## WP/19/209 — SME Financial Inclusion for Sustained Growth in the Middle East and Central Asia (Sections 1–3)

## Source details

**Canonical URL:** [WP/19/209 — SME Financial Inclusion for Sustained Growth in the Middle East and Central Asia (Sections 1–3)](https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019209-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2019/wpiea2019209-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2019/wpiea2019209-print-pdf.pdf.json)

---

### Abstract and introduction; scope and key summaries
- Authors: Mishel Ghassibe, Maximiliano Appendino, Samir Elsadek Mahmoudi.
- Main claim: Greater financial inclusion of small and medium enterprises (SMEs) can promote higher economic growth and employment, especially in the Middle East and Central Asia.
- Key empirical points:
  - Countries with higher SME financial inclusion exhibit more effective monetary policy transmission and tax collection.
  - Firm-level estimates show employment and labor productivity growth gains from access to credit, with gains higher for SMEs than for large firms.
  - Improved credit bureau coverage and insolvency regimes have a substantial positive impact on SME employment and labor productivity growth.
  - Cross-country aggregate evidence confirms employment and growth gains from SME financial inclusion; gains appear larger in the Middle East and Central Asia than in other regions.
- Regional SME prevalence (IMF, 2018c):
  - SMEs represent about 96 per cent of all registered companies in the region.
  - SMEs employ 48 per cent of the labor force.
- Paper structure:
  - Section II: SME financial inclusion and effectiveness of monetary and fiscal policy.
  - Section III: firm-level evidence and macroeconomic impact.
  - Section IV: cross-country aggregate confirmation.
  - Section V: conclusions and suggestions for future work.

### Measurement, data, and classification
- SME financial inclusion index:
  - Measured using the IMF (2018b) index; first principal component of World Bank Enterprise Surveys indicators (higher index values correspond to higher SME financial inclusion).
  - Correlations (world sample): -0.53 and -0.52 with two World Bank measures of credit constraints and rejected loan applications, respectively.
- Country baseline sample (data availability): Jordan, Lebanon, Egypt, Iraq, Morocco, Tunisia, Yemen, Armenia, Azerbaijan, Georgia, Kazakhstan, Kyrgyzstan, Tajikistan, Uzbekistan, Djibouti.
- High vs. Low SME financial inclusion:
  - High (indices above the mean): Morocco, Tunisia, Armenia, Georgia.
  - Low (below the mean): Jordan, Egypt, Iraq, Yemen, Azerbaijan, Kazakhstan, Kyrgyzstan, Tajikistan, Uzbekistan.

### II. Macroeconomic policies and SME financial inclusion — channels
- Monetary policy channel:
  - Greater access to formal borrowing and lending increases the role of the interest rate and strengthens monetary policy transmission; better linkage between real and nominal sides may reduce inflation volatility.
- Fiscal policy channel:
  - Greater SME financial inclusion can boost formalization and improve tax collection efficiency.
- Financial stability channel:
  - Under appropriate risk management and supervision, broader SME lending can diversify banks’ credit portfolios and potentially improve stability.

### II.A Monetary policy transmission — methods and quantitative results
- Data and empirical strategy:
  - Annual time series (1990–2017) on real GDP (output gap via HP filter), inflation, short-term nominal interest rate.
  - Panel VAR: Y_it = A Y_{i,t−1} + u_i + v_t + e_it, with Y_it = {output gap, inflation, nominal interest rate}.
  - Strength of monetary transmission measured by impulse response of output gap to a structural shock to the nominal interest rate; comparison across High vs. Low SME financial inclusion subsamples.
  - Price stability proxied by the ratio of unconditional variances of output and inflation.
- Key numerical findings:
  - Full sample: a 100bp structural increase in the nominal interest rate leads to a peak of 0.13 percentage point (pp) drop in the output gap (not statistically significant at 90 per cent confidence level).
  - High SME financial inclusion subsample: 0.15 pp peak drop in the output gap; statistically significant in the first two years following the shock.
  - Low SME financial inclusion subsample: output gap contraction close to and insignificantly different from zero.
- Interpretation and caveats:
  - Evidence supports stronger monetary policy transmission in higher-SME-inclusion countries in the region.
  - High SME financial inclusion countries have a greater ratio of output and inflation variances, interpreted as central banks better able to minimize price fluctuations (manifested in increased volatility of output gap).
  - Differences may reflect other heterogeneities (exchange rate flexibility, central bank independence); estimated benefits assume extra SME funding comes from traditional banking—alternative lenders (FinTech) may be less sensitive to monetary policy.

