TY - JOUR
T1 - The real output losses associated with modern banking crises
AU - Boyd, John H.
AU - Kwak, Sungkyu
AU - Smith, Bruce
PY - 2005/12
Y1 - 2005/12
N2 - This study investigates the real output losses associated with modern banking crises. We find a remarkable diversity of experience. In a number of instances banking crises have not been associated with any significant reduction in the growth of real, per capita GDP. Often, this has been the case in mature and developed economies. On the other hand, estimated output losses are extremely large for some other countries - amounting to several years of lost GDP. Interestingly, such large losses can be associated with banking crises that were designated as non-systemic. Our sample mean and median output loss estimates are also big. For the average sample country, the estimated present discounted value of crisis-related output losses is bounded between 63% and 302% of real, per capita GDP in the last year before the crisis onset. Average loss estimates are this large primarily because we find evidence that post-crisis economic slowdowns often persist long after the crisis is officially over. Such delayed costs have been largely overlooked in previous empirical work and resultantly our loss estimates are much larger than those that have appeared elsewhere.
AB - This study investigates the real output losses associated with modern banking crises. We find a remarkable diversity of experience. In a number of instances banking crises have not been associated with any significant reduction in the growth of real, per capita GDP. Often, this has been the case in mature and developed economies. On the other hand, estimated output losses are extremely large for some other countries - amounting to several years of lost GDP. Interestingly, such large losses can be associated with banking crises that were designated as non-systemic. Our sample mean and median output loss estimates are also big. For the average sample country, the estimated present discounted value of crisis-related output losses is bounded between 63% and 302% of real, per capita GDP in the last year before the crisis onset. Average loss estimates are this large primarily because we find evidence that post-crisis economic slowdowns often persist long after the crisis is officially over. Such delayed costs have been largely overlooked in previous empirical work and resultantly our loss estimates are much larger than those that have appeared elsewhere.
KW - Banking crises
UR - http://www.scopus.com/inward/record.url?scp=30344441785&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=30344441785&partnerID=8YFLogxK
U2 - 10.1353/mcb.2006.0002
DO - 10.1353/mcb.2006.0002
M3 - Article
AN - SCOPUS:30344441785
SN - 0022-2879
VL - 37
SP - 977
EP - 999
JO - Journal of Money, Credit and Banking
JF - Journal of Money, Credit and Banking
IS - 6
ER -