## _wp13246

## Source details

**Canonical URL:** [_wp13246](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13246.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13246.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13246.pdf.json)

---

### Introduction and motivation
- Addresses the quantitative impact of credit constraints on capital accumulation, noting limitations in prior empirical approaches:
  - The ‘investment-cash flow sensitivity’ approach (Fazzari et al. 1988) is viewed as an indirect measure of financial constraints with weak theoretical underpinnings.
  - The ‘survey-based’ qualitative approach (Kaplan and Zingales 1997) lacks quantitative rigor.
- Recent literature questions the link between cash flow sensitivity of investment and financial constraints (Chen and Chen 2012; Laeven 2003; Kaplan and Zingales 1997) and documents a reduced cash flow elasticity of investment (Chen and Chen, 2012; Brown and Petersen, 2009; Andersen and others, 2012; Guariglia and Poncet, 2007).
- Focus: MENA region using a unique firm-level dataset from Middle East and North Africa (MENA).

### Methodology: model and estimation approach
- Two-step parametric implementation of the Kiyotaki and Moore (1997) (KM) dynamic model:
  - Step 1: Stochastic Frontier Analysis (SFA) of equation (7) to estimate unobserved credit limit (CL) and distance from the credit limit (DL).
  - Step 2: Insert predicted changes in CL and DL into the dynamic capital accumulation equation (equation (5)) and estimate via dynamic pooled OLS.
- Credit limit estimation follows Herrala (2009) methodology and Fungacova and others (2013) application.
- Key empirical choices and controls:
  - Z1: change in real interest rate (R), time, country, and sector dummies, Arab Spring dummy, political unrest dummy, GCC dummy, and DL.
  - Z2: firm size (alternatives: total assets book value, equity capital, number of employees, SME dummy), firm age, pretax return on equity, and country/sector dummies.
  - DL distribution assumptions: half normal, truncated normal, exponential.
- Robustness checks include alternative DL distributions, alternative endogenous variables (short-term debt), alternative firm-size proxies (total assets), alternative interest-rate measures (short-term rates), individual country dummies, inclusion of firm profitability, errors-in-variables correction (Murphy and Topel, 1985), additional lags, higher order terms, country fixed effects, and changes in consumer prices.

### Data, sample, and coverage
- Firm-level data sources: Orbis and Zawya for 2007-2010.
- Sample size retained: 860 companies and 1,483 observations (unbalanced panel).
- Countries covered (data availability basis): Bahrain, Egypt, Iraq, Jordan, Kuwait, Lebanon, Morocco, Oman, Qatar, Saudi Arabia, Sudan, Syrian Arab Republic, Tunisia, United Arab Emirates, and West Bank and Gaza.
- Financial data units: thousands of US dollars (as noted).