### II.B SME financial inclusion and tax collection — methods and quantitative results
- Empirical strategy:
  - Panel VAR: Y_it = A Y_{i,t−1} + u_i + v_t + e_it, with Y_it = {Tax/GDP, output gap, inflation}.
  - Efficiency measured by impulse response of Tax/GDP to a structural positive shock to the output gap; VAR re-estimated for top and bottom quartiles of SME Financial Inclusion index.
- Key numerical findings:
  - Full sample: a 1 pp expansion in the output gap leads to about 0.4 pp increase in tax-to-GDP after one year (estimate significantly different from zero at the 90 percent confidence level for the first two years).
  - Subsample: effect holds in the top-quartile SME financial inclusion subsample; does not hold in the bottom-quartile subsample.
- Interpretation:
  - SME financial inclusion strengthens tax collection and is associated with more effective fiscal policy implementation via improved Tax/GDP responsiveness to output.

### III. Firm-level empirical approach and data
- Data: World Bank Enterprise Surveys (2008-2016) for firms in Azerbaijan, Armenia, Djibouti, Georgia, Iraq, Kazakhstan, Jordan, Kyrgyzstan, Lebanon, Morocco, Tajikistan, Tunisia, Egypt, Uzbekistan, Yemen.
- Firm size classification:
  - Large: more than 100 employees.
  - Medium: between 20 and 100 employees.
  - Small: less than 20 employees.
- Main firm-level specification:
  - ∆E_ijt = α F_ijt + β X_ijt + Z_jt + C_j + Y_t + ε_ijt where ∆E_ijt is annual employment growth or labor productivity growth, F_ijt indicates a formal loan outstanding.
- Policy-proxy specification:
  - ∆E_ijt = α S_jt + β X_ijt + Z_jt + C_j + Y_t + ε_ijt where S_jt is credit bureau coverage (share of covered adults) or insolvency regime index (0–100).

### III. Firm-level effects on employment and labor productivity — key estimates
- Access to formal finance (whole sample):
  - Employment growth: 2.07 pp increase for an average firm.
  - Labor productivity growth: 0.94 pp increase (whole sample).
- Heterogeneity by firm size (access to formal finance):
  - Employment growth gains concentrated in SMEs:
    - All firms: 2.07*** (0.31)
    - Large firms (>100): 0.89* (0.52)
    - SMEs (<100): 2.14*** (0.38)
    - Small (<20): 2.31*** (0.60)
    - Medium (20-100): 1.84*** (0.51)
  - Labor productivity growth:
    - All firms: 0.94 (0.62)
    - Large firms (>100): 0.43 (1.19)
    - SMEs (<100): 1.22* (0.74)
    - Small (<20): -0.26 (1.16)
    - Medium (20-100): 2.39** (0.97)
- Estimation notes:
  - Results exclude firms in top and bottom 1 percent of employment and productivity growth; results for top/bottom halves of SME Financial Inclusion were noisy.

### III. Policy interventions: credit bureaus and insolvency regimes — quantitative impacts
- Credit bureau coverage (percent):
  - All firms: 0.18*** (0.03) increase in employment growth per one pp increase in coverage.
  - Large firms: 0.08 (0.07) (insignificant).
  - SMEs: 0.23*** (0.04).
  - Labor productivity: 0.29*** (0.07) increase per one pp increase in coverage (All firms); subgroup estimates reported and vary with inclusion of survey year fixed effects.
  - With survey year fixed effects estimates are smaller and noisier; SME effects persist though magnitude varies.
- Insolvency regime index (0–100):
  - Employment growth:
    - All firms: 0.23*** (0.03) per unit increase.
    - Large firms: 0.12** (0.05).
    - SMEs: 0.30*** (0.04).
  - Labor productivity growth:
    - All firms: 0.57*** (0.06) per unit increase.
    - Large firms: 0.44*** (0.12).
    - SMEs: 0.60*** (0.07).
  - With survey year fixed effects some alternative specifications report larger or negative point estimates; results are sensitive to specification.
- Cumulative illustrative gains (moving to EMDE third quartile):
  - Credit bureau coverage:
    - SME employment growth would rise by close to 3 pp.
    - SME labor productivity growth would rise by close to 3.5 pp.
  - Insolvency regime index:
    - SME employment growth estimated to rise by 2.5 pp.
    - SME labor productivity growth estimated to rise by 4.5 pp.
  - Caveat: cumulative estimates assume linear effects and may be an upper bound if marginal returns diminish.

