## _wp0855 - 1. Details on the Financial Development Sub-indexes

## Source details

**Canonical URL:** [_wp0855 - 1. Details on the Financial Development Sub-indexes](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp0855.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp0855.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp0855.pdf.json)

---

### Introduction and motivation
- Credit growth accelerated globally; nominal credit growth to the private sector expanded by 32 percent on average in the seven largest Latin American countries in 2006.
- Chile: private credit increased by 18 percent in 2006; household debt increases described as moderate (poorest borrowers saw larger increases but represent only a small fraction of total debt).
- Banking-sector indicators in Chile: NPLs remain in the low single digits; NPLs are fully covered by provisions; banks enjoy strong capitalization.
- Non-bank lenders (department stores, credit cards) expanding market share; they do not take deposits and possess borrower information that has been used prudently.
- Research objective: explore whether financial development (reduction of financial frictions) in Chile and other countries can mitigate economic volatility and strengthen resilience to domestic and external shocks.
- Empirical approach: analyze determinants of credit cyclicality within a VAR framework for a sample of 18 emerging and advanced economies comparable to Chile; construct a financial development index and study its effect on impulse response amplitudes to shocks to output, terms of trade, global liquidity, and global risk appetite; use country-specific VARs and panel VARs.

### Theoretical premise: financial frictions and the financial accelerator
- Framework: financial accelerator — financial frictions amplify the business cycle via collateral/value of balance sheets affecting borrowing and thereby feeding back into economic activity.
- Key friction channels: weak creditor rights, poor credit information sharing, underdeveloped capital markets, over-indebted borrowers.
- Reducing financial frictions could make default less probable/costly and allow better risk-sharing via deeper capital markets, thereby reducing procyclicality.
- Alternative mechanism: credit constraints on demand side (Aghion and others, 2005) can increase procyclicality and volatility; motivates empirical testing of whether reducing frictions lowers volatility.

### Measuring financial friction: Financial development index construction
- Index structure:
  - Eight variables used to construct four sub-indexes capturing: creditor rights, size of the credit market, competition, and capital market development.
  - Each variable normalized on the “best” performer within the sample; variables range between 0 and 1; some variables inverted so higher = better.
  - Each of the four sub-indexes is the simple average of underlying normalized variables (credit market size and competition are only based on one variable).
  - Overall index is the simple average of the four sub-indexes, giving equal weight to all measured aspects.
- Sub-index grouping:
  - Creditor rights index built on survey-based indicators: cost/time to enforce debt, legal powers on collateral/bankruptcy.
  - “Arm’s length lending” index = average of the remaining three sub-indexes (credit market size, competition, capital market development).
- Country-level notes on Chile:
  - Creditor rights index is highest among Latin American countries in sample but low compared to the rest of the sample.
  - Legal enforcement of debt contracts is relatively slow and costly per World Bank Doing Business indicators.
  - Chile ranks near the top of emerging markets on credit liquidity and ahead of some industrialized countries on competition, but near the bottom on financial market development (derivatives markets, securitization, non-traditional banking).
  - Overall, Chile ranks higher than other Latin American countries in the sample; on par with South Africa and many Eastern European countries (except Hungary); still behind industrialized countries.