### Key empirical findings and key statistics
- Marginal effect of credit limits on capital accumulation:
  - Estimated marginal effect about 40 percent (reported range across specifications: between 0.392 and 0.402 in Table 7; interpreted in text as between 40 to 48 percent).
  - Authors’ interpretation: implies that between 40 to 48 percent of firms in MENA are credit constrained and that a change in credit limits by one unit contributes between 40 and 48 percent increase in capital accumulation in the following year.
- Lagged capital elasticity:
  - Fixed Assets, Lagged coefficients: 0.994***, 0.995***, 0.995***, 0.994*** (Table 7).
- Distance from limit (DL) change in capital accumulation regressions:
  - Lagged Change coefficients: -0.0235, -0.028, -0.0225, -0.0268 (not significant).
- SFA (credit limit) results (selected coefficients from Table 5):
  - Equity Capital coefficients (CL1-CL4): 1.039***, 1.043***, 1.051***, 1.015***.
  - Real Interest Rate coefficients: 0.0267***, 0.0269***, 0.0211**, 0.0249**.
  - Political Unrest coefficients: -0.398**, -0.374**, -0.522***, -0.265.
  - Arab Spring coefficients: 0.349***, 0.354***, 0.282***, 0.331***.
  - GCC coefficients: 0.298***, 0.302***, 0.305***, 0.281***.
- Use of credit relative to estimated limits (Table 6):
  - Average ratio of total debt to estimated credit limits across firms: 61.52 percent (Panel A average).
  - Country range for mean debt-to-credit-limit ratio: 55.04 (Lebanon) to 69.97 (West Bank & Gaza).
  - Sector average debt-to-credit-limit: Average 60.55 percent (Panel B).
- Descriptive regional and sample statistics (selected):
  - Only 10 percent of MENA firms make use of bank financing (World Bank Business Environment Survey citation).
  - Close to 40 percent of MENA firms identify access to finance as a major obstacle to growth.
  - IMF estimate: between 50 and 75 million new jobs are needed over the next decade in MENA to secure social and political stability.
  - Increase in unemployed people since onset of Arab Spring: “more than one million”; unemployment rates vary between 9 and 15 percent; youth unemployment reaching up to 30 percent in some countries (Ahmed, 2013).
  - Sample-level averages (Table 1 totals/averages):
    - Number of firms: 860.
    - Number of observations: 1,483.
    - Average Total Assets: 897,272.
    - Average Total Debt: 368,127.
    - Average Equity Capital: 483,802.
    - Average debt-to-equity ratio (narrative): 66 percent.
    - Average Current-Liabilities-to-Equity ratio (narrative): 58 percent.
- Regional domestic credit provision (Table 4 aggregates):
  - GCC domestic credit provided by banking sector (% of GDP): 47.73 (2007), 47.09 (2008), 64.55 (2009), 56.9 (2010).
  - Non-GCC: 71.54 (2007), 71.39 (2008), 71.07 (2009), 71.67 (2010).
- Skewness diagnostic (SFA residuals, Table 5): negative skew test statistics -14.35, -14.36, -13.97, -14.31 with significance 000.

### Interpretation and mechanisms
- Economic significance:
  - A one-unit improvement in credit limits yields a large contribution to capital accumulation (about 40 percent effect), indicating improved financing conditions are a key channel for private-sector growth and macroeconomic development in MENA.
- Political instability channel:
  - Political unrest significantly reduces estimated credit limits (negative coefficients in CL regressions).
  - Political unrest and Arab Spring dummies are insignificant in the capital accumulation regressions, implying political events affect capital accumulation indirectly via financial conditions rather than directly altering investment dynamics in the sample period.
  - Positive Arab Spring coefficient in CL regressions suggests Arab Spring countries had, on average, higher credit limits prior to the onset—consistent with the argument that revolutions may follow a ‘threshold’ level of economic or financial development.
- Financial-market frictions and institutional context:
  - Insignificance of DL in capital accumulation regressions aligns with KM: unused (non-binding) portions of credit limits do not affect investment.
  - Absence of DL effect may also reflect weak credit-risk assessment, limited judicial enforcement, limited credit-bureau coverage, and underdeveloped collateral/bankruptcy frameworks in the region.

### Policy implications and recommendations
- Improve financing conditions to foster private-sector development and inclusive growth:
  - Target financial inclusion by relaxing financing constraints on firms.
- Strengthen financial infrastructure to make credit markets more responsive and risk-sensitive:
  - Enhance credit information sharing (credit bureaus and public registries), judicial enforcement, and collateral and bankruptcy frameworks to improve risk pricing and the effectiveness of lending rates as a policy instrument.
- Monitor political stability effects on credit supply:
  - Policies to preserve macroeconomic stability and maintain credit availability during transitions are important given political unrest lowers credit limits.
- Broader research agenda:
  - The two-step empirical implementation of the KM framework is presented as a template for similar studies in other regions.

### Robustness and sensitivity
- Results robust to:
  - Alternative DL distributions (truncated normal, exponential, half normal).
  - Alternative endogenous variables (short-term debt), alternative firm-size indicators (total assets), alternative interest rate measures (short-term), inclusion of profitability, consumer prices, and country dummies.
  - Errors-in-variables correction for estimation error in CL and DL; qualitative results remain and quantitative impacts can be larger under corrected standard errors.
  - Inclusion of countries and sectors with few firms/observations.
- Additional diagnostics:
  - Negative skewness in SFA residuals consistent with an upper bound on firm debt imposed by credit limits.

