## _wp10115

## Source details

**Canonical URL:** [_wp10115](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10115.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10115.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10115.pdf.json)

---

### I. Research context, scope, and objectives
- Research context:
  - Since King and Levine (1993), more than 200 research papers have considered the finance–growth nexus, about half published in academic journals.
  - Mainstream view: banks stimulate growth by improving capital allocation, reducing transaction costs, and monitoring projects.
- WAEMU context and scope:
  - WAEMU comprises eight countries: Benin, Burkina Faso, Côte d’Ivoire, Guinea-Bissau, Mali, Niger, Senegal, and Togo; they share the same currency and a common central bank and bank regulator (BCEAO and the Regional Banking Commission).
  - Despite institutional homogeneity, disparities in financial development and maturity composition of private credit exist across member countries.
- Objectives:
  - Assess whether more financially developed WAEMU countries tend to grow faster.
  - Analyze whether banking systems that provide more short-term credit than long-term credit undermine the contribution of financial development to economic growth, and investigate why.

### II. Theoretical background and role of credit maturity
- Core mechanisms (Levine (1997) summarized):
  - Financial sector mobilizes resources and channels funds to highest-return projects, fosters technological innovation and capital accumulation.
  - Banks provide households safe and liquid deposit opportunities and mixes of liquid and illiquid investments that facilitate long-run investments.
  - Financial intermediaries lower transaction and information costs and monitor corporate managers, design contracts and collateral to lower monitoring/enforcement costs.
- Short- versus long-term credit:
  - Short-term credit: bank loans with maturity of less than a year; includes commercial paper, lines of credit, overdrafts; finances working capital and short-lived investments.
  - Long-term credit: loans repayable beyond a year without rollover provisions; suitable for new plant, R&D, technology adoption and other high-return, long-maturity investments.
  - Theoretical expectation: long-term credit has larger impact on productivity growth; scarcity of long-term credit increases growth volatility and reduces long-term growth; SMEs are disproportionately affected by constraints on long-term financing.

### III. Stylized facts and constraints in WAEMU
- Aggregate and maturity composition:
  - Average private credit to GDP stood at 12 percent for WAEMU countries over 1995–2006.
  - Short-term credit accounts for the bulk of bank financing to the private sector, averaging 70 percent of private credit.
  - Country ranges for short-term credit share: 60 percent in Togo and Senegal to nearly 90 percent in Guinea-Bissau; Guinea-Bissau: short-term credit accounts for over 90 percent of total credit.
  - With short-term credit dominating (about 70 percent), member countries are less able to reap the full benefit from financial development.
- Comparative and country-level indicators:
  - 15.6 percent for other SSA countries (excluding central African CFA countries).
  - Senegal: private credit = 21 percent of GDP in 2006.
  - Guinea-Bissau: private credit = 3.5 percent of GDP in 2006.
  - Côte d’Ivoire: private credit fell from 37 percent of GDP in 1990 to 13 percent in 2006.
  - Trade sector receives about 40 percent of credit while contributing less than 20 percent of GDP; agriculture has the lowest share of credit despite the highest share in GDP.
- Key constraints on supply of long-term loans:
  - Liquidity and maturity-regulation constraints: regulation requires at least 75 percent of a bank’s short-term liabilities be covered by short-term assets, and 75 percent of medium- and long-term bank assets be financed by resources of comparable maturity.
    - In 2007, 69 percent of banks complied with the liquidity ratio; 56 percent complied with the transformation ratio.
  - Limited access to international capital markets and a tiny long-term government bond market.
  - Weak legal enforcement and creditor rights; weak implementation of OHADA legislation in many countries.
  - Information gaps: about 3.5 percent of the adult population in the WAEMU is covered by a credit registry versus 7 percent in the rest of SSA.
  - Macroeconomic and political instability (inflation episodes during the 1994 CFA franc devaluation and during international food and fuel price hikes) discourage long-term contracting.
  - Excess short-term liquidity reduces banks’ incentives to mobilize long-term resources; statutory reserves are not remunerated in the WAEMU.
  - Human capital and bank expertise constraints: underdeveloped professional credit risk management on banks’ supply side and low-skilled entrepreneurs on the demand side.
- Institutional and financial-system features:
  - Financial development is shallow and banking-dominated; high concentration, excess liquidity (though easing with treasury bill and government bond market expansion), high NPL ratios and high overheads, yet comfortable net interest margins.
  - Long-term credit shares fell after the 1994 devaluation; sectoral shifts included increased long-term credit to export-oriented agriculture and services, decline for trade, and a flat or falling share for industry.

