## wp1709 — References

## Source details

**Canonical URL:** [wp1709 — References](https://www.imf.org/-/media/files/publications/wp/wp1709.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/wp1709.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/wp1709.pdf.json)

---

### I. Introduction — scope and key conclusion
- Purpose: assess price and wage flexibility in Hong Kong SAR from aggregate and industry-level evidence, compared with the United States, the United Kingdom and Singapore.
- Main conclusion: price and wage adjustments in Hong Kong SAR are as flexible as in Singapore, the United States or United Kingdom by most measures, and are even more flexible by some measures.
- Supporting context and prior studies:
  - Pauwels and Zhang (2008): extent of wage flexibility in Hong Kong SAR is similar to the United States.
  - Cheng and Ho (2009): wages and prices in Hong Kong SAR are flexible relative to other economies such as Canada.
  - Lai, Ha, and Leung (2002): prices and wages adjust more strongly in Hong Kong SAR than in Argentina.

### II. Aggregate evidence — three macroeconomic relationships
- Methods: compare Hong Kong SAR, Singapore, the United States and United Kingdom using:
  - Phillips curve (labor earnings growth vs. unemployment gap),
  - correlation between inflation and output gaps,
  - Okun’s Law (unemployment gap vs. output gap).
- Key numerical findings:
  - Inflation vs. output gap:
    - Hong Kong SAR: a 1-percent increase of the positive output gap corresponds to a 0.14 percent rise in inflation.
    - United States: 0.13 percent rise in inflation.
    - Singapore: 0.11 percent rise in inflation.
    - Interpretation: Hong Kong SAR’s price adjustment is slightly more flexible than the United States and Singapore by this measure.
  - Labor earnings growth vs. output gap (year-on-year comparisons):
    - Hong Kong SAR: a 1-percent increase of the positive output gap corresponds to a 0.49 percent increase in the growth rate of labor earnings.
    - Hong Kong SAR: inflation response reported as 0.34 percent (for comparison in same paragraph).
    - United States: labor earnings response to 1-percent output gap is about 0.28 percent; inflation response cited as 0.4 percent.
    - Interpretation: in Hong Kong SAR price adjustment is to a larger extent driven by change in labor costs compared with the United States.
  - Labor earnings growth vs. unemployment gap (Phillips-variant; 1990–2015):
    - Hong Kong SAR: a one-percent increase in the unemployment gap corresponds to a 1.8 percent drop in the growth rate of labor earnings.
    - United States: 0.7 percent drop.
    - United Kingdom: 1.5 percent drop.
    - Singapore: 3.1 percent drop.
    - Interpretation: labor cost adjustment more flexible in Hong Kong SAR than United States and United Kingdom, but less flexible than Singapore.
  - Unemployment gap vs. output gap (Okun’s Law; sample 1990:Q1–2015:Q4 except United Kingdom from 1995:Q1):
    - Hong Kong SAR: a 1-percent increase in the positive output gap corresponds to a 0.26 percent drop in the unemployment gap.
    - United Kingdom: 0.23 percent drop.
    - Singapore: 0.09 percent drop.
    - Note: elasticity larger in the United States than Hong Kong SAR; result may be influenced by volatility differences in output growth.
    - Caution: literature shows less consensus on linking Okun coefficient and labor market flexibility.

### III. Industry-level evidence — labor earnings distributions and skewness
- Rationale: downward nominal wage rigidity implies a pile-up near-zero earnings growth and a relatively short left tail; distributional shape (discontinuity at zero, skewness) is diagnostic.
- Data and samples:
  - Hong Kong SAR: industry-level year-on-year nominal earnings by quarter, sample covers 61 industries from 2000:Q1 to 2004:Q4 and 46 industries from 2005:Q1 to 2016:Q1.
  - Coverage: about 97 percent of total employees during 2000–04 and about 98 percent during 2005–16.
  - United States: 31 industries, representing about 84 percent of total nonfarm employment in 2015; time horizon 2000:Q1 to 2016:Q1.
  - Weighting: industry share of employees used as sample weight.
- Histogram and qualitative distributional findings:
  - Hong Kong SAR:
    - Distribution does not present substantial discontinuity around zero (bar [0, 1] not much higher than bar [-1, 0]).
    - Distribution has a long left tail; about one-third of observations are negative.
    - Interpretation: consistent with a flexible labor market.
  - United States:
    - Distribution exhibits more discontinuity around zero: bar [-1, 0] substantially lower than bar [0,1], indicating the existence of downward wage rigidity.
    - Hong Kong SAR distribution is more dispersed than U.S., possibly reflecting more volatile business cycles.
- Skewness (symmetry) results — low inflation years only:
  - Low inflation years defined as years with annual inflation rates less than 2 percent.
  - Skewness removes top/bottom outliers (top/bottom 1 percent and 5 percent variants).
  - Table 1 weighted distribution skewness:
    - Sample selection 1–99 percentile, 5–95 percentile
    - Hong Kong SAR: 0.17, 0.11
    - United States: 0.12, 0.12
  - Interpretation: Hong Kong SAR skewness is on par with the United States; distributions are broadly as symmetric as the United States; smaller skewness implies less downward rigidity.

