## 5. Relative Contribution of Each Causal Variable to the Size of the Informal

## Source details

**Canonical URL:** [5. Relative Contribution of Each Causal Variable to the Size of the Informal](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08102.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08102.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08102.pdf.json)

---

### Introduction and objectives
- Purpose: estimate the size of the informal economy and the relative contribution of each underlying factor in 32 mainly Latin American and Caribbean countries in the early 2000s.
- Definition adopted (following Feige (2005)):
  - illegal economy: income from activities in violation of legal statutes.
  - unreported economy: legal and illegal activities that evade fiscal rules.
  - informal economy: activities that circumvent costs and are excluded from benefits/rights under laws and administrative rules; summary measure is income generated by agents who operate informally.
- Motivations:
  - Examples linking informal economy and tax evasion:
    - Bolivia—informal economy ≈ 65 percent of GDP and VAT tax evasion ≈ 45 percent of GDP.
    - New Zealand—informal activity ≈ 12 percent and tax evasion ≈ 5 percent of GDP.
  - Informal economy affects social security coverage:
    - Netherlands—94 percent of labor force contributing.
    - Honduras—about 19 percent contributing.
- Methodological choice: MIMIC (Multiple Indicators, Multiple Causes) structural equation approach treating informal economy as a latent variable with multiple causes and indicators; cross-section approach across countries in early 2000s.

### Methods: alternatives and MIMIC approach
- Reviewed alternative measurement methods and key limitations:
  - Surveys: detailed but sensitive to questionnaire and response bias.
  - National income/expenditure discrepancy: vulnerable to omissions/errors.
  - Labor force discrepancy: weak indicator due to multiple explanations for participation changes.
  - Electricity approach: sectoral energy differences and variable electricity-GDP elasticity.
  - Transaction approach (Feige): assumption of constant k and payment-technology effects.
  - Currency demand approach (Tanzi/Cagan): assumes informal transactions use cash; unequal velocity; non-cash informal transactions; base-year assumptions.
- MIMIC model structure and estimation:
  - Indicators equation: y = λ IE + ε
  - Causes equation: IE = γ' x + υ
  - Reduced form: y = π x + μ, estimated by maximum likelihood with normalizations.
- Advantages for this study:
  - Cross-section MIMIC requires absolute-information for only one country to convert ordinal predictions into absolute values.
  - Enables use of real variables (avoids monetary series issues in highly dollarized economies).

### Data and variables
- Sample (32 countries, mostly early 2000s): Antigua and Barbuda, Argentina, Barbados, Belize, Brazil, Chile, Colombia, Costa Rica, Cyprus, Dominica, Dominican Republic, Ecuador, El Salvador, Fiji, Grenada, Guatemala, Guyana, Honduras, Jamaica, Malta, Mexico, Nicaragua, Panama, Paraguay, Peru, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, The Bahamas, Trinidad and Tobago, Uruguay, Venezuela.
- Cause variables (expected sign with IE):
  - Tax burden: average of corporate and personal marginal income tax rate (higher → larger IE).
  - Labor rigidities: two indices
    - Labor rigidity index #1 = ratio of annual minimum wage to GDP per capita.
    - Labor rigidity index #2 = average of normalized minimum wage ratio and mandated benefits (social security contribution rates as percent of wages), divided by GDP per capita.
  - Importance of agriculture: agricultural and food exports as percent of total exports (higher → larger IE).
  - Inflation rate (higher → larger IE).
  - Strength of enforcement system: average of three ICRG indicators—quality of bureaucracy, corruption in government, rule of law (stronger enforcement → lower IE). (Excluded from benchmark Model 1; included in alternative specifications.)
- Indicator variables (expected sign with IE):
  - Percentage of labor force contributing to social security (larger IE → lower contributors).
  - Degree of unionization (percentage of labor force in unions) (larger IE → lower unionization).
  - Gross enrollment ratio for secondary school (larger IE → lower enrollment).