### Cross-country analysis: SME share in bank lending and macro effects
- Stylized facts from IMF Financial Access Survey:
  - In Middle East and Central Asia, SMEs have consistently lower shares in total bank lending than other regions.
  - SMEs’ shares in total lending have tended to decrease over time; concurrent rise in share of governments and SOEs in bank lending.
- Dynamic panel specification:
  - U_it = δ U_it−1 + β1 SME_Share_it + β2 FD_it + β3 SME_Share_it × FD_it + θ X_it + γ_t + α_i + ε_it with Arellano and Bond (1991) GMM estimator.
  - Variables: U_it (unemployment or GDP per capita growth), FD_it (ln commercial bank loans to resident nonfinancial corporations and households relative to GDP), SME_Share_it (ln share of SMEs in total bank lending).
- Key macro results (selected coefficients from Table 7):
  - Unemployment:
    - U_it−1: 0.210** (0.100) in column (1); 0.278*** (0.0897) in column (2).
    - Share of SMEs in total bank lending: -0.664*** (0.245) in column (1); -0.692* (0.400) in column (2).
    - Interaction (SME Share × Total lending): 0.104* (0.0610) in column (1); -0.0848** (0.0389) in column (2).
  - GDP per capita growth:
    - U_it−1: 0.551*** (0.0664) in column (3).
    - Share of SMEs in total bank lending: 0.426*** (0.159) in column (3).
    - Loans with commercial Banks/GDP: 0.336*** (0.127) in column (3).
  - Interpretation:
    - Elasticities imply a 1 percent increase in the SME share in total bank lending leads to a 0.6–0.7 pp short-run decline in unemployment; long-run elasticity may rise to 1.
    - A 1 percent increase in SME share leads to a 0.4 pp short-run increase in GDP per capita growth; may rise to 0.6 pp in the long run.
    - Interaction terms indicate declining marginal returns to SME financial inclusion as financial depth increases.
  - Data coverage and estimation:
    - Observations: 263 (col 1); 265 (col 2); 294 (col 3).
    - Number of countries: 36 (cols 1–2); 43 (col 3).
    - Regressions include controls for credit to governments and SOEs, income, population, inflation; country and year fixed effects.

### Economic magnitude and policy implications
- Back-of-the-envelope growth accounting:
  - Additional 2.1 percent SME employment growth plus augmented labor productivity gain of 1.2 percent (assuming labor share equal to two-thirds) imply an additional 1.1 percent of GDP growth (assuming SMEs represent about half of the economy).
  - Authors stress this is a simple exercise indicating potential gains; precise estimates require more analysis.
- Policy recommendations:
  - Monitor both the aggregate amount of credit and its distribution across firm size.
  - Address structural constraints on SME access to finance, notably credit bureau coverage and insolvency regime quality.
  - Reforms to improve credit information and insolvency frameworks could yield measurable employment and labor productivity gains, particularly for SMEs in the Middle East and Central Asia.

### Future research and methodological notes
- Future work suggested:
  - Further analysis of channels through which credit distribution affects optimal monetary and fiscal policies.
  - Development of a theoretical framework to quantify channels and size of benefits more precisely.
  - Reference to IMF (2018c) for deeper policy discussion on SME Financial Inclusion in the region.
- Methodological notes:
  - Firm-level regressions report standard errors in parentheses; ***(**, *) denote significance at 1% (5%, 10%) levels where indicated.
  - Firm-level specifications typically include firm-level controls, country fixed effects, and vary by inclusion of survey year fixed effects.
  - Cross-country regressions include macro controls noted above and use Arellano and Bond GMM estimator to handle short T, wide N panels.