### IV. Estimating downward nominal wage rigidity — empirical model and results
- Model structure (Kimura and Ueda (2001) framework):
  - Firms face a lower bound d on nominal wage changes; fraction x of firms can fully adjust.
  - Observed average nominal wage growth dw(i,t):
    - dw(i, t) = dw*(i,t) if dw*(t) > d
    - dw(i, t) = x∙dw*(i,t) + (1-x)∙d otherwise
  - Equilibrium (flexible) wage growth dw*(i,t) specification:
    - dw*(i,t) = α u(i,t) + β inf(t) + γ g(i,t) + ε(i,t)
    - where u(i,t) = industry-level unemployment rates, inf(t) = aggregate inflation rates, g(i,t) = industry-level real GDP growth rates.
  - Parameters estimated jointly: x, d, α, β, γ using nonlinear least squares.
- Methodological notes:
  - Interpretation of parameters: larger x implies a larger proportion of firms capable of adjusting wages; smaller d implies firms less likely to be constrained by downward rigidity. Special case d = 0 implies firms cannot have negative wage adjustment.
  - Advantages: model controls for output fluctuations, inflation, and labor market conditions in explaining observed earnings growth distributions.
- Empirical results summary:
  - Signs of α, β and γ are as expected: labor earnings growth negatively correlated with unemployment rates, positively correlated with inflation rates and real GDP growth rates.
  - Regression results indicate growth of labor earnings in Hong Kong SAR responds more strongly to inflation and real GDP growth than the United States, consistent with aggregate evidence.
  - Overall model implication: downward wage rigidity is less binding in Hong Kong SAR than in the United States.

### V. Additional methodological and data remarks
- Choice of indicator: used “earnings” rather than “wage” because earnings include bonus and compensation for extra hours which respond to the business cycle.
- Seasonality and frequency: year-on-year growth of nominal earnings by quarter documented because official labor earnings series in Hong Kong SAR is not seasonally adjusted.
- Industry classification change: number of industries differs between sub-periods due to change from HISC 1.1 to HISC 2.0.

### Section II — Wage adjustment: key estimates and interpretation
- Estimated fraction of firms able to adjust wages over one year (coefficient x):
  - Hong Kong SAR: 0.53
    - 95% confidence interval: [0.51, 0.54]***
  - United States: 0.50
    - 95% confidence interval: [0.49, 0.51]***
- Estimated floor of annual wage growth (coefficient d):
  - Hong Kong SAR: 0.58
    - 95% confidence interval: [0.47, 0.68]**
  - United States: 2.04
    - 95% confidence interval: [2.00, 2.07]***
- Other regression coefficients (earnings growth regression):
  - Unemployment (α):
    - Hong Kong SAR: -0.13
      - 95% confidence interval: [-0.13, -0.12]***
    - United States: -0.13
      - 95% confidence interval: [-0.13, -0.12]***
  - Inflation (β):
    - Hong Kong SAR: 1.10
      - 95% confidence interval: [1.09, 1.11]***
    - United States: 0.50
      - 95% confidence interval: [0.50, 0.50]***
  - Industrial real GDP growth (γ):
    - Hong Kong SAR: 0.1624408
      - 95% confidence interval: [0.16, 0.17]***
    - United States: 0.0507812
      - 95% confidence interval: [0.05, 0.05]***
  - Constant term:
    - Hong Kong SAR: 0.06
      - 95% confidence interval: [0.01, 0.11]***
    - United States: 2.66
      - 95% confidence interval: [2.65, 2.67]***
- Interpretation:
  - About half of firms can freely change wages in both Hong Kong SAR and the United States (x ~ 0.53 vs. 0.50).
  - The much smaller floor of wage adjustment in Hong Kong SAR (d = 0.58 percent) versus the United States (d = 2.04 percent) suggests wage adjustment is more flexible in Hong Kong SAR.