### Empirical results — MIMIC estimation (benchmark and alternatives)
- Benchmark specification (Model 1):
  - Causes: labor rigidity index #1, tax burden, importance of agriculture, inflation.
  - Indicators: contributors to social security, degree of unionization, secondary gross enrollment ratio.
  - Standardized causal coefficients (effect on IE in standard-deviation units from a one standard-deviation increase):
    - Tax burden → IE: 0.27
    - Labor rigidities → IE: 0.52
    - Importance of agriculture → IE: 0.40
    - Inflation → IE: 0.47
  - Joint influence of the four causal variables explains approximately 79 percent of the variance of the informal economy.
  - Indicator variance explained by IE:
    - Contributors to social security: 76 percent of variance explained.
    - Degree of unionization: 35 percent of variance explained.
    - Secondary enrollment ratio: 57 percent of variance explained.
  - Statistical significance: coefficients have expected signs and are mostly statistically significant at the 1 or 5 percent level.
- Model-specific MIMIC results (standardized regression coefficients and fit):
  - Model 1 (t-values in parentheses):
    - Labor rigidity index #1: 0.519 (2.81)
    - Tax burden: 0.274 (2.02)
    - Importance of agriculture: 0.404 (2.41)
    - Inflation: 0.465 (2.73)
    - Workers contributing to social security: -0.874 (-3.51)
    - Degree of unionization: -0.593 (no t-test shown)
    - Gross enrollment ratio for secondary school: -0.756 (-3.24)
    - Share of variance explained by the informal economy: 76.3%
    - Informal economy’s share of variance explained by its causes: 78.6%
    - Model fit: CMIN (p-value) : 0.932; GFI : 0.973; AGFI : 0.907; RMSEA : 0
  - Model 2:
    - Labor rigidity index #2: 0.396 (2.09)
    - Tax burden: 0.242 (1.70)
    - Importance of agriculture: 0.412 (2.15)
    - Inflation: 0.450 (2.42)
    - Workers contributing to social security: -0.894 (-3.49)
    - Degree of unionization: -0.596 (no t-test shown)
    - Gross enrollment ratio for secondary school: -0.733 (-3.16)
    - Share of variance explained by the informal economy: 80.0%
    - Informal economy’s share of variance explained by its causes: 67.5%
    - Model fit: CMIN (p-value) : 0.951; GFI : 0.977; AGFI : 0.918; RMSEA : 0
  - Model 3 (includes strength of enforcement):
    - Labor rigidity index #1: 0.296 (2.09)
    - Tax burden: 0.163 (1.51)
    - Importance of agriculture: 0.226 (1.71)
    - Inflation: 0.203 (1.53)
    - Workers contributing to social security: -0.914 (-3.61)
    - Degree of unionization: -0.581 (no t-test shown)
    - Gross enrollment ratio for secondary school: -0.721 (-3.15)
    - Strength of enforcement system: -0.534 (-2.74)
    - Share of variance explained by the informal economy: 83.6%
    - Informal economy’s share of variance explained by its causes: 87.6%
    - Model fit: CMIN (p-value) : 0.914; GFI : 0.965; AGFI : 0.885; RMSEA : 0
- Robustness:
  - Estimated absolute sizes are similar across models:
    - Spearman rank correlation coefficients: Model 1 vs 2 = 0.89; Model 1 vs 3 = 0.98; Model 2 vs 3 = 0.85.
    - Null of independence of estimated absolute sizes across models rejected at the 1 percent level for all comparisons.
  - Model 3 shows strength of enforcement is an important determinant of IE.

### Estimated size of the informal economy (ordinal and absolute)
- Ordinal estimates from benchmark MIMIC converted to absolute values by normalizing Jamaica’s informal economy = 35.0 percent of total GDP (De La Roca et al. (2002) for Jamaica in 2000–01).
- Selected country absolute values (Model 1, early 2000s, percent of measured GDP):
  - The Bahamas: 15.9
  - Cyprus: 19.3
  - Grenada: 22.5
  - St. Kitts and Nevis: 24.2
  - Trinidad and Tobago: 24.4
  - Barbados: 24.5
  - Mexico: 28.2
  - Brazil: 28.4
  - Malta: 28.7
  - Antigua and Barbuda: 31.2
  - Chile: 32.1
  - Argentina: 32.9
  - Dominica: 34.2
  - Jamaica: 35.0
  - Uruguay: 36.2
  - El Salvador: 36.4
  - Guyana: 36.7
  - Peru: 38.1
  - St. Lucia: 41.5
  - Costa Rica: 41.8
  - Guatemala: 42.3
  - Venezuela: 43.0
  - Colombia: 43.5
  - Panama: 44.4
  - Dominican Republic: 44.8
  - Belize: 46.8
  - St. Vincent and the Grenadines: 50.6
  - Ecuador: 50.7
  - Honduras: 54.1
  - Fiji: 60.1
  - Nicaragua: 64.4
  - Paraguay: 68.2
- Sample statistics (Table 3):
  - Mean (absolute value): 38.3 (% of GDP)
  - Standard deviation (absolute value): 12.7
- Cross-model comparisons for Caribbean countries (absolute % of GDP, Table 4 examples):
  - The Bahamas: Model 1 = 15.9; Model 2 = 11.5; Model 3 = 15.1
  - Grenada: Model 1 = 22.5; Model 2 = 31.8; Model 3 = 22.9
  - Jamaica: Model 1 = 35.0; Model 2 = 35.0; Model 3 = 35.0
  - Guyana: Model 1 = 36.7; Model 2 = 57.3; Model 3 = 37.3
  - Belize: Model 1 = 46.8; Model 2 = 56.5; Model 3 = 47.4
  - St. Vincent and the Grenadines: Model 1 = 50.6; Model 2 = 58.4; Model 3 = 51.4