*Source: WP/19/209 — Sections 1–3 (IMF Working Paper, September 2019).*

### Section 1

### WP/19/209 — SME Financial Inclusion for Sustained Growth in the Middle East and Central Asia (Section 1)

### Abstract and Introduction
- Paper authors: Mishel Ghassibe, Maximiliano Appendino, Samir Elsadek Mahmoudi.
- Main claim: Greater financial inclusion of small and medium enterprises (SMEs) can promote higher economic growth and employment, especially in the Middle East and Central Asia regions.
- Key empirical findings summarized:
  - Countries with higher SME financial inclusion exhibit more effective monetary policy transmission and tax collection.
  - Firm-level estimates show substantial employment and labor productivity growth gains from access to credit, with gains higher for SMEs than for large firms.
  - Improved credit bureau coverage and insolvency regimes have a substantial positive impact on SME employment and labor productivity growth.
  - Cross-country aggregate evidence confirms employment and growth gains from SME financial inclusion; gains appear larger in the Middle East and Central Asia than in other regions.
- SME prevalence in the region (IMF, 2018c):
  - SMEs represent about 96 per cent of all registered companies in the region.
  - SMEs employ 48 per cent of the labor force.
- Paper scope and structure:
  - Section II: effect of SME financial inclusion on monetary and fiscal policy effectiveness.
  - Section III: firm-level evidence on employment and labor productivity growth gains and macroeconomic impact.
  - Section IV: cross-country aggregate confirmation.
  - Section V: conclusions and suggestions for future work.

### Measurement and Data
- SME financial inclusion index:
  - Measured using the IMF (2018b) index; calculated as the first principal component of a range of indicators from the World Bank Enterprise Surveys (higher index values correspond to higher SME financial inclusion).
  - Correlations with other measures (world sample): -0.53 and -0.52 with two World Bank measures of credit constraints and rejected loan applications, respectively.
- Country baseline sample (subject to data availability): Jordan, Lebanon, Egypt, Iraq, Morocco, Tunisia, Yemen, Armenia, Azerbaijan, Georgia, Kazakhstan, Kyrgyzstan, Tajikistan, Uzbekistan, Djibouti.
- High vs. Low SME financial inclusion classification:
  - High (indices above the mean): Morocco, Tunisia, Armenia, Georgia.
  - Low (below the mean): Jordan, Egypt, Iraq, Yemen, Azerbaijan, Kazakhstan, Kyrgyzstan, Tajikistan, Uzbekistan.

### II. Macroeconomic Policies and SME Financial Inclusion — Overview
- Hypothesized channels by which SME financial inclusion affects macroeconomic policy:
  - Monetary policy: greater access to formal borrowing and lending increases the role of the interest rate and strengthens monetary policy transmission; better linkage between real and nominal sides may reduce inflation volatility.
  - Fiscal policy: greater SME financial inclusion can boost formalization and improve tax collection efficiency.
  - Financial stability: under appropriate risk management and supervision, broader SME access to formal loans can diversify banks’ credit portfolios and potentially improve stability.

### II.A SME Financial Inclusion and Monetary Policy Transmission — Methods and Results
- Macroeconomic dataset:
  - Annual time series (1990–2017) on real GDP (transformed into output gap via HP filter), inflation, and short-term nominal interest rate.
  - SME financial inclusion index used to subdivide countries into High and Low groups.
- Empirical strategy:
  - Panel VAR: Y_it = A Y_{i,t−1} + u_i + v_t + e_it, with Y_it = {output gap, inflation, nominal interest rate}.
  - Strength of monetary transmission measured by impulse response of output gap to a structural shock to the nominal interest rate.
  - Comparison across High vs. Low SME financial inclusion subsamples.
  - Price stability measured by the ratio of unconditional variances of output and inflation.
- Key numerical findings:
  - Full sample: a 100bp structural increase in the nominal interest rate leads to a peak of 0.13 percentage point (pp) drop in the output gap (not statistically significant at 90 per cent confidence level).
  - High SME financial inclusion subsample: observed 0.15 pp peak drop in the output gap; statistically significant in the first two years following the shock.
  - Low SME financial inclusion subsample: output gap contraction close to and insignificantly different from zero.
- Interpretation:
  - Evidence supports that countries in the Middle East and Central Asia with higher SME financial inclusion benefit from stronger monetary policy transmission.
  - Price stability: High SME financial inclusion countries have a greater ratio of output and inflation variances, suggesting central banks in these countries are better able to minimize price fluctuations (manifested in increased volatility of output gap).
- Caveats:
  - Differences in estimated responses for High vs. Low groups may reflect other heterogeneities (e.g., exchange rate flexibility, central bank independence); validity depends on orthogonality to SME financial inclusion.
  - Estimated benefits assume extra SME funding comes from traditional banking; growing alternative lenders (FinTech) may be less sensitive to monetary policy, which could modify transmission.