### Consumer price adjustment: micro-level distributions
- Sample and coverage:
  - Quarterly price index data from 2000:Q1 to 2016:Q2.
  - Hong Kong SAR: 81 items, representing about 81 percent of the consumption basket.
  - United States: 41 items, representing about 89 percent of the consumption basket.
- Distributional features (all items):
  - For Hong Kong SAR, frequency of price changes in [-1, 0] is substantially smaller than in [0, 1], indicating some downward rigidity.
  - The extent of discontinuity around zero for Hong Kong SAR is on par with the United States.
- Measured density differences around zero (price-change "Density Difference around Zero" = probability(price change in (0%,1%)) − probability(price change in (-1%,0%))):
  - Values reported across panels: 0.039, 0.043, 0.050, 0.044.
- Skewness of price-change distributions during low inflation years (all items):
  - Sample selection 1-99 percentile:
    - Hong Kong SAR: -0.22
    - United States: -0.41
  - Sample selection 5-95 percentile:
    - Hong Kong SAR: -0.31
    - United States: -0.33
  - Interpretation: depending on outlier trimming (top/bottom 1% or 5%), the United States skewness is smaller or on par with Hong Kong SAR.

### Service items vs. all items: separating domestic and foreign influences
- Rationale: service-item price changes are more affected by domestic factors; goods (nonservice) items are more exposed to foreign factors (e.g., import prices, fuel).
- Skewness of price-change distributions (service items only), low inflation years:
  - Sample selection 1-99 percentile:
    - Hong Kong SAR: -0.48
    - United States: -0.17
  - Sample selection 5-95 percentile:
    - Hong Kong SAR: -0.45
    - United States: -0.35
- Interpretation:
  - After removing nonservice items, the skewness for Hong Kong SAR is smaller than for the United States regardless of outlier trimming, implying price adjustment for service items is likely more flexible in Hong Kong SAR.
  - This finding is consistent with the conclusion that Hong Kong SAR has a more flexible labor market.

### Aggregate and industry-level implications
- Aggregate evidence:
  - Broadly, price and wage adjust more strongly in Hong Kong SAR than in Singapore, the United States and the United Kingdom, which helps stabilize real aggregates such as output and employment.
- Industry-level evidence:
  - Distribution of labor earnings growth in Hong Kong SAR does not concentrate at zero and is not more positively skewed relative to the United States.
  - Estimates from the wage formation model suggest downward wage rigidity is less binding in Hong Kong SAR.
  - Price adjustments of service items (with more domestic content) are more flexible in Hong Kong SAR than in the United States.

### Data used for the downward wage rigidity model (sources and coverage)
- Hong Kong SAR:
  - Source: Census and Statistics Department of Hong Kong SAR via CEIC.
  - Time horizon: 2001:Q1 to 2016:Q1.
  - Industry-level wage growth dw(i,t): YoY change of quarterly industry-level nominal average payroll index (YoY used to eliminate seasonality).
  - Aggregate inflation and industry-level quarterly real GDP growth: YoY changes of CPI and industry-level real value-added.
  - Industries included (10 non-overlapping sectors):
    1. manufacturing
    2. import and export trade
    3. wholesale and retail trades
    4. services and accommodation, transport, storage, postal & courier
    5. information and communications
    6. financial services and insurance
    7. real estate
    8. professional and business services
    9. public administration
    10. social and personal services
  - Industry weights: share of real value-added in 2015.
- United States:
  - Industry-level unemployment rate and labor earnings growth: BLS via CEIC.
  - Industry-level real GDP: BEA.
  - Inflation series: Haver.
  - Time horizon: 2005:Q1 to 2016:Q1 (industry real value-added available from 2005).
  - Industries included (12):
    1. mining
    2. construction
    3. durable goods manufacturing
    4. nondurable goods manufacturing
    5. wholesale and retail trades
    6. transportation and warehousing
    7. information
    8. finance, insurance, real estate, rental and leasing
    9. professional and business services
    10. education services, health care and social assistance
    11. arts, entertainment, recreation, accommodation, and food services
    12. other nongovernment services
  - Industry weights: share of real value-added in 2015.
- Note: Only industries with accessible unemployment rate, payroll and real value-added series were included; therefore the number of industries here is smaller than in Section III.B.