### Relative contribution of each cause variable (cross-country patterns)
- Sample-average contributions to the overall size of the informal economy (Model 1, percent of total contribution):
  - Tax burden: 34.6
  - Labor rigidity index #1: 26.4
  - Importance of agriculture: 30.8
  - Inflation: 8.2
- Country heterogeneity — illustrative country-level shares (Tax burden / Labor rigidity index #1 / Importance of agriculture / Inflation) from Table 5:
  - The Bahamas: 0.0 / 54.6 / 42.3 / 3.1
  - Cyprus: 32.2 / 0.0 / 63.5 / 4.3
  - Grenada: 57.1 / 0.0 / 40.9 / 2.0
  - St. Kitts and Nevis: 34.0 / 32.4 / 28.1 / 5.5
  - Trinidad and Tobago: 61.4 / 26.5 / 6.5 / 5.6
  - Barbados: 65.6 / 0.0 / 31.2 / 3.2
  - Mexico: 52.4 / 14.4 / 5.4 / 27.8
  - Brazil: 31.1 / 19.6 / 27.5 / 21.8
  - Malta: 52.2 / 42.1 / 2.6 / 3.1
  - Antigua and Barbuda: 60.5 / 31.3 / 6.1 / 2.1
  - Chile: 36.1 / 27.6 / 30.2 / 6.0
  - Argentina: 45.6 / 15.3 / 38.3 / 0.7
  - Dominica: 43.2 / 24.7 / 30.7 / 1.4
  - Jamaica: 36.2 / 33.3 / 17.6 / 12.9
  - Uruguay: 22.8 / 15.4 / 43.0 / 18.9
  - El Salvador: 32.1 / 30.3 / 32.8 / 4.8
  - Guyana: 46.3 / 0.0 / 47.6 / 6.1
  - Peru: 31.9 / 36.7 / 24.4 / 7.0
  - St. Lucia: 32.9 / 16.4 / 48.7 / 2.0
  - Costa Rica: 30.8 / 35.6 / 22.0 / 11.6
  - Guatemala: 31.4 / 23.0 / 39.5 / 6.1
  - Venezuela: 33.9 / 24.9 / 1.1 / 40.1
  - Colombia: 36.4 / 35.3 / 15.2 / 13.1
  - Panama: 29.0 / 23.1 / 47.1 / 0.8
  - Dominican Republic: 23.9 / 44.4 / 26.3 / 5.4
  - Belize: 22.9 / 26.7 / 49.3 / 1.1
  - St. Vincent and the Grenadines: 33.8 / 23.9 / 41.0 / 1.2
  - Ecuador: 21.1 / 35.7 / 22.2 / 21.0
  - Honduras: 19.8 / 31.2 / 37.4 / 11.7
  - Fiji: 22.8 / 29.6 / 45.8 / 1.7
  - Nicaragua: 18.5 / 37.1 / 38.9 / 5.6
  - Paraguay: 10.4 / 52.4 / 32.7 / 4.5

### Impacts of informality on social indicators and correlations
- Informality reduces:
  - Labor unionization,
  - Number of contributors to social security schemes,
  - Enrollment rates in education.
- Correlations between cause and indicator variables (Table 2):
  - Tax burden vs (Workers contributing to social security / Gross enrollment ratio for secondary school / Degree of unionization): -0.14 / -0.12 / 0.07
  - Labor rigidity index #1: -0.59 / -0.60 / -0.39
  - Labor rigidity index #2: -0.59 / -0.53 / -0.36
  - Importance of agriculture: -0.39 / -0.32 / -0.31
  - Inflation: -0.40 / -0.29 / -0.30
  - Strength of enforcement system: 0.82 / 0.58 / 0.49

### Conclusions and policy implications
- Key determinants jointly explain around 79 percent of the variance in the informal economy: burdensome tax system, rigid labor markets, higher inflation, and dominance of agriculture.
- Cross-country variation is large: informal sector shares range from below 25 percent of GDP in several Caribbean economies to around 70 percent in some Latin American countries (Paraguay, Nicaragua).
- Policy recommendations (tailored to dominant cause in each country):
  - Where high tax burden drives informality: consider lowering and homogenizing effective tax rates across sectors to broaden the tax base and enhance compliance.
  - Where labor market rigidities drive informality: accelerate labor market reforms to enhance flexibility.
  - Where inflation drives informality: priority to tightening monetary policy and stabilizing prices.
  - Reducing the size of the informal economy should assist in bolstering fiscal and debt sustainability given high debt burdens in many Latin American and Caribbean countries.