### IV. Empirical approach, data, and estimation methods
- Data and model:
  - Panel data for WAEMU countries, annual observations for 1995–2006 (postreform period).
  - Financial development decomposed into short-term credit/GDP (F_short) and long-term credit/GDP (F_long).
  - Standard growth model: real GDP growth depends on financial development, initial GDP per capita, and controls (education, inflation, government expenditure/GDP, trade openness).
  - Determinants of the share of long-term credit (S) modeled with explanatory variables including GDP per capita (log), share of time and saving deposits in total deposits, inflation, human capital, bank liquidity, profitability, and institutional indicators.
- Estimation methods:
  - Fixed-effects estimator, System GMM (Blundell and Bond), and Fully Modified OLS (FMOLS) for panel cointegration.
  - Panel unit root tests: Im-Pesaran-Shin (2003) and Levin-Lin-Chu (2002); many series consistent with I(1).
  - Panel cointegration: Pedroni (1999); six of seven statistics indicate cointegration among real GDP growth rate, financial development indicators, initial GDP per capita, and inflation — supporting use of FMOLS.

### V. Main empirical results (preserve reported coefficients and statistics)
- Fixed-effects (Table 1, column 4):
  - Initial GDP per capita (log) = -0.464 (4.85)***.
  - Short-term credit/GDP = 0.590 (1.82)*.
  - Long-term credit/GDP = 1.325 (2.20)**.
  - Observations = 87; Number of countries = 8; R-squared = 0.39.
- System GMM (Table 2, column 8):
  - Initial GDP per capita (log) = -0.386 (2.62)***.
  - Short-term credit/GDP (insignificant in many specifications).
  - Long-term credit/GDP = 2.977 (2.00)**.
  - Observations = 87; Number of countries = 8; Hansen test (prob.) = 0.16; AR2 (prob.) = 0.48.
- FMOLS panel cointegration (Table 3, column 4):
  - Initial GDP per capita (log) = -0.643 (15.34)***.
  - Short-term credit/GDP = 0.354 (1.00).
  - Long-term credit/GDP = 1.682 (13.50)***.
  - Observations = 94; Number of countries = 8.
- Cross-estimator conclusion:
  - Financial development is positively associated with economic growth.
  - The contribution of long-term credit to growth is substantially larger and more robustly significant than that of short-term credit.

### VI. Determinants of the share of long-term credit (S)
- Positive and significant determinants:
  - Share of time deposits in total deposits:
    - 0.633 (2.28)** in column 1.
    - 0.568 (2.06)** in column 2.
    - 0.734 (2.51)** in column 6.
    - (Time deposits include savings deposits per table note.)
  - Institutional quality and stability:
    - Bureaucracy quality: 0.035 (2.46)** in column 7.
    - Political stability: 0.036 (1.90)* in column 8.
    - Freedom from corruption: 0.003 (2.74)*** in column 9.
  - Bank profitability:
    - Return on assets: 0.017 (1.91)* in column 11.
- Inflation and nonlinearity:
  - Evidence of an inverted U-shaped relationship between inflation and the share of long-term credit: inflation coefficient positive and inflation squared negative.
  - Long-term credit share tends to drop when inflation exceeds a threshold of 4 percent.
  - Threshold standard deviation = 0.007 implying a 95 percent confidence interval for the threshold between 2.8 percent and 5.6 percent.
- Bank liquidity and other factors:
  - Bank liquidity coefficient negative as expected but not statistically significant when controlling for mobilization of long-term resources.
  - Model explanatory power: R-squared up to 0.32 (column 11), implying other unobserved factors (access to international capital markets, size of long-term government bond market, information asymmetry) also matter.