### II.B SME Financial Inclusion and Tax Collection — Methods and Results
- Empirical strategy:
  - Panel VAR: Y_it = A Y_{i,t−1} + u_i + v_t + e_it, with Y_it = {Tax/GDP, output gap, inflation}.
  - Efficiency of tax collection measured by impulse response of Tax/GDP to a structural positive shock to the output gap.
  - VAR re-estimated separately for countries in the top and bottom quartiles of the SME Financial Inclusion index.
- Key numerical findings:
  - Full sample: a 1 pp expansion in the output gap leads to an increase in tax-to-GDP ratio of about 0.4 pp after one year (estimate significantly different from zero at the 90 percent confidence level for the first two years).
  - Subsample results:
    - Effect holds in the top-quartile SME financial inclusion subsample.
    - Effect does not hold in the bottom-quartile SME financial inclusion subsample.
- Interpretation:
  - SME financial inclusion strengthens tax collection and is associated with more effective fiscal policy implementation via improved tax-to-GDP responsiveness to output.

### III. Employment and Labor Productivity Gains from SME Financial Inclusion — Preview
- Section III aims to test whether SME financial inclusion boosts economic growth in the Middle East and Central Asia.
- Approach (as introduced in Section 1):
  - Firm-level micro-econometric estimates using World Bank Enterprise Surveys to quantify employment and labor productivity growth gains from SME access to formal finance.
  - Analysis of the impact of reforms that boost SME financial inclusion (credit bureau coverage, insolvency regimes).
  - Use of point estimates in a Solow growth accounting exercise to evaluate aggregate growth effects.
  - Cross-country panel estimates to confirm firm-level findings: increasing the share of total lending to SMEs lowers unemployment and increases economic growth, with larger gains in the Middle East and Central Asia.

*Source: WP/19/209 — Section 1 (IMF Working Paper, September 2019).*

### Section 2

### Section 2

### Empirical approach and data
- Firm-level data: World Bank Enterprise Surveys (2008-2016) covering firms in Azerbaijan, Armenia, Djibouti, Georgia, Iraq, Kazakhstan, Jordan, Kyrgyzstan, Lebanon, Morocco, Tajikistan, Tunisia, Egypt, Uzbekistan and Yemen.
- Firm size classification (Enterprise Survey definition):
  - Large: more than 100 employees.
  - Medium: between 20 and 100 employees.
  - Small: less than 20 employees.
- Main firm-level econometric specification (equation (3)):
  - ∆E_ijt = α F_ijt + β X_ijt + Z_jt + C_j + Y_t + ε_ijt
  - ∆E_ijt: annual employment growth or labor productivity growth for firm i in country j in survey year t.
  - F_ijt: indicator variable for a formal loan outstanding (α interpretable as the employment or productivity gain from access to formal finance).
  - X_ijt and Z_jt: firm-level and country-level controls.
  - C_j and Y_t: country and survey year fixed effects.
- Policy-proxy specification (equation (4)):
  - ∆E_ijt = α S_jt + β X_ijt + Z_jt + C_j + Y_t + ε_ijt
  - S_jt: either credit bureau coverage (share of covered adults) or insolvency regime index (from World Bank Doing Business).
  - α measures the employment or productivity growth gains from an extra percent covered by credit bureaus, or one unit change in the insolvency index.

### Firm-level effects on employment and labor productivity
- Access to formal finance (whole sample):
  - Access to formal finance leads to a 2.07 pp increase in the rate of employment growth for an average firm in the sample.
  - Access to formal finance leads to a 0.94 pp increase in labor productivity growth (whole sample).
- Heterogeneity by firm size:
  - Employment growth gains are concentrated in SMEs rather than large firms.
  - Labor productivity growth:
    - SMEs: 1.22 pp increase in productivity growth when given access to finance.
    - Large firms: 0.43 pp increase in productivity growth when given access to finance.
  - Within SMEs, most of the effect on productivity appears driven by medium firms rather than small firms.
- Notes on estimation:
  - Results for top and bottom halves of SME Financial Inclusion were noisy; authors compare more extreme subsamples.
  - Estimation follows Ayyagari et al. (2016) and excludes firms in top and bottom 1 percent of employment and labor productivity growth.

