the sector-region intermediate input linkages:
    - REER * shr_ir coefficients are consistently negative and highly significant in multiple specifications (e.g., -0.216  (0.036) contemporaneous; -0.223  (0.036) lagged).
- Sectoral and regional demand drivers:
  - China’s real GDP generally enters positively and significantly in cross-sector specifications (examples: 1.719  (0.000) in Table 3; 1.456  (0.143) in Table 4).
  - World demand shows positive and sometimes significant coefficients in several specifications (e.g., 0.139  (0.068) in Table 3; 0.108  (0.024) in Table 4).
- Role of intermediate inputs:
  - The share of nontradable intermediate inputs supplied to tradable sectors varies substantially across regions and input sectors, and the REER effect is amplified where these input linkages are stronger (negative REER * shr_ir coefficients).
- Dynamics and persistence:
  - Strong persistence in sector-region employment: ln L_Nir;t-1 coefficients near 0.94–0.95 in dynamic specifications (0.936  (0.031); 0.950  (0.029)).
- Interest rate and oil price:
  - Real interest rate and relative oil price appear with varying signs and significance across specifications; in several regional specifications real interest rate coefficients are large and significant (examples in Tables 8–11).

*Source: content from the provided PDF chapter/section _wp11148 (tables and explanatory notes reproduced exactly as presented in the source).*

### 1. Logarithm of real effective exchange rate

### 1. Logarithm of real effective exchange rate

### Time series scale and axis labels
- Y-axis tick labels recorded: 6, 5.8, 5.6, 5.4, 5.2, 5, 4.8, 4.6, 4.4, 4.2, 4, 4.
- X-axis years displayed: 1980, 1982, 1984, 1986, 1988, 1990, 1992, 1994, 1996, 1998, 2000, 2002, 2004, 2006, 2008.
- Data source: IFS.

### Key presentation features
- Multiple repetitions of the Y-axis ticks (6 down to 4) and the years 1980–2008 indicate overlaid or repeated panel plotting of Ln(REER) series.
- Caption: "Source: IFS".

---

### 2. Logarithms of real effective exchange rate and average USD/RMB rate

### Variables plotted
- Ln(REER)
- Ln(USD/RMB)

### Axes and scales
- Y-axis tick labels include negative values: -2.5, -2, -1.5, -1, -0.5, 0.
- Secondary Y-axis tick labels include 6, 5.8, 5.6, 5.4, 5.2, 5, 4.8, 4.6, 4.4, 4.2, 4.
- X-axis years displayed: 1980, 1982, 1984, 1986, 1988, 1990, 1992, 1994, 1996, 1998, 2000, 2002, 2004, 2006, 2008.
- Caption: "Source: IFS".

### Presentation notes
- The figure overlays Ln(REER) and Ln(USD/RMB) on shared time axis 1980–2008 with distinct vertical scales (one including negative values down to -2.5).
- Caption repeated: "Source: IFS".

---

### 3. Sectoral employment and the real exchange rate

### Variables and measures
- Series plotted (growth rates, detrended): REER, Manuf, Construction, Transport, Trade.
- Growth rate axis labels: -40%, -30%, -20%, -10%, 0%, 10%, 20%, 30%.
- Detrended axis tick labels include: 0.30, 0.40 and -0.50, -0.40, -0.30, -0.20, -0.10, 0.00, 0.10, 0.20, 0.30, 0.40.
- Caption: "Source: CEIC China Database, IFS".

### Key presentation features
- The figure presents growth-rate distributions (from -40% to 30%) alongside detrended values (from -0.50 to 0.40) for sectoral employment relative to REER movements.
- Legend recorded: "REER Manuf Construction Transport Trade".

---

### 4. Sectoral employees (urban units) and the real exchange rate

### Variables and measures
- Series plotted (growth rates, detrended): REER, Industry, Construction, Transport, Service, Banking.
- Growth rate axis labels: -40%, -30%, -20%, -10%, 0%, 10%, 20%.
- Year axis: 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008.
- Detrended axis tick labels include: 0.30, 0.40 and -0.30, -0.20, -0.10, 0.00, 0.10, 0.20, 0.30, 0.40.
- Caption: "Source: CEIC China Database, IFS".

### Key presentation features
- Time series 1990–2008 for urban-sector employment growth rates and detrended components relative to REER.
- Legend recorded: "REER Industry Construction Transport Service Banking".

---

### 5. China Regional Map

- Figure label: "Figure5. China Regional Map".
- Caption: "Source: China Data Online".

---

### 6. Employment growth in secondary and tertiary industry for selected regions

### Regions and variables
- Regions shown: Guangdong, Shanghai, Jiangxi, Fujian, Tianjin, Yunnan, Zhejiang, Anhui.
- Series labels: EmpInd2, EmpInd3.
- Year axis: 1994, 1996, 1998, 2000, 2002, 2004, 2006, 2008.
- Growth rate axes vary by panel; examples included: -10%, 0%, 10%, 20%, 30%, 40%, 50%; -30%, -20%, -10%, 0%, 10%, 20%; -5%, 0%, 5%, 10%, 15%; -40%, -20%, 0%, 20%, 40%.
- Caption: "Source: CEIC China Database."

### Key presentation features
- Panel plots for selected provinces showing EmpInd2 (secondary industry employment growth) and EmpInd3 (tertiary industry employment growth) over 1994–2008.
- Value ranges differ by province panel, indicating heterogeneity in employment growth series.

---

### 7. Intermediate input to tradables of the Wholesale/Retail sector vs. regional openness

### Variables and axes
- Vertical axis: "Int-Input(trade)/GOV(trade)" with tick values: 0, .1, .2, .3, .4, .5.
- Horizontal axis: "GOV(tradables)/GOV" with tick annotations ".3.4.5.6.7" present in the figure text.
- Listed regions plotted: anhui, beijing, fujian, gansu, guangdong, guangxi, guizhou, hainan, hebei, heilongjiang, henan, hubei, hunan, innermongolia, jiangsu, jiangxi, jilin, liaoning, ningxia, qinghai, shaanxi, shandong, shanghai, shanxi, sichuan, tianjin, xinjiang, yunnan, zhejiang.
- Data source: China input-output tables 2002.

### Caption and explanatory note
- Vertical axis shows intermediate input into tradable sectors as a share of the gross output value of the Wholesale/Retail/Catering Trade sector.
- Horizontal axis shows the tradable sectors’ gross output value as a share of regional total gross output value.
- Caption: "Source: China input-output tables 2002."

*Source: _wp11148 - 1. Logarithm of real effective exchange rate (figures and captions extracted from the supplied PDF content).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11148.pdf_