## ANNEX I. ANALYSIS OF INFORMALITY SHARES ## Source details **Canonical URL:** [ANNEX I. ANALYSIS OF INFORMALITY SHARES](https://www.imf.org/-/media/files/research/imf-and-g20/2022/g20-minimizing-scarring-from-the-pandemic-technical-annex.pdf) ## Other formats - [Markdown version](/-/media/files/research/imf-and-g20/2022/g20-minimizing-scarring-from-the-pandemic-technical-annex.pdf.md) - [Structured JSON version](/-/media/files/research/imf-and-g20/2022/g20-minimizing-scarring-from-the-pandemic-technical-annex.pdf.json) --- ### Context and purpose - Analysis supports the 2022 G-20 Background Note on Minimizing Scarring from the Pandemic and reports the data and methods behind Figures 6 and 7 of that publication. ### Data and country coverage - Countries analyzed: Brazil, Indonesia, Mexico, Russia, South Africa. - Data sources: - Brazil: National Household Sample Survey (PNAD). - Indonesia: Indonesia Statistics Office (BPS). - Mexico: National Survey of Occupation and Employment (ENOE). - Russia: Labor Force Survey (Statistics Russia). - South Africa: Quarterly Labor Force Survey (Statistics South Africa). ### Definitions of informality (country-specific) - Brazil: - Informal workers are those employees who do not hold a formal labor contract, those who are self-employed, and those who are domestic workers employed by households. - Indonesia: - Informal workers are those whose main job status is not defined as ”employers of formal workers” and “workers/employees” as provided by the Indonesia Statistics Office. - Mexico: - Informal workers are those who work in the informal sector as defined by INEGI, unpaid workers, salaried domestic workers working for households, and workers that are not enrolled or make no contributions to the social security system. - Russia: - Informal workers are persons employed by at least one informal sector unit of production during the reference period, irrespective of employment status or whether the job was primary or additional; informal sector units are defined by absence of government registration as a legal entity. - Informal sector employment includes: sole entrepreneurs; persons employed by sole entrepreneurs and individuals; family members assisting in a family business; persons working individually without registration as a sole entrepreneur; persons working as part of household production of agricultural, forestry, hunting and fishing goods intended to be sold or exchanged; persons working as part of household production of household goods intended to be sold or exchanged. - South Africa: - Informal workers are (i) employees working in establishments that employ fewer than five employees and do not deduct income tax from their salaries/wages; and (ii) employers, own-account workers, and persons helping without pay in their household business and who are not registered for either income tax or value-added tax. ### Measurement and decomposition of informality shares - Primary statistic: share of informal employment over total employment, using the country-specific informality definitions above. - Decomposition identity (change in informal-employment share): - I_t/E_t − I_{t−1}/E_{t−1} = sum_{s∈S_c} (E_{s,t−1}/E_{t−1}) * (I_{s,t}/E_{s,t} − I_{s,t−1}/E_{s,t−1}) + sum_{s∈S_c} (I_{s,t−1}/E_{s,t−1}) * (E_{s,t}/E_{t} − E_{s,t−1}/E_{t−1}) + sum_{s∈S_c} (I_{s,t}/E_{s,t} − I_{s,t−1}/E_{s,t−1}) * (E_{s,t}/E_{t} − E_{s,t−1}/E_{t−1}) - Interpretation of components: - Within‑sector component: sum_{s∈S_c} (E_{s,t−1}/E_{t−1}) * (I_{s,t}/E_{s,t} − I_{s,t−1}/E_{s,t−1}). - Between‑sector component: sum_{s∈S_c} (I_{s,t−1}/E_{s,t−1}) * (E_{s,t}/E_{t} − E_{s,t−1}/E_{t−1}). - Remaining (cross‑product) component: cross product between changes in employment shares and changes in informality rates within sectors; noted as small in practice and omitted from the chart. ### Sector aggregation and NACE codes - Sector aggregation varies by country (aggregated NACE sectors). Country-specific groupings: - Brazil and Mexico: A, BDE, C, F, G I, H, KLMN, PQ, RST, OU. - Indonesia: A, B, C, D, E, F, G, H, I, J, K, L, MN, O, P, Q, RSTU. - Russia: BDE, C, F, GI, HJ, KLMN, PQ, RSTUO. - South Africa: B, C, DE, F, GI, H, KLMN, RST. - NACE