### Cross-country analysis of impulse responses (household credit focus)
- VAR specification (country-specific):
  - 6-equation VAR in Cholesky order: (i) total M1 in U.S. dollars, euro, and yen (proxy for global liquidity, expressed in dollars); (ii) VIX (global risk appetite); (iii) terms of trade; (iv) industrial production index (IPI); (v) real lending rate; (vi) real credit.
  - Data frequency: monthly; all variables seasonally adjusted (Census X11) and expressed as deviations from an HP trend except VIX and lending rate (unadjusted, in levels).
  - Lag length: two lags used for all countries (Schwarz criterion suggested one or two lags).
  - Impulse responses normalized to a one percent shock for cross-country comparability; analysis focused on 6-month and 12-month horizons.
- Robustness checks:
  - Core VAR with fewer variables produced similar responses.
  - Exchange rate excluded for parsimony.
  - Cholesky ordering relatively unimportant.
  - Different sample periods produced moderate changes; Australia sensitive to period choice.
- Stylized empirical patterns:
  - 6-month household credit impulse responses to output shocks plotted against overall financial index show a downward-sloping trend: countries with lower financial development scores tend to have larger positive credit responses to output shocks.
  - Subsample excluding outliers (Argentina, Australia, Brazil, Hungary, Mexico): downward slope remains for the creditor rights sub-index but disappears for the arm’s length sub-index.
  - Impulse responses to terms-of-trade and VIX shocks show some relation with financial development; no evidence that financial development influences the response of credit to global liquidity shocks.
- Cross-country regressions:
  - Regress country-specific 6-month impulse responses on financial index and controls; account for differing standard errors of impulse responses by bootstrapping 500 samples of impulse responses and repeating regressions 500 times.
  - Key results:
    - Financial index is statistically significant with expected (negative) sign even after controlling for GDP per capita.
    - Both creditor rights and arm’s length sub-indexes are significant with expected sign when all countries included.
    - Household debt/GDP point estimate indicates higher debt is associated with greater procyclicality, but not statistically significant (household debt data available for only 12 countries).
    - Number of items on financial statements (proxy for creditor information; available for 14 countries) is significant with expected sign.
  - Regression notes:
    - Bootstrapped confidence intervals estimated using Monte Carlo simulations.
    - Numbers in bold in the source indicate statistical significance within two standard deviations.

### Estimating panel VARs (top half vs bottom half of financial index)
- Purpose: increase degrees of freedom by splitting sample into top half and bottom half of the financial index and estimating panel VARs with same specification as country VARs.
- Normalization: impulse responses normalized to a one percent shock; 2 standard deviation confidence intervals displayed.
- Findings — household credit:
  - Top half of the financial index shows reduced procyclicality of household credit relative to output and terms-of-trade shocks versus bottom half.
  - Differences between top and bottom half are statistically significant for household credit.
  - No significant impact found for VIX or global liquidity in panel VARs for household credit.
- Findings — corporate credit:
  - Higher financial development implies less procyclical corporate credit response to output shocks, but not to terms-of-trade shocks.
- Robustness to income correlation:
  - Financial development correlates with GDP per capita; authors regress financial development index on GDP per capita and split sample between over-performers and under-performers relative to income; key results remain unchanged.

### Variance decomposition: drivers of recent household credit volatility (2002–present)
- Method: country-specific VARs used to decompose variance of credit developments since 2002 into contributions from VAR variables.
- General performance: VAR system performs well; own (direct) shocks to credit account for a reasonably small share for most countries (less satisfactory for Mexico, Poland, Czech Republic).
- Chile — contributions to household credit volatility (narrative percentages from source):
  - Global liquidity and the VIX account for about half of the volatility in household credit in the current cycle.
  - Domestic interest rate shocks contributed about one-quarter of total household credit variability.
  - Domestic output contributes only a couple of percent to total credit variations.
  - Terms-of-trade shocks have contributed little to Chilean credit volatility despite copper’s importance (partly due to mining sector being an enclave and countercyclical fiscal policies).
- Comparative observations:
  - Eastern European countries’ credit cycles appear strongly influenced by global liquidity.
  - Countries most affected by VIX changes tend to have high current account deficits or be at ends of carry trades.