### VII. Policy-relevant implications and recommendations
- Core policy objective:
  - Increase the supply and effective provision of long-term credit to finance investments that raise productive capacity, support SMEs, and amplify the growth benefits of financial development.
- Mobilize long-term resources:
  - Increase the share of time and savings deposits and develop markets/instruments to provide banks with stable long-term funding.
  - Consider opening up the banking system to foreign participation to support banks’ mobilization of long-term resources.
  - Develop a vibrant regional capital market and expand long-term government bond markets to provide yield curve information and avenues for banks to match maturities.
- Legal, institutional, and information reforms:
  - Strengthen legal protections for creditors and effectively apply the laws; improve implementation of OHADA where weak.
  - Improve judicial systems and enforcement to strengthen creditor rights and collateral enforcement.
  - Improve credit information systems: expand credit registry coverage and record both positive and negative information to reduce information asymmetry.
- Regulatory and supervisory adjustments:
  - Align prudential norms better with best practices and better enforce prudential norms (IMF 2008 WAEMU Financial System Stability Assessment recommendations).
  - Loosen regulation on the transformation ratio by taking into account demand deposits in the calculation of the transformation ratio to improve compliance and increase banks’ ability to extend long-term loans, while maintaining supervision to prevent excessive risk-taking.
  - Address reserve remuneration and financial market development to reduce costly excess short-term liquidity disincentives.
- Capacity building and firm-level measures:
  - Improve professional credit risk management in banks and foster entrepreneur capacity to design bankable projects.
  - Promote nonbanking institutions (leasing companies, venture capital funds) to ease constraints on long-term financing for firms.
- Macroeconomic policy:
  - Preserve macroeconomic and political stability; keep inflation low and stable to encourage long-term contracting (WAEMU peg to the euro cited as a stabilizing factor).

*IMF working paper content as provided in the supplied PDF excerpt.*

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

### _wp10115 - References ................................................................................................

### I. Introduction
- Research context:
  - Since King and Levine (1993), more than 200 research papers have considered the finance–growth nexus, about half published in academic journals.
  - Mainstream view: banks stimulate growth by improving capital allocation, reducing transaction costs, and monitoring projects.
- Gaps in the literature:
  - Past studies commonly used monetary aggregates or credit to the private sector to measure financial development.
  - Less attention has been paid to the type of banking system credit (short-, medium-, and long-term) and its suitability to private sector needs.
  - Mismatch between credit supply and demand (especially scarcity of medium- and long-term credit) can constrain high-return investments.
- Relevance to WAEMU:
  - Issue is acute in WAEMU where banks either lack access to long-term resources or avoid long-term commitments because contract enforcement is weak.
  - WAEMU is a relatively homogenous group: eight countries (Benin, Burkina Faso, Côte d’Ivoire, Guinea-Bissau, Mali, Niger, Senegal, and Togo) share the same currency, a common central bank and bank regulator (the Banque Centrale des Etats de l’Afrique de l’Ouest (BCEAO) and the Regional Banking Commission), and a similar legal environment.
  - Despite homogeneity, disparities in financial development and distribution of private credit between short- and long-term are noticeable.
- Objectives of the paper:
  - Assess whether more financially developed WAEMU countries tend to grow faster than others.
  - Analyze whether banking systems that provide more short-term credit than long-term undermine the contribution of financial development to economic growth, and why.

### II. Theoretical background and mechanisms
- Core theoretical mechanisms (Levine (1997) summarized):
  - Financial sector mobilizes financial resources and channels funds to highest-return projects, fostering technological innovation and capital accumulation.
  - Banks offer households safe and liquid deposit opportunities and provide appropriate mixes of liquid and illiquid investments, facilitating long-run investments.
  - Financial intermediaries lower transaction and information costs and stand in for savers to ensure capital is put to best use.
  - Banks ease asymmetric information problems by monitoring corporate managers, designing financial contracts and collateral mechanisms to lower monitoring and enforcement costs.
- Role of credit maturity structure:
  - What matters is not just the volume of credit to the private sector but also the type/maturity of credit.
  - Scarcity of long-term credit makes it harder for firms—especially small and medium enterprises (SME)—to use external sources to finance investments that increase production capacity.
  - If firms rely on short-term credit for long-term projects, risk of project delay or halt increases if credit lines are not renewed.
- Distributional implications:
  - Large firms can raise funds abroad or self-finance, while SMEs—constituting the backbone of the productive sector—are hit hardest by the lack of long-term credit because they have fewer financial resources and undertake riskier innovative projects.

