2005年-世界发展银行全球_Quantitative_Analysis_of_Crisis___Crisis_Identification_and_Causality_26页_2mb

Summary of "Quantitative Analysis of Crisis: Crisis Identification and Causality"

Core Content

This paper provides a quantitative analysis of different types of crises in emerging economies, focusing on crisis identification and causality. It addresses the lack of consensus in defining crises and the resulting inconsistencies in crisis identification across studies. The paper proposes a standardized approach to identify crisis periods and episodes, using quantitative indicators and thresholds to ensure objective and consistent analysis.

Main Types of Crises Analyzed

The paper examines seven types of crises:

1. Liquidity type banking crises
2. Solvency type banking crises
3. Balance of payments crises
4. Currency crises
5. Debt crises
6. Growth rate crises
7. Financial crises

These types are analyzed across 15 emerging economies between Q1 1980 and Q4 2002 on a quarterly basis.

Crisis Identification Method

To identify crisis periods and episodes, the paper uses standardized scores, which are calculated using the following formula:

$$
Z = \frac {X - \mu}{\mathrm {d}} \tag {1}
$$

Where:

• $X$ is the value of the indicator
• $\mu$ is the mean of the indicator
• $d$ is the standard deviation of the indicator

The standardized score measures how far the indicator deviates from its mean in terms of standard deviations. A threshold of 2.0 is used to identify crisis periods, as it corresponds to a 2.5 percent probability of exceeding the threshold, and is considered an average of currency crisis studies.

The degree of a crisis episode is calculated as the sum of the crisis depths and lengths, using the following formula:

$$
\sum_ {i = 1} ^ {m} \sum_ {t = 1} ^ {n} a _ {it} \tag {2}
$$

Where:

• $a_{it}$ is the crisis depth of $m$ different crisis types at period $t$
• $n$ is the length of the crisis episode

This method allows for the measurement of crisis severity based on both depth and duration.

Key Issues in Crisis Identification

• Conceptual vs. operational definitions: While there is general agreement on conceptual definitions, operational definitions vary widely. This leads to inconsistent crisis identifications even when using the same analytical framework.
• Multidimensionality: Recent crises are often multidimensional, with multiple types occurring simultaneously or in quick succession. Most studies, however, focus on only a few types, which may not fully capture the complexity of crises.
• Data limitations: Many studies use annual data, which may not be sufficient to identify crises that occur and end within the same year. Quarterly data is preferred for more accurate identification.

Operational Definitions and Indicators

Granger Causality Tests

Granger causality tests are conducted in five Asian countries: Indonesia, the Republic of Korea, Malaysia, the Philippines, and Thailand. The results indicate that:

• Currency crises tend to trigger other types of crises.
• This suggests that exchange rate management is a critical policy tool in preventing or mitigating broader financial crises.

Conclusion

The paper emphasizes the importance of using consistent and standardized definitions and indicators to identify crises. It highlights the need for a multidimensional approach to capture the complexity of modern crises and argues that currency crises often act as catalysts for other crises. The proposed methodology provides a framework for more accurate and comparable crisis identification across countries and time periods.