### Conclusions, policy implications, and caveats
- Main conclusion: financial development matters for credit cyclicality — stronger creditor rights and greater arm’s length lending are associated with reduced credit procyclicality; evidence strongest for creditor rights.
- Household credit results are more robust than corporate credit results (consistent with IMF (2006)).
- Policy implications suggested:
  - Strengthen creditor rights to reduce credit-induced volatility.
  - Enhance credit information sharing (evidence suggests credit information can reduce procyclicality, but data limitations exist).
  - Develop deeper markets in securitized products and create a credit derivatives market to improve risk sharing and reduce credit-induced volatility.
  - Consider policies that promote countercyclical credit behavior; in some frameworks (permanent income models) countercyclical credit may be optimal.
- Caveats:
  - Financial frictions are not the only determinants of credit cyclicality: public ownership of banks, foreign vs domestic ownership, exchange rate regime, degree of dollarization, dependence on commodity exports, access to global financial markets, and confidence/permanent income shocks also matter.
  - Data limitations affect measurement (e.g., lack of a consistent high-frequency indicator for credit information across countries).

### Appendix 1 — Raw data (country-level indicators)
- Column headings: Country Leg_rights Days Procedures Cost   Cred_GDP Spread Fee_inc Stock mkt turn.
- Data rows (preserve formatting of numeric values):
  - ARGENTINA 3 5 20 33 15.00 0.16 6.43 4.60 0.07
  - AUSTRALIA 9 15 7 19 12.80 0.91 5.32 8.80 0.76
  - BRAZIL 2 5 6 64 21.50 0.32 40.32 5.10 0.34
  - CANADA 7 3 4 6 17 12.01 1.05 3.62 6.70 0.63
  - CHILE 4 3 0 5 33 16.30 0.64 3.92 1.40 0.10
  - CZECH REP. 6 3 0 0 21 14.10 0.37 4.52 5.30 0.52
  - DENMARK 8 8 3 15 6.51 1.38 5.01 1.30 0.67
  - HUNGARY 6 3 6 52 19.60 0.36 2.32 6.80 0.55
  - KOREA 6 7 5 29 5.50 0.87 1.72 7.02 2.34
  - LATVIA 8 1 8 92 11.80 0.28 4.52 8.70 0.16
  - MEXICO 2 4 2 13 7 20.00 0.15 5.52 2.00 0.21
  - NORWAY 6 8 7 14 9.00 0.71 2.31 9.70 0.86
  - NEW ZEALAND 9 5 0 28 10.91 1.13 4.61 7.90 0.38
  - POLAND 4 10 0 04 110.00 0.27 4.93 3.30 0.26
  - SOUTH AFRICA 5 2 7 72 6 11.50 0.63 4.75 5.40 0.45
  - SWITZERLAND 6 1 7 02 21 11.01 1.60 2.54 1.50 0.89
  - SWEDEN 6 2 0 8 19 5.90 0.92 3.21 2.51 1.12
  - TAIWAN 4 2 10 28 16.60 0.98 2.94 1.91 1.84
- Summary statistics:
  - Mean 5.629 6.125 5.91 11.90 0.76 0.02 5.00 0.7
  - Stdv 2.1230 0.48 8.63 3.90 0.48 0.79 7.70 0.6

### Cross correlations (matrix)
- Variables: LEG_RIGHTS, DAYS, PROCEDURES, COST, CRED_GDP, SPREAD, FEE_INC, STOCK_MKT_TU
  - LEG_RIGHTS: 1.00 -0.62 -0.75 -0.57 0.55 -0.43 -0.20 0.15
  - DAYS: -0.62 1.00 0.69 0.31 -0.62 0.34 0.27 -0.45
  - PROCEDURES: -0.75 0.69 1.00 0.51 -0.53 0.51 0.23 -0.24
  - COST: -0.57 0.31 0.51 1.00 -0.44 0.29 0.32 -0.42
  - CRED_GDP: 0.55 -0.62 -0.53 -0.44 1.00 -0.28 -0.03 0.47
  - SPREAD: -0.43 0.34 0.51 0.29 -0.28 1.00 -0.02 -0.24
  - FEE_INC: -0.20 0.27 0.23 0.32 -0.03 -0.02 1.00 0.19
  - STOCK_MKT_TU: 0.15 -0.45 -0.24 -0.42 0.47 -0.24 0.19 1.00