### III. Financial development in WAEMU: stylized facts from the paper
- Aggregate indicators:
  - Financial development in the WAEMU is low, though comparable with other sub-Saharan African (SSA) countries.
  - Average private credit to GDP stood at 12 percent for WAEMU countries over 1995–2006.
- Maturity composition of credit:
  - Short-term credit accounts for the bulk of bank financing to the private sector, averaging 70 percent of private credit.
  - Country ranges: 60 percent in Togo and Senegal to nearly 90 percent in Guinea-Bissau.
- Trends and distributions analyzed in the paper (tables and figures listed in source):
  - Tables: impact of short- and long-term credit on economic growth (various specifications including System GMM and panel cointegration), determinants of the share of long-term credit.
  - Figures and annexes cover financial development and economic growth (1989–2006), share of short-term credit and growth (1995–2006), financial soundness indicators (1995–2006), trend in private credit over GDP (1989–2006), short- and long-term credit shares (1989–2006), and sectoral shares in GDP and credit to private sector.
  - Annexes include panel unit root and cointegration tests, summary statistics and correlation matrices, and variable definition and sources.

### IV. Empirical approach and intended analysis (as described)
- Empirical strategy elements described in the source:
  - Model to assess whether financial development correlates with faster growth across WAEMU countries.
  - Explicit attention to composition of bank credit by maturity (short- vs. long-term).
  - Use of panel estimation methods including System GMM and panel cointegration approaches (as indicated by table titles).
  - Investigation of determinants of the share of long-term credit in total credit.

### V. Policy-relevant implications highlighted
- Importance of long-term bank financing:
  - Long-term credit is critical to finance investments that increase productive capacity and support SMEs engaged in innovative, higher-risk projects.
- Constraints identified that undermine provision of long-term credit in WAEMU:
  - Weak contract enforcement.
  - Banks’ limited access to long-term resources or reluctance to take long-term commitments.
- Anticipated policy focus (as reflected in paper’s structure and motivation):
  - Policies to improve contract enforcement and strengthen institutions to reduce banks’ reluctance to provide long-term credit.
  - Measures to enhance access to long-term funding for banks so they can match the maturity of credit with firms’ investment needs.
  - Targeted support to alleviate SMEs’ financing constraints, given their limited access to external finance relative to large firms.

*Source: Content from _wp10115 - References ................................................................................................*

### 15.6 percent for other SSA countries (excluding central African CFA countries).

### _wp10115 - 15.6 percent for other SSA countries (excluding central African CFA countries).

### Short versus long-term credit: roles and theoretical rationale
- Short-term credit defined as bank loans with maturity of less than a year; includes commercial paper, lines of credit, overdrafts; finances working capital and short-lived investments.
- Long-term credit defined as loans repayable beyond a year without rollover provisions; suitable for high-return projects requiring continuing capital commitment (new plant, R&D, technology adoption).
- Theoretical expectations:
  - Long-term credit likely has a larger impact on productivity growth because it finances higher-return, longer-maturity investments.
  - Constraints on long-term borrowing increase growth volatility and reduce long-term growth because investment becomes more sensitive to cash flow shocks.
  - SMEs are disproportionately affected by constraints on long-term financing.

### Constraints on supply of long-term loans in the WAEMU
- Key factors constraining long-term lending:
  - Liquidity risk and lack of long-term resources: regulation requires at least 75 percent of a bank’s short-term liabilities be covered by short-term assets, and 75 percent of medium- and long-term bank assets be financed by resources of comparable maturity.
    - In 2007, 69 percent of banks complied with the liquidity ratio; 56 percent complied with the transformation ratio.
  - Limited access to international capital markets and a tiny long-term government bond market.
  - Weak legal enforcement and creditor rights; weak implementation of OHADA legislation in many countries.
  - Information asymmetry and limited credit information: World Bank estimate that about 3.5 percent of the adult population in the WAEMU is covered by a credit registry versus 7 percent in the rest of SSA.
    - Recording both positive and negative information improves default prediction and extends credit.
  - Macroeconomic and political instability discourage long-term contracting; inflation episodes occurred during the 1994 CFA franc devaluation and during international food and fuel price hikes.
  - Excess short-term liquidity reduces banks’ incentives to mobilize long-term resources; statutory reserves are not remunerated in the WAEMU.
  - Human capital and bank expertise constraints: underdeveloped professional credit risk management on banks’ supply side and low-skilled entrepreneurs on the demand side.