*Source: Section II, wp1709.*

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

### wp1709 - References

### I. Introduction — scope and key conclusion
- Purpose: assess price and wage flexibility in Hong Kong SAR from aggregate and industry-level evidence, compared with the United States, the United Kingdom and Singapore.
- Main conclusion: price and wage adjustments in Hong Kong SAR are as flexible as in Singapore, the United States or United Kingdom by most measures, and are even more flexible by some measures.
- Supporting context and prior studies:
  - Pauwels and Zhang (2008): extent of wage flexibility in Hong Kong SAR is similar to the United States.
  - Cheng and Ho (2009): wages and prices in Hong Kong SAR are flexible relative to other economies such as Canada.
  - Lai, Ha, and Leung (2002): prices and wages adjust more strongly in Hong Kong SAR than in Argentina.

### II. Aggregate evidence — three macroeconomic relationships
- Methods: compare Hong Kong SAR, Singapore, the United States and United Kingdom using:
  - Phillips curve (labor earnings growth vs. unemployment gap),
  - correlation between inflation and output gaps,
  - Okun’s Law (unemployment gap vs. output gap).
- Key numerical findings (preserve exact values):
  - Inflation vs. output gap:
    - Hong Kong SAR: a 1-percent increase of the positive output gap corresponds to a 0.14 percent rise in inflation.
    - United States: 0.13 percent rise in inflation.
    - Singapore: 0.11 percent rise in inflation.
    - Interpretation: Hong Kong SAR’s price adjustment is slightly more flexible than the United States and Singapore by this measure.
  - Labor earnings growth vs. output gap (year-on-year comparisons):
    - Hong Kong SAR: a 1-percent increase of the positive output gap corresponds to a 0.49 percent increase in the growth rate of labor earnings.
    - Hong Kong SAR: inflation response reported as 0.34 percent (for comparison in same paragraph).
    - United States: labor earnings response to 1-percent output gap is about 0.28 percent; inflation response cited as 0.4 percent.
    - Interpretation: in Hong Kong SAR price adjustment is to a larger extent driven by change in labor costs compared with the United States.
  - Labor earnings growth vs. unemployment gap (Phillips-variant; 1990–2015):
    - Hong Kong SAR: a one-percent increase in the unemployment gap corresponds to a 1.8 percent drop in the growth rate of labor earnings.
    - United States: 0.7 percent drop.
    - United Kingdom: 1.5 percent drop.
    - Singapore: 3.1 percent drop.
    - Interpretation: labor cost adjustment more flexible in Hong Kong SAR than United States and United Kingdom, but less flexible than Singapore.
  - Unemployment gap vs. output gap (Okun’s Law; sample 1990:Q1–2015:Q4 except United Kingdom from 1995:Q1):
    - Hong Kong SAR: a 1-percent increase in the positive output gap corresponds to a 0.26 percent drop in the unemployment gap.
    - United Kingdom: 0.23 percent drop.
    - Singapore: 0.09 percent drop.
    - Note: elasticity larger in the United States than Hong Kong SAR; result may be influenced by volatility differences in output growth.
    - Caution: literature shows less consensus on linking Okun coefficient and labor market flexibility.

### III. Industry-level evidence — labor earnings distributions and skewness
- Rationale: downward nominal wage rigidity implies a pile-up near-zero earnings growth and a relatively short left tail; distributional shape (discontinuity at zero, skewness) is diagnostic.
- Data and samples:
  - Hong Kong SAR: industry-level year-on-year nominal earnings by quarter, sample covers 61 industries from 2000:Q1 to 2004:Q4 and 46 industries from 2005:Q1 to 2016:Q1.
  - Coverage: about 97 percent of total employees during 2000–04 and about 98 percent during 2005–16.
  - United States: 31 industries, representing about 84 percent of total nonfarm employment in 2015; time horizon 2000:Q1 to 2016:Q1.
  - Weighting: industry share of employees used as sample weight.
- Histogram and qualitative distributional findings:
  - Hong Kong SAR:
    - Distribution does not present substantial discontinuity around zero (bar [0, 1] not much higher than bar [-1, 0]).
    - Distribution has a long left tail; about one-third of observations are negative.
    - Interpretation: consistent with a flexible labor market.
  - United States:
    - Distribution exhibits more discontinuity around zero: bar [-1, 0] substantially lower than bar [0,1], indicating the existence of downward wage rigidity.
    - Hong Kong SAR distribution is more dispersed than U.S., possibly reflecting more volatile business cycles.
- Skewness (symmetry) results — low inflation years only:
  - Low inflation years defined as years with annual inflation rates less than 2 percent.
  - Skewness removes top/bottom outliers (top/bottom 1 percent and 5 percent variants).
  - Table 1 weighted distribution skewness (exact values preserved):
    - Sample selection: 1–99 percentile, 5–95 percentile
    - Hong Kong SAR: 0.17, 0.11
    - United States: 0.12, 0.12
  - Interpretation: Hong Kong SAR skewness is on par with the United States; distributions are broadly as symmetric as the United States; smaller skewness implies less downward rigidity.