### Other data sources (as listed)
- Credit data from national authorities and IMF desks
- Global liquidity (US/Japan/Euro area M1): Haver
- Terms of trade: Haver and IFS
- CPI (deflator for real interest rates, real credit): Haver, IFS
- Industrial production index: Haver
- Lending interest rate: IFS

*Source: _wp0855 - Details on the Financial Development Sub-indexes — https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp0855.pdf*

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

### _wp0855 - References

### Table
- 1.  Cross-Country Regressions .........................................................................................13

### Figures
- 1. Credit to the Private Sector  ...........................................................................................3
- 2. Household debt Indicators  ............................................................................................3
- 3. Correlation, Real GDP Growth .....................................................................................4
- 4. Financial Development Index  .......................................................................................7
- 5. Overall Financial Development Index  ..........................................................................8
- 6. Household Credit Impulse Responses to Output Shocks, Full Sample  ......................11
- 7. Household Credit Impulse Responses to Output Shocks, Subsample  ........................11
- 8. Household Credit Impulse Responses, Full Sample  ...................................................12
- 9. Household Cridt Impulse Responses, Panel VAR  ......................................................14
- 10. Corporate Credit Impulse Responses, Panel VAR  .....................................................15
- 11. Variance Decomposition – Explaining Credit Volatility since 2002 ..........................16

### Appendix
- Appendix

*Source: _wp0855 - References — https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp0855.pdf*

### 1. Details on the Financial Development Sub-indexes ....................................................20

### REFERENCES

### _wp0855 - REFERENCES

### References
- Aghion, P., G-M, Angeletos, A. Banerjee, and K. Manova, (2004), “Volatility and Growth: Credit Constraints and Productivity-Enhancing Investment”, mimeo, available at: http://www.economics.harvard.edu/faculty/aghion/papers.html
- Bernanke, B., M. Gertler, and S. Gilchrist (1996), “The Financial Accelerator and the Flight to Quality”, The Review of Economics and Statistics, Vol. 78, No. 1.
- Braun, M. and B. Larrain (2005), “Finance and the Business Cycle: International, Inter-Industry Evidence”, The Journal of Finance”, Vol. LX. No. 3.
- Nicoló, Luc Laeven, and Kenichi Ueda, 2006, “Corporate Governance Quality in Asia: Comparative Trends and Impact” (unpublished).
- Djankov, S., C. McLiesh, and A. Shleifer (2005), “Private Credit in 129 Countries”, NBER Working Paper 11078.
- Fuentes, R., and C. Maquieira (2001), “Why Borrowers Repay: Understanding High Performance in Chile’s Financial Market”, in Defusing Default, M. Pagano (ed.), IADB, Washington DC.
- Galindo, A. and A. Micco (2001), “Credito Protection and Financial Cycles”, Inter-American Development Bank Working Paper No. 443.
- IMF (2006), World Economic Outlook, Chapter IV, September, Washington DC.
- Kiyotaki, N. and J. Moore (1997), “Credit Cycles”, Journal of Political Economy, Vol. 105, No. 2.
- Love, I., and L. Ziccino (2006) “Financial Development and Dynamic Investment Behaviour: evidence from Panel VAR”, The Quarterly Review of Economics and Finance, Vol. 46.
- Micco, A., and U. Panizza (2004), “Bank Ownership and Lending Behavior”, Inter-American Development Bank Working Paper No. 520.
- Miniane, J., and J. Rogers (2003), Capital Controls and the International Transmission of U.S. Money Shocks”, Board of Governors of the Federal Reserve System, International Finance Discussion Paper No. 778.