### Stylized facts for WAEMU financial systems
- Financial development is shallow and banking-dominated; high concentration, excess liquidity (though easing with treasury bill and government bond market expansion), high NPL ratios and high overheads, yet comfortable net interest margins.
- Country-level private credit to GDP (examples):
  - Senegal: private credit = 21 percent of GDP in 2006.
  - Guinea-Bissau: private credit = 3.5 percent of GDP in 2006.
  - Côte d’Ivoire: private credit fell from 37 percent of GDP in 1990 to 13 percent in 2006.
- Credit composition and sectoral allocation:
  - Credit is mostly short-term; in most WAEMU countries the share of long-term credit in total credit is low and declining.
  - Guinea-Bissau: short-term credit accounts for over 90 percent of total credit.
  - Trade sector receives about 40 percent of credit while contributing less than 20 percent of GDP; agriculture has the lowest share of credit despite the highest share in GDP.
  - Long-term credit shares fell after the 1994 devaluation; sectoral shifts included increased long-term credit to export-oriented agriculture and services, decline for trade, and a flat or falling share for industry.

### Empirical estimations: models, data, and main results
- Data and model:
  - Panel data for WAEMU countries, annual observations for 1995–2006.
  - Financial development indicator decomposed into short-term credit/GDP (F_short) and long-term credit/GDP (F_long).
  - Standard growth model: real GDP growth depends on financial development, initial GDP per capita, and controls (education, inflation, government expenditure/GDP, trade openness).
  - Also estimated determinants of the share of long-term credit in total credit (S) with explanatory variables including GDP per capita (log), share of time and saving deposits in total deposits, inflation, human capital, bank liquidity, profitability, and institutional indicators.
- Estimation methods used:
  - Fixed-effects estimator, System GMM (Blundell and Bond), and Fully Modified OLS (FMOLS) for panel cointegration.

- Key empirical findings (preserve coefficients and significance levels as reported):
  - Fixed-effects (Table 1, column 4):
    - Initial GDP per capita (log) = -0.464 (4.85)***.
    - Short-term credit/GDP = 0.590 (1.82)*.
    - Long-term credit/GDP = 1.325 (2.20)**.
    - Observations = 87; Number of countries = 8; R-squared = 0.39.
  - System GMM (Table 2, column 8):
    - Initial GDP per capita (log) = -0.386 (2.62)***.
    - Short-term credit/GDP (insignificant in many specifications).
    - Long-term credit/GDP = 2.977 (2.00)**.
    - Observations = 87; Number of countries = 8; Hansen test (prob.) = 0.16; AR2 (prob.) = 0.48.
  - FMOLS panel cointegration (Table 3, column 4):
    - Initial GDP per capita (log) = -0.643 (15.34)***.
    - Short-term credit/GDP = 0.354 (1.00).
    - Long-term credit/GDP = 1.682 (13.50)***.
    - Observations = 94; Number of countries = 8.
  - Across estimators: financial development positively associated with economic growth, and the contribution of long-term credit to growth is substantially larger than that of short-term credit.

### Determinants of the share of long-term credit in total credit
- Main results from fixed-effects regressions (Table 4):
  - Share of time deposits in total deposits strongly positively correlated with the share of long-term credit (example coefficients and significance):
    - 0.633 (2.28)** in column 1.
    - 0.568 (2.06)** in column 2.
    - 0.734 (2.51)** in column 6.
    - Time deposits include savings deposits (note in table).
  - Evidence of an inverted U-shaped relationship between inflation and the share of long-term credit:
    - Inflation coefficient positive and inflation squared negative; long-term credit share tends to drop when inflation exceeds a threshold of 4 percent.
    - Threshold standard deviation = 0.007 implying a 95 percent confidence interval for the threshold between 2.8 percent and 5.6 percent.
  - Institutional and stability variables:
    - Bureaucracy quality positive and significant (example: 0.035 (2.46)** in column 7).
    - Political stability positive and significant (example: 0.036 (1.90)* in column 8).
    - Freedom from corruption positive and significant (example: 0.003 (2.74)*** in column 9).
  - Bank health:
    - Return on assets positive and significant: 0.017 (1.91)* in column 11.
  - Bank liquidity coefficient negative as expected but not statistically significant in specifications controlling for mobilization of long-term resources.
  - Model explanatory power: R-squared up to 0.32 (column 11), suggesting other unobserved factors (access to international capital markets, size of the long-term government bond market, information asymmetry) also matter.