### Policy interventions: credit bureaus and insolvency regimes
- Credit bureau coverage (share of covered adults):
  - Without survey year fixed effects: a one pp increase in credit bureau coverage leads to a 0.18 pp increase in employment growth (statistically significant).
  - When estimated separately:
    - Large firms: effect small and insignificant.
    - SMEs: 0.23 pp increase in employment growth (statistically significant).
  - With survey year fixed effects: estimates are smaller and noisier, but larger effects for SMEs persist.
  - Labor productivity: a one pp increase in credit bureau coverage increases labor productivity growth by 0.29 pp (statistically significant without survey year fixed effects); estimated effect comparable for large firms and SMEs in these regressions.
- Insolvency regime index (Doing Business resolving insolvency index; varies from 0 to 100):
  - Without survey year fixed effects: a unit increase in the insolvency regime index leads to a 0.23 pp increase in employment growth (statistically significant).
    - Large firms: 0.12 pp.
    - SMEs: 0.30 pp.
  - With survey year fixed effects: effects become larger and the SME effect remains bigger than for large firms.
  - Labor productivity: a unit increase in the insolvency regime index increases labor productivity growth by 0.57 pp (no survey year fixed effects).
    - SMEs: 0.60 pp.
    - Large firms: 0.44 pp.
- Cumulative gains moving to EMDE third quartile:
  - Moving to the third quartile among EMDEs for credit bureau coverage:
    - SME employment growth would rise by close to 3 pp.
    - SME labor productivity growth would rise by close to 3.5 pp.
  - Moving to the third quartile among EMDEs for the insolvency regime index:
    - SME employment growth estimated to rise by 2.5 pp.
    - SME labor productivity growth estimated to rise by 4.5 pp.
  - Caveat: these cumulative estimates assume linear effects and may be an upper bound if marginal returns diminish.

### Cross-country analysis: SME share in bank lending and macro effects
- Data: IMF Financial Access Survey (FAS) cross-country data; sample of countries listed in the source text.
- Stylized facts:
  - In Middle East and Central Asia economies, SMEs have consistently lower shares in total bank lending than in other regions.
  - SMEs’ shares in total lending have tended to decrease over time in the Middle East and Central Asia and in the overall FAS sample, coinciding with an increase in the share of governments and SOEs in bank lending.
- Dynamic panel specification (equation (5)):
  - U_it = δ U_it−1 + β1 SME_Share_it + β2 FD_it + β3 SME_Share_it × FD_it + θ X_it + γ_t + α_i + ε_it
  - U_it: real sector outcome (e.g., unemployment or GDP per capita growth).
  - FD_it: financial depth = ln(outstanding commercial bank loans to resident nonfinancial corporations and households relative to GDP).
  - SME_Share_it: ln(share of SMEs in total bank lending).
  - X_it: country-level time-varying controls (credit to governments and SOEs, income, population, inflation).
  - Estimation method: Arellano and Bond (1991) GMM estimator to address short time dimension and wide cross-section.
- Key macro results (Table 7 summary):
  - Employment/unemployment:
    - Columns (1) and (2): elasticity range of -0.6–0.7 implying a 1 percent increase in the share of SMEs in total bank lending leads to a 0.6–0.7 pp decline in the unemployment rate in the short run; long-run elasticity may rise to 1.
    - Interaction term (SME share × financial depth) is positive, indicating marginal employment benefits of financial inclusion are lower at higher levels of financial development (declining returns to inclusion as depth increases).
  - Economic growth:
    - Column (3): a 1 percent increase in the share of SMEs in total bank lending leads to a 0.4 pp increase in GDP per capita growth in the short run; may increase to 0.6 pp in the long run.
    - Negative interaction term between financial inclusion and financial depth suggests declining returns to financial inclusion as financial depth increases.
- Interpretation:
  - Cross-country results confirm firm-level findings of positive employment and growth gains from higher SME financial inclusion.
  - Given lower and declining SME shares in bank lending in the Middle East and Central Asia, gains from improving SME financial inclusion could be especially pronounced in the region.

