|Practice area:||Finance | Financial markets and Institutions|
|Client:||EC DG Economic & Financial Affairs|
|Keywords:||modelling quantitative analysis|
This study investigates the effectiveness of state guarantee schemes and ad hoc guarantees given to banks over 2008-10.
The study analyses the market value of state guarantees as reflected in the costs banks faced in raising funds on wholesale financial markets and whether state guarantees influenced the funding and (ultimately) lending behaviour of banks. The report presents a comprehensive ex-post evaluation of one of the main tools used to restore the functioning of wholesale financial markets after the Lehman bankruptcy and onset of the financial crisis.
The results suggest that state guarantees were successful in lowering the costs of bond issuance for participating banks while having relatively few distortionary impacts on the conditions faced by non-participating banks on wholesale financial markets. Impacts on bank behaviour, especially in terms of lending and leverage were less clear.
In detail, difference-in-differences / differences approaches were applied, with the main results scrutinised against a range of robustness tests. Given this background, the study made the following findings.
On the market value of state guarantees: The direct effect of state guarantees was limited to approximately 30 basis points ceteris paribus, i.e., investors were prepared to accept 30 basis points less for a comparable debt security issued by a bank with a state guarantee than without one. Moreover, there was a second, upgrade effect, insofar as lower-rated banks were able to issue bonds with a superior rating. This effect had a sizeable and statistically significant impact on the issuing costs of bonds rated BBB-/BBB/BBB+, in the order of over 200bps. In contrast to other research, there was no evidence of differences across government guarantors having had a significant impact on issuing costs.
On spill-over: While each individual state guaranteed bond issue benefited from a state guarantee through a lower issue cost, all bond issues (state guaranteed and non-state guaranteed) were impacted negatively by the overall volume of state guaranteed bond issues (i.e., spill-over). Quantitatively, each bond issue faced a higher issuance cost of 10.5bps, on average, as a result of the overall volume of state guaranteed issuance activity in a given month. However, this spill-over effect was entirely domestic.
On bank behaviour: The econometric analysis showed that banks that issued state guaranteed bonds: a) lent less; and b) reduced their leverage ratios by more than banks that were eligible to issue state guaranteed bonds but did not do so. These results are, however, less clear cut.