### Policy-relevant implications and recommendations (drawn from analysis)
- Mobilize long-term resources:
  - Increase the share of time and savings deposits and develop markets/instruments to provide banks with stable long-term funding.
- Strengthen legal and judicial enforcement:
  - Improve implementation and funding of judicial systems to strengthen creditor rights and collateral enforcement (weak OHADA implementation cited).
- Improve credit information systems:
  - Expand credit registry coverage and record both positive and negative information to reduce information asymmetry and support lending.
- Develop domestic long-term capital markets:
  - Expand long-term government bond markets to provide yield curve information and avenues for banks to match maturities.
- Reduce excess short-term liquidity disincentives:
  - Address reserve remuneration and financial market development to reduce costly excess liquidity and encourage long-term resource mobilization.
- Enhance bank risk assessment capacity and borrower human capital:
  - Improve professional credit risk management and entrepreneur capacity to design bankable projects.
- Preserve macroeconomic and political stability:
  - Keep inflation low and stable (WAEMU peg to the euro cited as a stabilizing factor) and foster political stability to encourage long-term commitments by banks.

*Italicized source attribution: IMF working paper content as provided in the supplied PDF excerpt.*

### CONCLUSION AND POLICY RECOMMENDATIONS

### CONCLUSION AND POLICY RECOMMENDATIONS

### Key empirical findings
- Study period: 1995–2006 (postreform period).
- Estimators applied: fixed-effects, System GMM, and FMOLS.
- Overall result: financial development generally supports growth in the WAEMU.
- Crucial distinction: long-term credit underpins the favorable impact of financial development on growth; short-term credit appears in most cases not to be correlated with growth.
- Structure of bank financing: short-term credit accounts for about 70 percent of bank financing in the WAEMU.
- Implication: with short-term credit dominating (about 70 percent), member countries are less able to reap the full benefit from improvements in their financial systems.

### Determinants of long-term credit provision
- Factors that encourage banks to provide long-term financing to firms:
  - macroeconomic stability
  - a creditor-friendly environment
  - political stability
  - availability of long-term financial resources

### Policy recommendations
- Continue to maintain a stable macroeconomic environment.
- Strengthen legal protections for creditors and effectively apply the laws.
- Implement recommendations of the 2007 WAEMU Financial System Stability Assessment (IMF, 2008), including:
  - align prudential norms better with best practices;
  - better enforce prudential norms;
  - improve the legal and judicial frameworks;
  - improve the quality of financial and accounting information.
- Loosen regulation on the transformation ratio by taking into account demand deposits in the calculation of the transformation ratio to:
  - improve compliance;
  - increase banks’ ability to extend long-term loans;
  - maintain adequate supervision and regulatory framework to prevent excessive risk-taking.
- Consider opening up the banking system to foreign participation to support bank mobilization of long-term resources.
- Foster firms’ access to long-term financing through other financial institutions:
  - develop a vibrant regional capital market;
  - promote nonbanking institutions, such as leasing companies and venture capital funds, to ease constraints on long-term financing for firms.

### Supporting methodological notes (concise)
- Panel unit root tests used: Im-Pesaran-Shin (2003) and Levin-Lin-Chu (2002); results mixed across specifications, many series consistent with I(1).
- Panel cointegration: Pedroni (1999) procedure; six of seven statistics indicate cointegration among real GDP growth rate, financial development indicators, initial GDP per capita, and inflation — supporting use of FMOLS for long-run relationships.

*Source: CONCLUSION AND POLICY RECOMMENDATIONS, _wp10115 - CONCLUSION AND POLICY RECOMMENDATIONS*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10115.pdf_