*Source: _wp08102 - 5. Relative Contribution of Each Causal Variable to the Size of the Informal (IMF working paper content provided).*

### References...................................................................................................     16

### _wp08102 - References...................................................................................................     16

### Appendix
- Data Construction and Sources .................................................................... 20

### Figures
- 1. MIMIC Estimation Results, Model 1 ......................................................... 22
- 2. MIMIC Estimation Results, Model 2 ......................................................... 23
- 3. MIMIC Estimation Results, Model 3 ......................................................... 24
- 4. Estimated Size of the Informal Economy.................................................... 25
- 5. Caribbean: Contribution of Each Cause Variable to the Size of the Informal 
  Economy.................................................................................................. 25

### Tables
- 1. Size of the Informal Economy and VAT Tax Evasion .......................................    26
- 2. Correlations Between Cause and Indicator Variables .........................................    26
- 3. Estimated Size of Informal Economy: Standardized and Absolute Values................     27
- 4. Caribbean: Estimated Absolute Size of the Informal Economy Under Alternative 
  Model Specifications....................................................................................     28

*Source: _wp08102 - References...................................................................................................     16*

### 5. Relative Contribution of Each Causal Variable to the Size of the Informal

### 5. Relative Contribution of Each Causal Variable to the Size of the Informal

### Introduction and objectives
- Purpose: estimate the size of the informal economy and the relative contribution of each underlying factor in 32 mainly Latin American and Caribbean countries in the early 2000s.
- Definition adopted (following Feige (2005)):
  - illegal economy: income from activities in violation of legal statutes.
  - unreported economy: legal and illegal activities that evade fiscal rules.
  - informal economy: activities that circumvent costs and are excluded from benefits/rights under laws and administrative rules; summary measure is income generated by agents who operate informally.
- Motivations:
  - Link between informal economy and tax evasion (examples): Bolivia—informal economy ≈ 65 percent of GDP and VAT tax evasion ≈ 45 percent of GDP; New Zealand—informal activity ≈ 12 percent and tax evasion ≈ 5 percent of GDP.
  - Informal economy affects social security coverage (examples): Netherlands—94 percent of labor force contributing; Honduras—about 19 percent contributing.
- Methodological choice: MIMIC (Multiple Indicators, Multiple Causes) structural equation approach treating informal economy as a latent variable with multiple causes and indicators; cross-section approach across countries in early 2000s.

### Methods: alternatives and MIMIC approach
- Alternative measurement methods reviewed:
  - Surveys (micro approaches): detailed but sensitive to questionnaire and response bias.
  - Discrepancy between national expenditure and income statistics: vulnerable to omissions/errors.
  - Discrepancy between official and actual labor force: weak indicator due to multiple explanations for participation changes.
  - Electricity approach (Kaufmann and Kaliberda): uses difference between electricity growth and official GDP growth; limitations include sectoral energy differences and variable electricity-GDP elasticity.
  - Transaction approach (Feige): uses Money*Velocity = k(official GDP + informal economy); weaknesses include assumption of constant k and payment-technology effects.
  - Currency demand approach (Tanzi/Cagan): assumes informal transactions use cash; problems include unequal velocity, non-cash informal transactions, and base-year assumptions.
- MIMIC approach rationale:
  - Uses multiple observable causes and multiple observable indicators (symptoms) to estimate unobserved latent IE (informal economy).
  - Model structure:
    - y = λ IE + ε (indicators equation)
    - IE = γ' x + υ (causes equation)
    - Reduced form: y = π x + μ, estimated by maximum likelihood with appropriate normalizations.
  - Advantages for this study:
    - Cross-section MIMIC requires absolute-information for only one country (benchmark) to convert ordinal predictions into absolute values.
    - Enables use of real variables (avoids monetary series issues in highly dollarized economies).