*Source: _wp13246 - 1. Summary statistics by country, 2007-2010; Section 6 concludes.*

### 1. Summary statistics by country, 2007-2010 ..........................................................................10

### _wp13246 - 1. Summary statistics by country, 2007-2010 ..........................................................................10

### Introduction and motivation
- The paper addresses the quantitative impact of credit constraints on capital accumulation, noting limitations in prior empirical approaches:
  - The ‘investment-cash flow sensitivity’ approach (Fazzari et al. 1988) is viewed as an indirect measure of financial constraints and has weak theoretical underpinnings.
  - The ‘survey-based’ qualitative approach (Kaplan and Zingales 1997) lacks quantitative rigor.
- Recent literature questions the link between cash flow sensitivity of investment and financial constraints (Chen and Chen 2012; Laeven 2003; Kaplan and Zingales 1997) and documents a reduced cash flow elasticity of investment (Chen and Chen, 2012; Brown and Petersen, 2009; Andersen and others, 2012; Guariglia and Poncet, 2007).
- The study focuses on the MENA region and uses a unique firm-level dataset from Middle East and North Africa (MENA).

### Methodology: model and estimation approach
- The paper builds on the Kiyotaki and Moore (KM, 1997) model and reformulates it for empirical estimation:
  - Credit constraints are not directly observable; the paper estimates unobservable credit constraints (denoted as credit limits) using stochastic frontier analysis (SFA) of the loan distribution for a sample of firms.
  - The estimated credit limits are then incorporated into a dynamic regression framework to quantify their marginal effect on capital accumulation.
- Credit limit estimation follows methodology presented by Herrala (2009) and more recently employed by Fungacova and others (2013).
- The two-step estimation strategy circumvents difficulties of bringing the KM model to data by first uncovering firm-level credit limits and then estimating their effect on capital accumulation.

### Key empirical findings
- Impact of credit constraints on capital accumulation:
  - The marginal effect of a change in credit limits on capital accumulation is estimated at about 40 percent.
  - The effect is both economically and statistically significant.
  - Results are robust to changes in model specification.
- Political instability and financial constraints:
  - Credit limits get tighter amid prolonged political uncertainty.
  - The dynamic effects of continued political unrest on capital accumulation are insignificant, consistent with Bloom (2009).
- Regional and pre-unrest conditions:
  - Firms operating in Arab Spring countries seemed to enjoy higher credit limits prior to the revolution, on average.
  - No evidence of significant differences in capital accumulation between Arab Spring countries and other countries in the region can be detected empirically.
  - The paper suggests revolutions may be triggered after some ‘threshold’ level of economic or financial development is reached, in line with Acemoglu and Robertson (2012).

### Contributions and implications
- Methodological contribution:
  - Provides a parametric, two-step empirical implementation of a KM-based framework that yields quantifiable measures of credit constraints (credit limits) and their effect on capital accumulation.
- Substantive contributions:
  - Demonstrates a quantitatively large role for credit constraints in capital accumulation in MENA economies.
  - Shows that political instability tightens credit limits, posing a challenge to maintaining a well-functioning financial system during prolonged uncertainty.
- Policy-relevant implication:
  - Financial development in MENA countries—measured by a relaxing of financial constraints—is key to macroeconomic development in the region.

### Paper organization (as provided)
- Section 2: Derives estimable equations of credit limits and capital accumulation, building on the KM model.
- Section 3: Provides an overview of key macroeconomic and financial conditions in the MENA region, sociopolitical challenges, and describes the unique dataset.
- Section 4: Discusses estimation results.
- Section 5: Conducts robustness checks.