sector codes (used for aggregation): - A Agriculture, forestry and fishing - B Mining and quarrying - C Manufacturing - D Electricity, gas, steam, and air conditioning supply - E Water supply; sewerage, waste management and remediation activities - F Construction - G Wholesale and retail trade; repair of motor vehicles and motorcycles - H Transportation and storage - I Accommodation and food service activities - J Information and communication - K Financial and insurance activities - L Real estate activities - M Professional, scientific, and technical activities - N Administrative and support service activities - O Public administration and defense; compulsory social security - P Education - Q Human health and social work activities - R Arts, entertainment, and recreation - S Other service activities - T Activities of households as employers; undifferentiated goods- and services-producing activities of households for own use - U Activities of extraterritorial organizations and bodies --- ### 17. Columns 4 to 6 in Table A III.3: Specification, results, and interpretation ### Specification and comparison - Specification compares the relationship between employment growth and labor productivity with the measure of sectoral reallocation held at the median level. - Comparison across economies at the 20th percentile and 80th percentile of job retention policy spending. - Estimated effects are qualitatively similar to those reported in the main text (Figure 17, right panel). ### Figure AIII.1 (captioned results) - Sources: Orbis; OECD; IMF staff calculations. - Outcome displayed: Change in employment growth from a 10 percent increase in labor productivity (percent; policy spending at 20th and 80th percentiles respectively, median sectoral reallocation). - Displayed percentiles: 20th percentile, 80th percentile. - Vertical scale values shown: 0.0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8. ### Selected regression coefficients and statistics (columns (1)–(6) where reported) - Dependent variable: DEmp (reported for six regressions). - Constant coefficients and standard errors: - Column (1): -0.684*** (0.0279) - Column (2): -0.676*** (0.0291) - Column (3): -0.684*** (0.0277) - Column (4): -0.687*** (0.0275) - Column (5): -0.676*** (0.0284) - Column (6): -0.686*** (0.0275) - lag(labprod) coefficients and standard errors: - Column (1): 0.0777*** (0.00498) - Column (2): 0.0630*** (0.00357) - Column (3): 0.0715*** (0.00371) - Column (4): 0.0724*** (0.00377) - Column (5): 0.0640*** (0.00266) - Column (6): 0.0711*** (0.00320) - Interaction terms (selected coefficients and standard errors): - lag(labprod)*lag(reten): -0.0300*** (0.0113) and -0.0236** (0.0113) - lag(labprod)*lag(realloc): -0.00932 (0.00739) and 0.00434 (0.00790) - lag(labprod)*lag(reten)*(emp realloc US): -0.0966*** (0.0277) and -0.0932*** (0.0270) - lag(labprod)*lag(realloc)*(emp realloc US): -0.00807 (0.0135) and 0.00884 (0.0103) ### Sample size, fit, and estimation details - Sample sizes (N) by column: - Column (1): 8,488,366 - Column (2): 10,066,956 - Column (3): 8,553,921 - Column (4): 8,579,656 - Column (5): 10,042,533 - Column (6): 8,652,979 - R-squared: 0.054 0.054 0.054 0.054 0.054 0.054 (reported for columns (1) through (6)) - Fixed effects: Country-sector FE: YES; Country-year FE: YES for all columns. - Standard errors: Clustered standard errors (country-sector) reported in parentheses. - Significance notation: *** p<0.01, ** p<0.05, * p<0.1. ### Key interpretation points reported - The interaction of lagged labor productivity with measures of job retention (lag(reten)) and sectoral reallocation influences employment growth responses to a 10 percent productivity shock, with magnitude and significance varying by specification and by inclusion of an "emp realloc US" term. - Effects evaluated at the 20th and 80th percentiles of job retention policy spending indicate heterogeneous impacts on employment growth given median sectoral reallocation. *International Monetary Fund, Annex I to the technical annex of the 2022 G-20 Background Note on Minimizing Scarring from the Pandemic.