### Data and variables
- Sample: 32 countries (mainly Latin American and Caribbean) in the early 2000s: Antigua and Barbuda, Argentina, Barbados, Belize, Brazil, Chile, Colombia, Costa Rica, Cyprus, Dominica, Dominican Republic, Ecuador, El Salvador, Fiji, Grenada, Guatemala, Guyana, Honduras, Jamaica, Malta, Mexico, Nicaragua, Panama, Paraguay, Peru, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, The Bahamas, Trinidad and Tobago, Uruguay, Venezuela.
- Note: Most data based on 2002 or 2003 information.
- Cause variables (expected sign with IE):
  - Tax burden: average of corporate and personal marginal income tax rate (higher → larger IE).
  - Labor rigidities: two indices
    - Labor rigidity index #1 = ratio of annual minimum wage to GDP per capita.
    - Labor rigidity index #2 = average of normalized minimum wage ratio and mandated benefits (social security contribution rates as percent of wages), divided by GDP per capita.
  - Importance of agriculture: agriculture and food exports as percent of total exports (higher → larger IE).
  - Inflation rate (higher → larger IE).
  - Strength of enforcement system: average of three ICRG indicators—quality of bureaucracy, corruption in government, rule of law (stronger enforcement → lower IE). (Excluded from benchmark Model 1 due to endogeneity concerns; included in alternative specifications.)
- Indicator variables (expected sign with IE):
  - Percentage of labor force contributing to social security (larger IE → lower contributors).
  - Degree of unionization (percentage of labor force in unions) (larger IE → lower unionization).
  - Gross enrollment ratio for secondary school (larger IE → lower enrollment).

### Empirical results — MIMIC estimation
- Benchmark specification (Model 1):
  - Cause variables: labor rigidity index #1, tax burden, importance of agriculture, inflation.
  - Indicator variables: contributors to social security, degree of unionization, secondary gross enrollment ratio.
  - Standardized causal coefficients (effect on IE in standard-deviation units from a one standard-deviation increase):
    - Tax burden → IE: 0.27
    - Labor rigidities → IE: 0.52
    - Importance of agriculture → IE: 0.40
    - Inflation → IE: 0.47
  - Joint influence of the four causal variables explains approximately 79 percent of the variance of the informal economy.
  - Indicator variance explained by IE:
    - Contributors to social security: 76 percent of variance explained.
    - Degree of unionization: 35 percent of variance explained.
    - Secondary enrollment ratio: 57 percent of variance explained.
  - Coefficients on causal and indicator variables have expected signs and are mostly statistically significant at the 1 or 5 percent level.
- Robustness / alternative specifications:
  - Models 2 and 3 include alternative labor rigidity measure and strength of enforcement system.
  - Results consistent with benchmark; Model 3 shows strength of enforcement is an important determinant of IE.
  - Estimated absolute sizes are similar across models:
    - Spearman rank correlation coefficients: Model 1 vs 2 = 0.89, Model 1 vs 3 = 0.98, Model 2 vs 3 = 0.85.
    - Null of independence of estimated absolute sizes across models rejected at the 1 percent level for all comparisons.
- Comparison with other studies:
  - For 15 common countries, correlation with Schneider (2002) estimates for the late 1990s: Pearson = 0.37; Spearman rank test = 0.44 (rejects zero correlation at the 10 percent level).

### Estimated size of the informal economy (ordinal and absolute)
- Ordinal estimates produced by the benchmark MIMIC model.
- Conversion to absolute values: normalized by setting Jamaica’s informal economy = 35 percent of total GDP (based on De La Roca et al. (2002) study for Jamaica in 2000–01).
- Selected country ranges (absolute values, early 2000s, as percent of measured GDP):
  - The Bahamas, Cyprus, Grenada, St. Kitts and Nevis, Trinidad and Tobago, Barbados: approximately 15 to 25 percent of GDP (among the smallest in sample).
  - St. Vincent and the Grenadines, Belize, Dominican Republic: approximately 45 to 51 percent of GDP.
  - Paraguay and Nicaragua: around 65 to 70 percent of GDP (among the largest in sample).
- Caution: absolute values depend on the Jamaica benchmark; ordinal ranking is independent of that extra information.

### Relative contribution of each cause variable (cross-country patterns)
- Sample-average contributions to the overall size of the informal economy:
  - Tax burden: around 35 percent.
  - Labor rigidity: around 26 percent.
  - Importance of agriculture: around 31 percent.
  - Inflation: around 8 percent.
- Heterogeneity across countries — illustrative examples:
  - Antigua and Barbuda, Barbados, Trinidad and Tobago: tax burden is the main component influencing IE (example: Antigua and Barbuda had marginal statutory corporate tax rate of 40 percent in the late 1990s).
  - St. Vincent and the Grenadines, St. Lucia, Belize: importance of agriculture is a decisive factor (about 75 percent of exports concentrated in agriculture and food products in some of these countries).
  - Paraguay and Dominican Republic: labor rigidities are crucial (minimum wages representing 170 percent and 90 percent, respectively, of GDP per capita).
  - Inflation: generally not an important determinant for most economies in the sample (likely due to relative price stability in the second half of the 1990s).