*Source: _wp13246 - 1. Summary statistics by country, 2007-2010 (PDF chapter).*

### Section 6 concludes.

### _wp13246 - Section 6 concludes.

### Methodology and data
- Empirical approach: two-step parametric implementation of Kiyotaki and Moore (1997) (KM) dynamic model.
  - Step 1: Stochastic Frontier Analysis (SFA) of equation (7) to estimate unobserved credit limit (CL) and distance from the credit limit (DL).
  - Step 2: Insert predicted changes in CL and DL into dynamic capital accumulation equation (equation (5)) and estimate via dynamic pooled OLS.
- Sample and coverage:
  - Firm-level data from Orbis and Zawya for the period 2007-2010.
  - Sample size retained: 860 companies and 1,483 observations (unbalanced panel).
  - Countries covered (data availability basis): Bahrain, Egypt, Iraq, Jordan, Kuwait, Lebanon, Morocco, Oman, Qatar, Saudi Arabia, Sudan, Syrian Arab Republic, Tunisia, United Arab Emirates, and West Bank and Gaza.
- Key empirical choices and controls:
  - Z1 includes change in real interest rate (R), time, country, and sector dummies, Arab Spring dummy, political unrest dummy, GCC dummy, and DL (distance) to test for pre-binding effects.
  - Z2 includes firm size (alternatives: total assets book value, equity capital, number of employees, SME dummy), firm age, pretax return on equity, and country/sector dummies.
  - DL distribution assumptions considered: half normal, truncated normal, exponential.
- Robustness steps:
  - Alternative distributions for DL, alternative endogenous variables (short-term debt), alternative firm-size proxies (total assets), alternative interest-rate measures (short-term rates), individual country dummies, inclusion of firm profitability.
  - Errors-in-variables correction for estimation error in predicted CL and DL using Murphy and Topel (1985) approach.
  - Results tested with additional lags, higher order terms, country fixed effects, and changes in consumer prices.

### Main empirical findings
- Credit constraints and capital accumulation:
  - The estimated marginal effect of a change in credit limits on capital accumulation is about 40 percent (reported range across specifications: between 0.392 and 0.402 in Table 7; interpreted as between 40 to 48 percent in text).
    - Authors’ interpretation: implies that between 40 to 48 percent of firms in MENA are credit constrained and that a change in credit limits by one unit contributes between 40 and 48 percent increase in capital accumulation in the following year.
  - Lagged capital elasticity (capital’s own elasticity) is very close to unity across models (Fixed Assets, Lagged coefficients: 0.994***, 0.995***, 0.995***, 0.994*** as reported in Table 7).
  - Distance from limit (DL) change is insignificant in capital accumulation regressions (Distance from Limit, Lagged Change coefficients: -0.0235, -0.028, -0.0225, -0.0268; not significant).
- Credit limits (SFA results):
  - Equity capital has a positive and highly significant marginal effect on credit limits; coefficient slightly above unity in CL1-CL4 (Equity Capital coefficients: 1.039***, 1.043***, 1.051***, 1.015*** in Table 5).
  - Real Interest Rate effect on credit limit is unexpectedly positive and significant in CL specifications (Real Interest Rate coefficients: 0.0267***, 0.0269***, 0.0211**, 0.0249** in Table 5).
  - Political unrest dummy is associated with lower credit limits (Political Unrest coefficients: -0.398**, -0.374**, -0.522***, -0.265 in Table 5).
  - Arab Spring dummy is positive and highly significant in CL regressions (Arab Spring coefficients: 0.349***, 0.354***, 0.282***, 0.331*** in Table 5).
  - GCC dummy is positive and highly significant in CL regressions (GCC coefficients: 0.298***, 0.302***, 0.305***, 0.281*** in Table 5).
- Use of credit relative to estimated limits:
  - Average ratio of total debt to estimated credit limits across firms: 61.52 percent (Panel A average in Table 6: Average 61.52).
  - Country range for mean debt-to-credit-limit ratio: from 55.04 (Lebanon) to 69.97 (West Bank & Gaza) as reported in Table 6 Panel A.
  - Sector average debt-to-credit-limit: Average 60.55 percent (Panel B in Table 6).
- Descriptive and contextual statistics (selected):
  - Only 10 percent of MENA firms make use of bank financing (World Bank Business Environment Survey cited).
  - Close to 40 percent of MENA firms identify access to finance as a major obstacle to growth.
  - IMF estimate cited: between 50 and 75 million new jobs are needed over the next decade in MENA to secure social and political stability.
  - Increase in unemployed people since onset of Arab Spring: “more than one million”; unemployment rates vary between 9 and 15 percent; youth unemployment reaching up to 30 percent in some countries (Ahmed, 2013).
  - Sample-level balance-sheet averages (from Table 1 totals/averages):
    - Total / Average: Number of firms 860; Number of observations 1,483.
    - Average Total Assets: 897,272 (financial data are in thousands of US dollars as noted).
    - Average Total Debt: 368,127 (thousands of US dollars).
    - Average Equity Capital: 483,802 (thousands of US dollars).
    - Average Debt/Equity: 66.8 percent (table text reports average debt-to-equity ratio is 66 percent in narrative; Table 1 last row shows 65.8?; preserve narrative: “average debt-to-equity ratio is 66 percent”).
    - Average Current-Liabilities-to-Equity ratio: 58 percent (narrative).
  - Regional domestic credit provision (Table 4 highlights):
    - GCC average domestic credit provided by banking sector (% of GDP) reported as 47.73 for 2007, 47.09 for 2008, 64.55 for 2009, 56.9 for 2010 (Table 4 aggregates for GCC row).
    - Non-GCC: 71.54 (2007), 71.39 (2008), 71.07 (2009), 71.67 (2010) (Table 4 aggregates for Non-GCC row).