### Economic magnitude and policy implications
- Simple growth accounting exercise (based on firm-level point estimates and assumptions):
  - Additional 2.1 percent in SME employment growth plus augmented labor productivity gain of 1.2 percent (assuming labor share equal to two-thirds) imply an additional 1.1 percent of GDP growth (assuming SMEs represent about half of the economy).
  - Authors note this is a simple exercise providing indications of gains; more precise estimates require further analysis.
- Policy implications emphasized:
  - Policymakers should monitor not only the aggregate amount of credit but also the distribution of credit across firm size.
  - Structural constraints on SME access to finance matter, including credit bureau coverage and insolvency regime quality.
  - Reforms to improve credit information and insolvency frameworks could yield measurable employment and labor productivity gains, particularly for SMEs in the Middle East and Central Asia.

*Source: wpiea2019209-print-pdf - Section 2*

### Section 3

### wpiea2019209-print-pdf - Section 3

### Future research and policy gaps
- Future academic work could address further the channels through which these distributional issues matter for the formulation of optimal monetary and fiscal policies.
- Future academic work could provide an adequate theoretical framework to quantify more precisely the channels and size of these benefits.
- See IMF (2018c) for a deeper discussion on policy issues related to SME Financial Inclusion in the Middle East and Central Asia.

### Key empirical findings (firm-level regressions)
- Table 1 — Effect of Access to Formal Finance on Employment Growth (dependent variable: employment growth (percent))
  - Access to formal finance: 2.07*** (0.31) — All firms
  - Access to formal finance: 0.89* (0.52) — Large firms (>100)
  - Access to formal finance: 2.14*** (0.38) — SMEs (<100)
  - Access to formal finance: 2.31*** (0.60) — Small (<20)
  - Access to formal finance: 1.84*** (0.51) — Medium (20-100)
  - R2: 0.09 (All firms); 0.08 (Large); 0.10 (SMEs); 0.11 (Small); 0.10 (Medium)
  - No. of observations: 10,402 (All firms); 2,059 (Large); 8,343 (SMEs); 4,286 (Small); 4,057 (Medium)
  - Firm-level controls: Yes; Country fixed effects: Yes; Survey year fixed effects: Yes

- Table 2 — Effect of Access to Formal Finance on Labor Productivity Growth (dependent variable: labor productivity growth (percent))
  - Access to formal finance: 0.94 (0.62) — All firms
  - Access to formal finance: 0.43 (1.19) — Large firms (>100)
  - Access to formal finance: 1.22* (0.74) — SMEs (<100)
  - Access to formal finance: -0.26 (1.16) — Small (<20)
  - Access to formal finance: 2.39** (0.97) — Medium (20-100)
  - R2: 0.21 (All firms); 0.07 (Large); 0.24 (SMEs); 0.28 (Small); 0.20 (Medium)
  - No. of observations: 7,999 (All firms); 1,613 (Large); 6,386 (SMEs); 3,241 (Small); 3,145 (Medium)
  - Firm-level controls: Yes; Country fixed effects: Yes; Survey year fixed effects: Yes

- Table 3 — Effect of Credit Bureau Coverage on Employment Growth (dependent variable: employment growth (percent))
  - Credit Bureau coverage (percent): 0.18*** (0.03) — All firms
  - Credit Bureau coverage (percent): 0.08 (0.07) — Large (>100)
  - Credit Bureau coverage (percent): 0.23*** (0.04) — SMEs (<100)
  - Alternative specifications (with/without survey year fixed effects) report coefficients: 0.04 (0.06); -0.02 (0.15); 0.07 (0.06)
  - R2: 0.08 (All firms); 0.08 (Large); 0.09 (SMEs) across specifications
  - No. of observations: 10,585 (All firms); 2,126 (Large); 8,459 (SMEs)
  - Firm-level controls: Yes; Country fixed effects: Yes; Survey year fixed effects: mixed (No/Yes depending on column)