* ### ANNEX I. ANALYSIS OF INFORMALITY SHARES_____________________________________________________ 3 ### ANNEX I. ANALYSIS OF INFORMALITY SHARES ### Context - The analysis supports the 2022 G-20 Background Note on Minimizing Scarring from the Pandemic and reports the data and methods behind Figures 6 and 7 of that publication. ### Data - Countries analyzed: Brazil, Indonesia, Mexico, Russia, South Africa. - Data sources: - Brazil: National Household Sample Survey (PNAD). - Indonesia: Indonesia Statistics Office (BPS). - Mexico: National Survey of Occupation and Employment (ENOE). - Russia: Labor Force Survey (Statistics Russia). - South Africa: Quarterly Labor Force Survey (Statistics South Africa). - Definitions of informality by country: - Brazil: informal workers are those employees who do not hold a formal labor contract, those who are self-employed, and those who are domestic workers employed by households. - Indonesia: informal workers are those whose main job status is not defined as ”employers of formal workers” and “workers/employees” as provided by the Indonesia Statistics Office. - Mexico: informal workers are those who work in the informal sector as defined by INEGI, unpaid workers, salaried domestic workers working for households, and workers that are not enrolled or make no contributions to the social security system. - Russia: informal workers are persons employed by at least one informal sector unit of production during the reference period, irrespective of employment status or whether the job was primary or additional; informal sector units are defined by absence of government registration as a legal entity. Informal sector employment includes: sole entrepreneurs; persons employed by sole entrepreneurs and individuals; family members assisting in a family business; persons working individually without registration as a sole entrepreneur; persons working as part of household production of agricultural, forestry, hunting and fishing goods intended to be sold or exchanged; persons working as part of household production of household goods intended to be sold or exchanged. - South Africa: informal workers are (i) employees working in establishments that employ fewer than five employees and do not deduct income tax from their salaries/wages; and (ii) employers, own-account workers, and persons helping without pay in their household business and who are not registered for either income tax or value-added tax. ### Results — Measurement of Informality Shares - Primary statistic reported: share of informal employment over total employment, using the country-specific informality definitions above. - Decomposition of change in informal-employment share (notation): - I_t/E_t − I_{t−1}/E_{t−1} = sum_{s∈S_c} (E_{s,t−1}/E_{t−1}) * (I_{s,t}/E_{s,t} − I_{s,t−1}/E_{s,t−1}) + sum_{s∈S_c} (I_{s,t−1}/E_{s,t−1}) * (E_{s,t}/E_{t} − E_{s,t−1}/E_{t−1}) + sum_{s∈S_c} (I_{s,t}/E_{s,t} − I_{s,t−1}/E_{s,t−1}) * (E_{s,t}/E_{t} − E_{s,t−1}/E_{t−1}) - Interpretation of decomposition components: - Within‑sector component: sum over sectors of (E_{s,t−1}/E_{t−1}) times the change in informality rate within each sector: (I_{s,t}/E_{s,t} − I_{s,t−1}/E_{s,t−1}). - Between‑sector component: sum over sectors of (I_{s,t−1}/E_{s,t−1}) times the change in sectoral employment shares: (E_{s,t}/E_{t} − E_{s,t−1}/E_{t−1}). - Remaining (cross‑product) component: cross product between changes in employment shares and changes in informality rates within sectors; noted as small in practice and omitted from the chart. - Sector aggregation: - Sums over sectors are taken over country‑specific aggregates (S_c), where sector aggregation varies by country depending on data availability (aggregated NACE sectors). - Country-specific sector groupings used: - Brazil and Mexico: A, BDE, C, F, G I, H, KLMN, PQ, RST, OU. - Indonesia: A, B, C, D, E, F, G, H, I, J, K, L, MN, O, P, Q, RSTU. - Russia: BDE, C, F, GI, HJ, KLMN, PQ, RSTUO. - South Africa: B, C, DE, F, GI, H, KLMN, RST. ### NACE sector codes (used for aggregation) - A Agriculture, forestry and fishing - B Mining and quarrying - C Manufacturing - D Electricity, gas, steam, and air conditioning supply - E Water supply; sewerage, waste management and remediation