CP10 (Full Document) - European Banking Authority
CP10 (Full Document) - European Banking Authority
CP10 (Full Document) - European Banking Authority
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
Pooled data sets<br />
for good predictive power (Annex VII, Part 4, Paragraph 31 (a)) is<br />
fulfilled. The institution should ensure, however, that deviations<br />
from the definitions provided in the CRD are not too great and<br />
that the definition of default used results in a ranking of obligors<br />
or exposures similar to that which would result from the CRD’s<br />
definition.<br />
314. Annex VII, Part 4, Paragraph 57(b) of the CRD requires that “the<br />
pool shall be representative for the portfolio for which the pooled<br />
data is used.” The interpretation of ‘representative’ in the context of<br />
data pools includes the following points:<br />
· Representativeness should be demonstrated according to<br />
plausible criteria that have been set in advance in a general<br />
policy.<br />
· These criteria could include the comparability of populations of<br />
exposures in the relevant parts of the pool.<br />
· In the case of pooled data, the default definition used by an<br />
institution for any part of the pool should be similar to the default<br />
definition used for that part by the other institutions in the pool.<br />
Any underestimation of risk that might result from applying<br />
similar default definitions must be avoided. Institutions should<br />
particularly avoid curtailing their use of ‘unlikeliness to pay’<br />
indicators in order to achieve this similarity in default definitions.<br />
· Where applicable, the use of appropriate statistical methods to<br />
demonstrate representativeness should be recommended.<br />
Use of data histories<br />
315. Annex VII, Part 4, Paragraph 50 requires institutions to demonstrate that<br />
their estimates of the risk parameters are representative of longrun<br />
experience. Representativeness, in the context of Paragraph 50, means<br />
that the estimates based a future time horizon of one year nevertheless<br />
reflect at least the relevant events experienced by the institution in the<br />
past over a longer time period.<br />
3.4.4. Data sources and definition of default<br />
316. If institutions use external data sets for risk parameter estimation,<br />
they should apply appropriate adjustments to their estimates in<br />
order to achieve broad equivalence to the reference definition in the<br />
CRD. In the case of pooled data, the definition of default used by the<br />
institution for any part of the pool/data set should be similar to the<br />
default definition used for that part by the other institutions in the<br />
pool, as required by Annex VII, Part 4, Paragraph 57(a).<br />
317. Special considerations for default definitions of pooled data sets: for<br />
data sets collected from the same market, the similarity of the<br />
definition of default could be checked by examining the wording of<br />
Page 73 of 123