### Interpretation and mechanisms
- Economic significance:
  - A one-unit improvement in credit limits yields a large contribution to capital accumulation (about 40 percent effect), implying that improved financing conditions are a key channel for private-sector growth and macroeconomic development in MENA.
- Political instability channel:
  - Political unrest has a significant negative effect on estimated credit limits (reducing credit availability).
  - Political unrest and Arab Spring dummies are insignificant in the capital accumulation regressions, implying political events affect capital accumulation indirectly via financial conditions rather than directly altering investment dynamics in the sample period.
  - The positive Arab Spring coefficient in the CL regressions suggests that countries that experienced Arab Spring had, on average, higher credit limits prior to the onset—consistent with the argument that revolutions are more likely after some level of financial and economic development has been reached.
- Financial-market frictions and institutional context:
  - Insignificance of DL in capital accumulation regressions aligns with KM: unused portions of credit limits (non-binding constraints) do not affect investment.
  - The absence of a negative DL effect may also reflect weak credit-risk assessment, limited judicial enforcement, limited credit-bureau coverage, and underdeveloped collateral/bankruptcy frameworks in the region.

### Policy implications and recommendations (as emphasized by authors)
- Improve financing conditions to foster private-sector development and inclusive growth:
  - Target financial inclusion by relaxing financing constraints on firms—this is highlighted as an important means to foster private sector development and inclusive growth.
- Strengthen financial infrastructure to make credit markets more responsive and risk-sensitive:
  - Enhance credit information sharing (credit bureaus and public registries), judicial enforcement, and collateral and bankruptcy frameworks to improve risk pricing and the effectiveness of lending rates as a policy instrument.
- Monitor political stability effects on credit supply:
  - Given political unrest lowers credit limits, policies to preserve macroeconomic stability and maintain credit availability during transitions are important for sustaining investment.
- Broader research agenda:
  - The empirical two-step approach that quantifies the KM model is presented as a framework for similar studies in other regions.

### Robustness and sensitivity
- Results robust to:
  - Alternative DL distributions (truncated normal, exponential, half normal).
  - Alternative endogenous variables (short-term debt), alternative firm-size indicators (total assets), alternative interest rate measures (short-term), and inclusion of additional controls (profitability, consumer prices, country dummies).
  - Errors-in-variables correction for estimation error in CL and DL; qualitative results remain and quantitative impact can be larger under corrected standard errors.
  - Inclusion of countries and sectors with few firms/observations.
- Skewness diagnostic:
  - Negative skewness tests on residuals from SFA indicate significant negative skew consistent with an upper bound on firm debt imposed by credit limits (negative skew test statistics reported in Table 5: -14.35, -14.36, -13.97, -14.31 with significance 000).