### Conclusions and policy implications
- Key determinants: burdensome tax system, rigid labor markets, higher inflation, and dominance of agriculture together explain around 79 percent of the variance in the informal economy.
- Informality reduces:
  - Labor unionization,
  - Number of contributors to social security schemes,
  - Enrollment rates in education.
- Cross-country variation is large: informal sector shares range from below 25 percent of GDP in several Caribbean economies to around 70 percent in some Latin American countries (Paraguay, Nicaragua).
- Policy implications (tailored to dominant cause in each country):
  - Where high tax burden drives informality: consider lowering and homogenizing effective tax rates across sectors to broaden the tax base and enhance compliance.
  - Where labor market rigidities drive informality: accelerate labor market reforms to enhance flexibility.
  - Where inflation drives informality: priority to tightening monetary policy and stabilizing prices.
  - Reducing the size of the informal economy should assist in bolstering fiscal and debt sustainability given high debt burdens in many Latin American and Caribbean countries.

*Italic: Source — _wp08102 - 5. Relative Contribution of Each Causal Variable to the Size of the Informal (IMF working paper content provided).*

### REFERENCES

### _wp08102 - REFERENCES

### Data construction and sources (Appendix)
- Causal variables and their measurement:
  - Tax burden: proxy = average of corporate and personal marginal income tax rates; highest rate used when multiple rates; normalized between 0 and 100; data correspond mostly for 2003; sources: World Development Indicators 2006, Bain and dos Santos (2004).
  - Labor rigidity indices:
    - Labor rigidity index #1: ratio of minimum wage and GDP per capita normalized between 0 and 100; minimum wage corresponds to most general regime; data correspond to 2002; main source: “Country Reports on Human Rights Practices” (2002).
    - Labor rigidity index #2: normalized average of two components divided by real GDP per capita; components are labor rigidity index #1 and mandated benefits measured by contribution rates (percent of salaries) for all social security programs for employee and employer; Belize exception uses normalized legal number of days of maternity leave with full pay; social security contribution data correspond mostly to year 2003 from “Social Security Programs Throughout the World”; maternity leave information correspond to average of 1999–2002.
  - Importance of agriculture: measured by agricultural raw material and food exports (percent of total exports); World Development Indicators 2006; correspond mainly for 2000 (Dominican Republic 2001; Fiji 2002).
  - Inflation: annual average consumer prices inflation for the period 1995–99; data mainly from World Development Indicators 2006; exception Antigua and Barbuda use IMF data.
  - Strength of enforcement system: average of three ICRG subjective indicators for 2002 — quality of bureaucracy, corruption in government, and rule of law; ICRG publication of Political Risk Services.

- Indicator variables and sources:
  - Workers contributing to social security: active contributors to old-age pension schemes, percent of the labor force; predominantly for 2002; sources: social security agencies, household surveys, IMF desk information.
  - Degree of unionization: total union membership as percent of the labor force; mainly from “Country Reports on Human Rights Practices” (2002) and national authorities.
  - Gross enrollment ratio for secondary school: total secondary enrollment as percent of official school-age population; mostly for 2001; sources: Human Development Report (2005) and World Bank (2005).

### MIMIC estimation results (Models and key coefficients)
- Note on identification: the non-standardized coefficient associated with Degree of unionization was set to -1 in all models; t-test cannot be performed on that coefficient.
- Model 1 (Figure 1, overall model fit):
  - Size of the informal economy — standardized regression coefficients (t-values in parentheses):
    - Labor rigidity index #1: 0.519 (2.81)
    - Tax burden: 0.274 (2.02)
    - Importance of agriculture: 0.404 (2.41)
    - Inflation: 0.465 (2.73)
    - Workers contributing to social security: -0.874 (-3.51)
    - Degree of unionization: -0.593 (no t-test shown)
    - Gross enrollment ratio for secondary school: -0.756 (-3.24)
  - Share of variance explained by the informal economy: 76.3%
  - Informal economy’s share of variance explained by its causes: 78.6%
  - Overall model fit:
    - Discrepancy function (CMIN) (p-value) : 0.932
    - Goodness of fit index (GFI) : 0.973
    - Adjusted goodness of fit index (AGFI) : 0.907
    - Root mean square error of approximation (RMSEA) : 0