- Table 4 — Effect of Insolvency Regime on Employment Growth (dependent variable: employment growth (percent))
  - Insolvency regime (index): 0.23*** (0.03) — All firms
  - Insolvency regime (index): 0.12** (0.05) — Large (>100)
  - Insolvency regime (index): 0.30*** (0.04) — SMEs (<100)
  - Additional reported coefficients in alternative specifications: 0.57*** (0.13); 0.60* (0.32); 0.63*** (0.15)
  - R2: 0.09 (All firms); 0.08 (Large); 0.10 (SMEs)
  - No. of observations: 10,585 (All firms); 2,126 (Large); 8,459 (SMEs)
  - Firm-level controls: Yes; Country fixed effects: Yes; Survey year fixed effects: mixed (No/Yes)

- Table 5 — Effect of Credit Bureau Coverage on Labor Productivity Growth (dependent variable: labor productivity growth (percent))
  - Credit Bureau coverage (percent): 0.29*** (0.07) — All firms
  - Credit Bureau coverage (percent): 0.32** (0.15) — Large (>100)
  - Credit Bureau coverage (percent): 0.27*** (0.07) — SMEs (<100)
  - Alternative specification coefficients (with survey year fixed effects) report negative estimates: -0.42*** (0.12); -0.89** (0.37); -0.30** (0.13)
  - R2: 0.20 (All firms); 0.06 (Large); 0.22 (SMEs) in initial columns; 0.22/0.08/0.24 in alternative columns
  - No. of observations: 8,120 (All firms); 1,657 (Large); 6,463 (SMEs)
  - Firm-level controls: Yes; Country fixed effects: Yes; Survey year fixed effects: mixed (No/Yes)

- Table 6 — Effect of Insolvency Regime on Labor Productivity Growth (dependent variable: labor productivity growth (percent))
  - Insolvency regime (index): 0.57*** (0.06) — All firms
  - Insolvency regime (index): 0.44*** (0.12) — Large (>100)
  - Insolvency regime (index): 0.60*** (0.07) — SMEs (<100)
  - Alternative specification coefficients (with survey year fixed effects) report negative point estimates for All firms and subgroups: -0.67*** (0.29); -1.33 (0.87); -0.45 (0.31)
  - R2: 0.20 (All firms); 0.06 (Large); 0.23 (SMEs) in initial columns; 0.22/0.08/0.24 in alternative columns
  - No. of observations: 8,120 (All firms); 1,657 (Large); 6,463 (SMEs)
  - Firm-level controls: Yes; Country fixed effects: Yes; Survey year fixed effects: mixed (No/Yes)

### Cross-country evidence (macro regressions)
- Table 7 — Cross-Country Evidence on Effect of SME Financial Inclusion on Unemployment and GDP per Capita Growth
  - Additional controls include measures of credit to governments and SOEs, income, population and inflation.
  - Unemployment last period, U_it−1:
    - Coefficient: 0.210** (0.100) in column (1)
    - Coefficient: 0.278*** (0.0897) in column (2)
  - GDP per capita growth, U_it−1:
    - Coefficient: 0.551*** (0.0664) in column (3)
  - Share of SMEs in total Bank Lending:
    - -0.664*** (0.245) in column (1)
    - -0.692* (0.400) in column (2)
    - 0.426*** (0.159) in column (3)
  - Loans with commercial Banks/GDP:
    - -0.420 (0.428) in column (1)
    - 0.336*** (0.127) in column (3)
  - Interaction of SME Share and Total lending:
    - 0.104* (0.0610) in column (1)
    - -0.0848** (0.0389) in column (2)
  - Private credit/GDP:
    - -0.507 (0.318) reported in one specification
  - Interaction of SME Share and private credit/GDP:
    - 0.199* (0.116) in one specification
  - Country fixed effects: Yes; Year fixed effects: Yes
  - Observations: 263 (col 1); 265 (col 2); 294 (col 3)
  - Number of countries: 36 (cols 1–2); 43 (col 3)

### Methodological notes and controls
- Regressions report standard errors in parentheses; ***(**, *) denotes statistical significance from zero at 1% (5%, 10%) level where indicated.
- Firm-level specifications typically include firm-level controls, country fixed effects, and vary by inclusion of survey year fixed effects.
- Cross-country regressions include additional macro controls noted above.

*Source: wpiea2019209-print-pdf - Section 3*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019209-print-pdf.pdf_