*Source: Section 6 and supporting sections of the provided IMF working paper content (_wp13246 - Section 6 concludes.).*

### References

### _wp13246 - References

### Academic studies on financial constraints, investment, and credit
- Abdallah, C. S., and W. D. Lastrapes, 2012, “Home Equity Lending and Retail Spending: Evidence from a Natural Experiment in Texas”, American Economic Journal: Macroeconomics, 4(4), 94-125.
- Andersen, T. B., S. Jones, and F. Tarp, 2012, “The Finance-Growth Thesis: A Skeptical Assessment,” Journal of African Economies, Vol 21, AERC Supplement 1, i57-i88.
- Bernanke, B. S., M. Gertler, and S. Gilchrist, 1999, “The Financial Accelerator in a Quantitative Business Cycle Framework,” In: J. B. Taylor & M. Woodford (eds.), Handbook of Macroeconomics, Ed. 1, Vol. 1, Ch. 21, 1341-1393, Elsevier.
- Bhaumik, S. K., P. K. Das, and S. C. Kumbakhar, 2012, “A Stochastic Frontier Approach to Modeling Financial Constraints in Firms: An Application to India,” Journal of Banking and Finance, 36, 1311-1319.
- Bloom, N., 2009, “The Impact of Uncertainty Shocks,” Econometrica, Vol. 77, No. 3, 623-685.
- Brown, J. R. and B. C. Petersen, 2009, “Why has the Investment-Cash Flow Sensitivity Declined so Sharply? Rising R&D and Equity Market Developments,” Journal of Banking and Finance, 33, 971-984.
- Chen, H. and S. Chen, 2012, “Investment-cash-flow Sensitivity Cannot be a Good Measure of Financial Constraints: Evidence from the Time Series,” Journal of Financial Economics, 103,393-410.
- Fazzari, S. M., R. G. Hubbard, B. Petersen, A. Blinder, and J. M. Poterba, 1988, “Financing Constraints and Corporate Investment,” Quarterly Journal of Economics, 10988(1), 141-206.
- Fungacova, Z., R. Herrala, and L. Weill, 2013, “The Influence of Bank ownership on Credit Supply: Evidence from the Recent Financial Crisis,” Emerging Markets Review, 15, 136-147.
- Guariglia, A. and S. Poncet, 2008, “Could Financial Distortions be no Impediment to Economic Growth After All? Evidence from China,” Journal of Comparative Economics, 36, 633-657.
- Hubbard, R. G., 1998, “Capital Market Imperfections and Investment,” Journal of Economic ‎Literature, 36(1), 193-225.
- Kaplan, S. N. and L. Zingales, 1997, “Do Investment-Cash Flow Sensitivities Provide Useful Measures of Financing Constraints?” The Quarterly Journal of Economics, 112(1), 169-215.
- Kiyotaki, N. and J. Moore, 1997, “Credit Cycles” Journal of Political Economy 105, 211-248.
- Rocha, R., Z. Arvai, and S. Farazi, 2011, “Financial Access and Stability: A Road Map for the Middle East and North Africa,” MENA Development Report, the World Bank, Washington D.C.

### Methodological and econometric references
- Jondrow, J., C. A. K. Lovell, I. S. Materov, and P. Schmidt, 1982, “On the Estimation of Technical Inefficiency in the Stochastic Frontier Production Function Model,” Journal of Econometrics 19, 233-238.
- Murphy, K. M. and R. H. Topel, 1985, “Estimation and Inference in Two-Step Econometric Models,” Journal of Business and Economic Statistics, 3, 370–379.

### Regional and policy-oriented reports, data sources, and commentary
- Ahmed, M., 2013, “Growth and Jobs to Meet the Aspirations of the Arab Countries in Transition”, IMF Commentary published in Almasry Alyoum and available at:  http://www.imf.org/external/np/vc/2013/051913.htm .
- Herrala, R., 2009, “Credit Crunch? An Empirical Test of Cyclical Credit Policy,” Bank of Finland Research Discussion Papers 10/09.
- International Monetary Fund, Regional Economic Outlook Update Middle East and Central Asia, 2013, International Monetary Fund, April, Washington D.C.
- World Bank, World Business Environment Survey (WBES), available at http://econ.worldbank.org/WBSITE/EXTERNAL/EXTDEC/EXTRESEARCH/0,,contentMDK:20699364~pagePK:64214825~piPK:64214943~theSitePK:469382,00.html

*Source: _wp13246 - References*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13246.pdf_