- Model 2 (Figure 2, overall model fit):
  - Size of the informal economy — standardized regression coefficients (t-values):
    - Labor rigidity index #2: 0.396 (2.09)
    - Tax burden: 0.242 (1.70)
    - Importance of agriculture: 0.412 (2.15)
    - Inflation: 0.450 (2.42)
    - Workers contributing to social security: -0.894 (-3.49)
    - Degree of unionization: -0.596 (no t-test shown)
    - Gross enrollment ratio for secondary school: -0.733 (-3.16)
  - Share of variance explained by the informal economy: 80.0%
  - Informal economy’s share of variance explained by its causes: 67.5%
  - Overall model fit:
    - Discrepancy function (CMIN) (p-value) : 0.951
    - Goodness of fit index (GFI) : 0.977
    - Adjusted goodness of fit index (AGFI) : 0.918
    - RMSEA : 0

- Model 3 (Figure 3, overall model fit):
  - Size of the informal economy — standardized regression coefficients (t-values):
    - Labor rigidity index #1: 0.296 (2.09)
    - Tax burden: 0.163 (1.51)
    - Importance of agriculture: 0.226 (1.71)
    - Inflation: 0.203 (1.53)
    - Workers contributing to social security: -0.914 (-3.61)
    - Degree of unionization: -0.581 (no t-test shown)
    - Gross enrollment ratio for secondary school: -0.721 (-3.15)
    - Strength of enforcement system: -0.534 (-2.74) (included in Model 3)
  - Share of variance explained by the informal economy: 83.6%
  - Informal economy’s share of variance explained by its causes: 87.6%
  - Overall model fit:
    - Discrepancy function (CMIN) (p-value) : 0.914
    - Goodness of fit index (GFI) : 0.965
    - Adjusted goodness of fit index (AGFI) : 0.885
    - RMSEA : 0

### Estimated sizes of the informal economy (early 2000s) — country estimates and summary statistics
- Selected absolute values (% of GDP) from Table 3 (Model 1 results; country, absolute value):
  - The Bahamas: 15.9
  - Cyprus: 19.3
  - Grenada: 22.5
  - St. Kitts and Nevis: 24.2
  - Trinidad and Tobago: 24.4
  - Barbados: 24.5
  - Mexico: 28.2
  - Brazil: 28.4
  - Malta: 28.7
  - Antigua and Barbuda: 31.2
  - Chile: 32.1
  - Argentina: 32.9
  - Dominica: 34.2
  - Jamaica: 35.0
  - Uruguay: 36.2
  - El Salvador: 36.4
  - Guyana: 36.7
  - Peru: 38.1
  - St. Lucia: 41.5
  - Costa Rica: 41.8
  - Guatemala: 42.3
  - Venezuela: 43.0
  - Colombia: 43.5
  - Panama: 44.4
  - Dominican Republic: 44.8
  - Belize: 46.8
  - St. Vincent and the Grenadines: 50.6
  - Ecuador: 50.7
  - Honduras: 54.1
  - Fiji: 60.1
  - Nicaragua: 64.4
  - Paraguay: 68.2
- Sample statistics (Table 3):
  - Mean (standardized value mean): 0.000
  - Mean (absolute value): 38.3 (% of GDP)
  - Standard deviation (standardized): 1.000
  - Standard deviation (absolute value): 12.7

- Cross-model comparisons for Caribbean countries (Table 4; absolute % of GDP):
  - The Bahamas: MIMIC Model 1 = 15.9; Model 2 = 11.5; Model 3 = 15.1
  - Grenada: Model 1 = 22.5; Model 2 = 31.8; Model 3 = 22.9
  - St. Kitts and Nevis: Model 1 = 24.2; Model 2 = 24.6; Model 3 = 24.4
  - Trinidad and Tobago: Model 1 = 24.4; Model 2 = 25.2; Model 3 = 24.8
  - Barbados: Model 1 = 24.5; Model 2 = 36.6; Model 3 = 24.3
  - Antigua and Barbuda: Model 1 = 31.2; Model 2 = 29.7; Model 3 = 31.7
  - Dominica: Model 1 = 34.2; Model 2 = 38.8; Model 3 = 35.0
  - Jamaica: Model 1 = 35.0; Model 2 = 35.0; Model 3 = 35.0
  - Guyana: Model 1 = 36.7; Model 2 = 57.3; Model 3 = 37.3
  - St. Lucia: Model 1 = 41.5; Model 2 = 52.0; Model 3 = 41.8
  - Dominican Republic: Model 1 = 44.8; Model 2 = 46.1; Model 3 = 45.3
  - Belize: Model 1 = 46.8; Model 2 = 56.5; Model 3 = 47.4
  - St. Vincent and the Grenadines: Model 1 = 50.6; Model 2 = 58.4; Model 3 = 51.4

