Comparing logit-based early warning systems: Does the duration of systemic banking crises matter?

Giovanni Caggiano; Pietro Calice; Leone Leonida; George Kapetanios

doi:10.1016/j.jempfin.2016.01.005

Comparing logit-based early warning systems: Does the duration of systemic banking crises matter?

Giovanni Caggiano, Pietro Calice, Leone Leonida, George Kapetanios

Research output: Contribution to journal › Article › peer-review

41 Citations (Scopus)

520 Downloads (Pure)

Abstract

This paper compares the performance of binomial and multinomial logit models in the context of building early warning systems (EWS) for systemic banking crises. We test the hypothesis that the predictive performance of binomial logit models is hampered by what we define as the crisis duration bias, arising from the decision to either treat crisis years after the onset of a crisis as non-crisis years or remove them altogether from the sample. In line with our hypothesis, results from a large sample of world economies suggest that i) the multinomial logit outperforms the binomial logit model in predicting systemic banking crises, and ii) the longer the average duration of the crisis in the sample, the larger the improvement.

Original language	English
Pages (from-to)	104-116
Journal	Journal of Empirical Finance
Volume	37
Issue number	0
Early online date	16 Feb 2016
DOIs	https://doi.org/10.1016/j.jempfin.2016.01.005
Publication status	Published - 1 Jun 2016

Keywords

Banking crises
Systemic risk
Early warning systems
Logit estimation
Financial regulation

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1016/j.jempfin.2016.01.005

Comparing logit-based early warning systems_CAGGIANO_Accepted5January2016_GREEN AAMAccepted author manuscript, 434 KBLicence: CC BY-NC-ND

Cite this

@article{35a201106ef046a59ed9bafc2ec33268,

title = "Comparing logit-based early warning systems: Does the duration of systemic banking crises matter?",

abstract = "This paper compares the performance of binomial and multinomial logit models in the context of building early warning systems (EWS) for systemic banking crises. We test the hypothesis that the predictive performance of binomial logit models is hampered by what we define as the crisis duration bias, arising from the decision to either treat crisis years after the onset of a crisis as non-crisis years or remove them altogether from the sample. In line with our hypothesis, results from a large sample of world economies suggest that i) the multinomial logit outperforms the binomial logit model in predicting systemic banking crises, and ii) the longer the average duration of the crisis in the sample, the larger the improvement.",

keywords = "Banking crises, Systemic risk, Early warning systems, Logit estimation, Financial regulation",

author = "Giovanni Caggiano and Pietro Calice and Leone Leonida and George Kapetanios",

year = "2016",

month = jun,

day = "1",

doi = "10.1016/j.jempfin.2016.01.005",

language = "English",

volume = "37",

pages = "104--116",

journal = "Journal of Empirical Finance",

issn = "0927-5398",

publisher = "Elsevier",

number = "0",

}

TY - JOUR

T1 - Comparing logit-based early warning systems

T2 - Does the duration of systemic banking crises matter?

AU - Caggiano, Giovanni

AU - Calice, Pietro

AU - Leonida, Leone

AU - Kapetanios, George

PY - 2016/6/1

Y1 - 2016/6/1

N2 - This paper compares the performance of binomial and multinomial logit models in the context of building early warning systems (EWS) for systemic banking crises. We test the hypothesis that the predictive performance of binomial logit models is hampered by what we define as the crisis duration bias, arising from the decision to either treat crisis years after the onset of a crisis as non-crisis years or remove them altogether from the sample. In line with our hypothesis, results from a large sample of world economies suggest that i) the multinomial logit outperforms the binomial logit model in predicting systemic banking crises, and ii) the longer the average duration of the crisis in the sample, the larger the improvement.

AB - This paper compares the performance of binomial and multinomial logit models in the context of building early warning systems (EWS) for systemic banking crises. We test the hypothesis that the predictive performance of binomial logit models is hampered by what we define as the crisis duration bias, arising from the decision to either treat crisis years after the onset of a crisis as non-crisis years or remove them altogether from the sample. In line with our hypothesis, results from a large sample of world economies suggest that i) the multinomial logit outperforms the binomial logit model in predicting systemic banking crises, and ii) the longer the average duration of the crisis in the sample, the larger the improvement.

KW - Banking crises

KW - Systemic risk

KW - Early warning systems

KW - Logit estimation

KW - Financial regulation

U2 - 10.1016/j.jempfin.2016.01.005

DO - 10.1016/j.jempfin.2016.01.005

M3 - Article

SN - 0927-5398

VL - 37

SP - 104

EP - 116

JO - Journal of Empirical Finance

JF - Journal of Empirical Finance

IS - 0

ER -

Comparing logit-based early warning systems: Does the duration of systemic banking crises matter?

Abstract

Keywords

UN SDGs

Access to Document

Fingerprint

Cite this