### Appendix 1. Details on the financial development sub-indexes
- 1. Creditor rights
  - Legal rights index (World Bank, Doing Business Database)
  - Number of days to enforce debt contract in court (Djankov, McLiesh, and Shleifer, 2005).
  - Number of procedures for lawsuit (World Bank, Doing Business Database)
  - Cost of legal procedure/debt value (World Bank, Doing Business Database)
- 2. Credit market liquidity
  - Private credit/GDP (IFS, average 2000-06)
- 3. Competition
  - Spread between borrowing and lending interest rates (IFS, average 2000-06)
- 4. Financial markets development
  - Banks’ non-interest income/total income (Bankscope, 2003)
  - Stock market turnover/stock market capitalization (World Bank, A new database on financial development and structure, 2003)

### Raw data (country-level indicators)
- Column headings: Country Leg_rights Days Procedures Cost   Cred_GDP Spread Fee_inc Stock mkt turn.
- Data rows (preserve formatting of numeric values):
  - ARGENTINA 3 5 20 33 15.00 0.16 6.43 4.60 0.07
  - AUSTRALIA 9 15 7 19 12.80 0.91 5.32 8.80 0.76
  - BRAZIL 2 5 6 64 21.50 0.32 40.32 5.10 0.34
  - CANADA 7 3 4 6 17 12.01 1.05 3.62 6.70 0.63
  - CHILE 4 3 0 5 33 16.30 0.64 3.92 1.40 0.10
  - CZECH REP. 6 3 0 0 21 14.10 0.37 4.52 5.30 0.52
  - DENMARK 8 8 3 15 6.51 1.38 5.01 1.30 0.67
  - HUNGARY 6 3 6 52 19.60 0.36 2.32 6.80 0.55
  - KOREA 6 7 5 29 5.50 0.87 1.72 7.02 2.34
  - LATVIA 8 1 8 92 11.80 0.28 4.52 8.70 0.16
  - MEXICO 2 4 2 13 7 20.00 0.15 5.52 2.00 0.21
  - NORWAY 6 8 7 14 9.00 0.71 2.31 9.70 0.86
  - NEW ZEALAND 9 5 0 28 10.91 1.13 4.61 7.90 0.38
  - POLAND 4 10 0 04 110.00 0.27 4.93 3.30 0.26
  - SOUTH AFRICA 5 2 7 72 6 11.50 0.63 4.75 5.40 0.45
  - SWITZERLAND 6 1 7 02 21 11.01 1.60 2.54 1.50 0.89
  - SWEDEN 6 2 0 8 19 5.90 0.92 3.21 2.51 1.12
  - TAIWAN 4 2 10 28 16.60 0.98 2.94 1.91 1.84
- Summary statistics:
  - Mean 5.629 6.125 5.91 11.90 0.76 0.02 5.00 0.7
  - Stdv 2.1230 0.48 8.63 3.90 0.48 0.79 7.70 0.6

### Cross correlations
- Matrix (variables: LEG_RIGHTS, DAYS, PROCEDURES, COST, CRED_GDP, SPREAD, FEE_INC, STOCK_MKT_TU)
  - LEG_RIGHTS: 1.00 -0.62 -0.75 -0.57 0.55 -0.43 -0.20 0.15
  - DAYS: -0.62 1.00 0.69 0.31 -0.62 0.34 0.27 -0.45
  - PROCEDURES: -0.75 0.69 1.00 0.51 -0.53 0.51 0.23 -0.24
  - COST: -0.57 0.31 0.51 1.00 -0.44 0.29 0.32 -0.42
  - CRED_GDP: 0.55 -0.62 -0.53 -0.44 1.00 -0.28 -0.03 0.47
  - SPREAD: -0.43 0.34 0.51 0.29 -0.28 1.00 -0.02 -0.24
  - FEE_INC: -0.20 0.27 0.23 0.32 -0.03 -0.02 1.00 0.19
  - STOCK_MKT_TU: 0.15 -0.45 -0.24 -0.42 0.47 -0.24 0.19 1.00

### Other data sources
- Credit data from national authorities and IMF desks
- Global liquidity (US/Japan/Euro area M1): Haver
- Terms of trade: Haver and IFS
- CPI (deflator for real interest rates, real credit): Haver, IFS
- Industrial production index: Haver
- Lending interest rate: IFS

*Source: _wp0855 - REFERENCES*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp0855.pdf_