### Contributions of causes to the informal economy (country-level shares)
- Figure 5 (Caribbean contributions; Model 1 base): displays percent contributions by Tax burden, Labor rigidity #1, Importance of agriculture, Inflation for Caribbean countries; note: only contributions higher than 7 percent display the associated number on figure.
- Table 5 (Relative contribution of each causal variable; Model 1; values expressed as percent of total contribution):
  - Sample country-specific breakdowns (Tax burden / Labor rigidity index #1 / Importance of agriculture / Inflation):
    - The Bahamas: 0.0 / 54.6 / 42.3 / 3.1
    - Cyprus: 32.2 / 0.0 / 63.5 / 4.3
    - Grenada: 57.1 / 0.0 / 40.9 / 2.0
    - St. Kitts and Nevis: 34.0 / 32.4 / 28.1 / 5.5
    - Trinidad and Tobago: 61.4 / 26.5 / 6.5 / 5.6
    - Barbados: 65.6 / 0.0 / 31.2 / 3.2
    - Mexico: 52.4 / 14.4 / 5.4 / 27.8
    - Brazil: 31.1 / 19.6 / 27.5 / 21.8
    - Malta: 52.2 / 42.1 / 2.6 / 3.1
    - Antigua and Barbuda: 60.5 / 31.3 / 6.1 / 2.1
    - Chile: 36.1 / 27.6 / 30.2 / 6.0
    - Argentina: 45.6 / 15.3 / 38.3 / 0.7
    - Dominica: 43.2 / 24.7 / 30.7 / 1.4
    - Jamaica: 36.2 / 33.3 / 17.6 / 12.9
    - Uruguay: 22.8 / 15.4 / 43.0 / 18.9
    - El Salvador: 32.1 / 30.3 / 32.8 / 4.8
    - Guyana: 46.3 / 0.0 / 47.6 / 6.1
    - Peru: 31.9 / 36.7 / 24.4 / 7.0
    - St. Lucia: 32.9 / 16.4 / 48.7 / 2.0
    - Costa Rica: 30.8 / 35.6 / 22.0 / 11.6
    - Guatemala: 31.4 / 23.0 / 39.5 / 6.1
    - Venezuela: 33.9 / 24.9 / 1.1 / 40.1
    - Colombia: 36.4 / 35.3 / 15.2 / 13.1
    - Panama: 29.0 / 23.1 / 47.1 / 0.8
    - Dominican Republic: 23.9 / 44.4 / 26.3 / 5.4
    - Belize: 22.9 / 26.7 / 49.3 / 1.1
    - St. Vincent and the Grenadines: 33.8 / 23.9 / 41.0 / 1.2
    - Ecuador: 21.1 / 35.7 / 22.2 / 21.0
    - Honduras: 19.8 / 31.2 / 37.4 / 11.7
    - Fiji: 22.8 / 29.6 / 45.8 / 1.7
    - Nicaragua: 18.5 / 37.1 / 38.9 / 5.6
    - Paraguay: 10.4 / 52.4 / 32.7 / 4.5
  - Mean contribution across sample (Table 5):
    - Tax burden: 34.6
    - Labor rigidity index #1: 26.4
    - Importance of agriculture: 30.8
    - Inflation: 8.2

### Additional relevant numeric comparisons
- Table 1 (Size of the Informal Economy and VAT Tax Evasion, early 1990s; percent of GDP):
  - New Zealand: Informal economy 12%; VAT tax evasion 5%
  - Sweden: Informal economy 16%; VAT tax evasion 6%
  - Argentina: Informal economy 21%; VAT tax evasion 30%
  - Honduras: Informal economy 47%; VAT tax evasion 35%
  - Bolivia: Informal economy 66%; VAT tax evasion 44%
  - Sources cited in table: Schneider and Enste (2000), and Silvani and Brondolo (1993).

### Correlations between cause and indicator variables (Table 2)
- Correlation matrix (cause variables vs. indicator variables: Workers contributing to social security / Gross enrollment ratio for secondary school / Degree of unionization):
  - Tax burden: -0.14 / -0.12 / 0.07
  - Labor rigidity index #1: -0.59 / -0.60 / -0.39
  - Labor rigidity index #2: -0.59 / -0.53 / -0.36
  - Importance of agriculture: -0.39 / -0.32 / -0.31
  - Inflation: -0.40 / -0.29 / -0.30
  - Strength of enforcement system: 0.82 / 0.58 / 0.49

*Source: _wp08102 - REFERENCES (Appendix, figures, and tables) — Author’s calculations as presented in the source content.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08102.pdf_