### IV. Estimating downward nominal wage rigidity — empirical model and results
- Model structure (Kimura and Ueda (2001) framework):
  - Firms face a lower bound d on nominal wage changes; fraction x of firms can fully adjust.
  - Observed average nominal wage growth dw(i,t):
    - dw(i, t) = dw*(i,t) if dw*(t) > d
    - dw(i, t) = x∙dw*(i,t) + (1-x)∙d otherwise
  - Equilibrium (flexible) wage growth dw*(i,t) specification:
    - dw*(i,t) = α u(i,t) + β inf(t) + γ g(i,t) + ε(i,t)
    - where u(i,t) = industry-level unemployment rates, inf(t) = aggregate inflation rates, g(i,t) = industry-level real GDP growth rates.
  - Parameters estimated jointly: x, d, α, β, γ using nonlinear least squares.
- Key methodological notes:
  - Interpretation of parameters: larger x implies a larger proportion of firms capable of adjusting wages; smaller d implies firms less likely to be constrained by downward rigidity. Special case d = 0 implies firms cannot have negative wage adjustment.
  - Advantages: model controls for output fluctuations, inflation, and labor market conditions in explaining observed earnings growth distributions.
- Empirical results summary (qualitative, as exact parameter estimates beyond table mention are not provided in the supplied content):
  - Signs of α, β and γ are as expected: labor earnings growth negatively correlated with unemployment rates, positively correlated with inflation rates and real GDP growth rates.
  - Regression results indicate growth of labor earnings in Hong Kong SAR responds more strongly to inflation and real GDP growth than the United States, consistent with aggregate evidence.
  - Overall model implication: downward wage rigidity is less binding in Hong Kong SAR than in the United States.

### V. Additional methodological and data remarks
- Choice of indicator: used “earnings” rather than “wage” because earnings include bonus and compensation for extra hours which respond to the business cycle.
- Seasonality and frequency: year-on-year growth of nominal earnings by quarter documented because official labor earnings series in Hong Kong SAR is not seasonally adjusted.
- Industry classification change: number of industries differs between sub-periods due to change from HISC 1.1 to HISC 2.0.

*Source: wp1709 - References.*

### Section II.

### Section II.

### Wage adjustment: key estimates and interpretation
- Estimated fraction of firms able to adjust wages over one year (coefficient x):
  - Hong Kong SAR: 0.53
    - 95% confidence interval: [0.51, 0.54]***
  - United States: 0.50
    - 95% confidence interval: [0.49, 0.51]***
- Estimated floor of annual wage growth (coefficient d):
  - Hong Kong SAR: 0.58
    - 95% confidence interval: [0.47, 0.68]**
  - United States: 2.04
    - 95% confidence interval: [2.00, 2.07]***
- Other regression coefficients (from the earnings growth regression):
  - Unemployment (α):
    - Hong Kong SAR: -0.13
      - 95% confidence interval: [-0.13, -0.12]***
    - United States: -0.13
      - 95% confidence interval: [-0.13, -0.12]***
  - Inflation (β):
    - Hong Kong SAR: 1.10
      - 95% confidence interval: [1.09, 1.11]***
    - United States: 0.50
      - 95% confidence interval: [0.50, 0.50]***
  - Industrial real GDP growth (γ):
    - Hong Kong SAR: 0.1624408
      - 95% confidence interval: [0.16, 0.17]***
    - United States: 0.0507812
      - 95% confidence interval: [0.05, 0.05]***
  - Constant term:
    - Hong Kong SAR: 0.06
      - 95% confidence interval: [0.01, 0.11]***
    - United States: 2.66
      - 95% confidence interval: [2.65, 2.67]***
- Interpretation:
  - About half of firms can freely change wages in both Hong Kong SAR and the United States (x ~ 0.53 vs. 0.50).
  - The much smaller floor of wage adjustment in Hong Kong SAR (d = 0.58 percent) versus the United States (d = 2.04 percent) suggests wage adjustment is more flexible in Hong Kong SAR.