activities - F Construction - G Wholesale and retail trade; repair of motor vehicles and motorcycles - H Transportation and storage - I Accommodation and food service activities - J Information and communication - K Financial and insurance activities - L Real estate activities - M Professional, scientific, and technical activities - N Administrative and support service activities - O Public administration and defense; compulsory social security - P Education - Q Human health and social work activities - R Arts, entertainment, and recreation - S Other service activities - T Activities of households as employers; undifferentiated goods- and services-producing activities of households for own use - U Activities of extraterritorial organizations and bodies *International Monetary Fund, Annex I to the technical annex of the 2022 G-20 Background Note on Minimizing Scarring from the Pandemic.* ### 17. Columns 4 to 6 in Table A III.3 reports ### 17. Columns 4 to 6 in Table A III.3 reports ### Specification and comparison - The specification evaluated compares the relationship between employment growth and labor productivity with the measure of sectoral reallocation held at the median level. - The comparison examines economies at the 20th percentile and 80th percentile of job retention policy spending. - The estimated effects are qualitatively similar to those reported in the main text (Figure 17, right panel). ### Figure AIII.1 — Impact of Policies on Reallocation (captioned results) - Sources: Orbis; OECD; IMF staff calculations. - Change in employment growth from a 10 percent increase in labor productivity (percent; policy spending at 20th and 80th percentiles respectively, median sectoral reallocation) - Displayed percentiles: 20th percentile, 80th percentile - Vertical scale values shown: 0.0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8 ### Regression results (selected coefficients and statistics from columns (4) to (6) and related columns) - VARIABLES reported for six regressions: DEmp DEmp DEmp DEmp DEmp DEmp - Constant coefficients: - Column (1): -0.684*** - Column (2): -0.676*** - Column (3): -0.684*** - Column (4): -0.687*** - Column (5): -0.676*** - Column (6): -0.686*** - Standard errors (parentheses) corresponding to each: (0.0279) (0.0291) (0.0277) (0.0275) (0.0284) (0.0275) - lag(labprod) coefficients: - Column (1): 0.0777*** - Column (2): 0.0630*** - Column (3): 0.0715*** - Column (4): 0.0724*** - Column (5): 0.0640*** - Column (6): 0.0711*** - Standard errors: (0.00498) (0.00357) (0.00371) (0.00377) (0.00266) (0.00320) - Interaction terms: - lag(labprod)*lag(reten): - Coefficients shown: -0.0300*** (column not explicitly numbered in snippet) and -0.0236** - Standard errors: (0.0113) (0.0113) - lag(labprod)*lag(realloc): - Coefficients shown: -0.00932 and 0.00434 - Standard errors: (0.00739) (0.00790) - lag(labprod)*lag(reten)*(emp realloc US): - Coefficients shown: -0.0966*** and -0.0932*** - Standard errors: (0.0277) (0.0270) - lag(labprod)*lag(realloc)*(emp realloc US): - Coefficients shown: -0.00807 and 0.00884 - Standard errors: (0.0135) (0.0103) ### Sample size, fit, and fixed effects - N: 8,488,366; 10,066,956; 8,553,921; 8,579,656; 10,042,533; 8,652,979 (reported for columns (1) through (6) respectively) - R-squared: 0.054 0.054 0.054 0.054 0.054 0.054 - Country-sector FE: YES for all columns - Country-year FE: YES for all columns - Clustered standard errors (country-sector) in parentheses - Significance notation: *** p<0.01, ** p<0.05, * p<0.1 ### Key interpretation points (as presented) - The interaction of lagged labor productivity with measures of job retention (lag(reten)) and sectoral reallocation is estimated to influence employment growth responses to a 10 percent productivity shock, with the magnitude and significance varying by specification and inclusion of an "emp realloc US" term. - Effects evaluated at the 20th and 80th percentiles of job retention policy spending indicate heterogeneous impacts on employment growth given median sectoral reallocation. --- _Source: https://www.imf.org/-/media/files/research/imf-and-g20/2022/g20-minimizing-scarring-from-the-pandemic-technical-annex.pdf_