### Consumer price adjustment: micro-level distributions
- Sample and coverage:
  - Quarterly price index data from 2000:Q1 to 2016:Q2.
  - Hong Kong SAR: 81 items, representing about 81 percent of the consumption basket.
  - United States: 41 items, representing about 89 percent of the consumption basket.
- Distributional features (all items):
  - For Hong Kong SAR, frequency of price changes in [-1, 0] is substantially smaller than in [0, 1], indicating some downward rigidity.
  - The extent of discontinuity around zero for Hong Kong SAR is on par with the United States.
- Measured density differences around zero (price-change "Density Difference around Zero" = probability(price change in (0%,1%)) − probability(price change in (-1%,0%))):
  - Values reported across panels: 0.039, 0.043, 0.050, 0.044 (as shown in Figure 6).
- Skewness of price-change distributions during low inflation years (all items):
  - Sample selection 1-99 percentile:
    - Hong Kong SAR: -0.22
    - United States: -0.41
  - Sample selection 5-95 percentile:
    - Hong Kong SAR: -0.31
    - United States: -0.33
  - Interpretation: depending on outlier trimming (top/bottom 1% or 5%), the United States skewness is smaller or on par with Hong Kong SAR.

### Service items vs. all items: separating domestic and foreign influences
- Rationale: service-item price changes are more affected by domestic factors; goods (nonservice) items are more exposed to foreign factors (e.g., import prices, fuel).
- Skewness of price-change distributions (service items only), low inflation years:
  - Sample selection 1-99 percentile:
    - Hong Kong SAR: -0.48
    - United States: -0.17
  - Sample selection 5-95 percentile:
    - Hong Kong SAR: -0.45
    - United States: -0.35
- Interpretation:
  - After removing nonservice items, the skewness for Hong Kong SAR is smaller than for the United States regardless of outlier trimming, implying price adjustment for service items is likely more flexible in Hong Kong SAR.
  - This finding is consistent with the conclusion that Hong Kong SAR has a more flexible labor market.

### Aggregate and industry-level implications
- Aggregate evidence:
  - Broadly, price and wage adjust more strongly in Hong Kong SAR than in Singapore, the United States and the United Kingdom, which helps stabilize real aggregates such as output and employment.
- Industry-level evidence:
  - Distribution of labor earnings growth in Hong Kong SAR does not concentrate at zero and is not more positively skewed relative to the United States.
  - Estimates from the wage formation model suggest downward wage rigidity is less binding in Hong Kong SAR.
  - Price adjustments of service items (with more domestic content) are more flexible in Hong Kong SAR than in the United States.

### Data used for the downward wage rigidity model (sources and coverage)
- Hong Kong SAR:
  - Source: Census and Statistics Department of Hong Kong SAR via CEIC.
  - Time horizon: 2001:Q1 to 2016:Q1.
  - Industry-level wage growth dw(i,t): YoY change of quarterly industry-level nominal average payroll index (YoY used to eliminate seasonality).
  - Aggregate inflation and industry-level quarterly real GDP growth: YoY changes of CPI and industry-level real value-added.
  - Industries included (10 non-overlapping sectors):
    1. manufacturing
    2. import and export trade
    3. wholesale and retail trades
    4. services and accommodation, transport, storage, postal & courier
    5. information and communications
    6. financial services and insurance
    7. real estate
    8. professional and business services
    9. public administration
    10. social and personal services
  - Industry weights: share of real value-added in 2015.
- United States:
  - Industry-level unemployment rate and labor earnings growth: BLS via CEIC.
  - Industry-level real GDP: BEA.
  - Inflation series: Haver.
  - Time horizon: 2005:Q1 to 2016:Q1 (industry real value-added available from 2005).
  - Industries included (12):
    1. mining
    2. construction
    3. durable goods manufacturing
    4. nondurable goods manufacturing
    5. wholesale and retail trades
    6. transportation and warehousing
    7. information
    8. finance, insurance, real estate, rental and leasing
    9. professional and business services
    10. education services, health care and social assistance
    11. arts, entertainment, recreation, accommodation, and food services
    12. other nongovernment services
  - Industry weights: share of real value-added in 2015.
- Note: Only industries with accessible unemployment rate, payroll and real value-added series were included; therefore the number of industries here is smaller than in Section III.B.

*Source: Section II, wp1709.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/wp1709.pdf_