The calculation of downturn LGD is subject to specific requirements
Institutions characterise an economic downturn in terms of economic and credit indicators. This is done on the basis of the observed evolution of these indicators over a historical period. When analysing historical series in order to characterise a downturn period, institutions should take the following into consideration.
- The length of the historical dataset of economic indicators should be at least the most recent 20 years.
- As a minimum, and where relevant, institutions should consider (for all exposure types) indicators (analysed separately) such as GDP growth, unemployment rates, interest rates, inflation rates, system-wide default rates and credit losses, complemented by internal series (i.e. default rates, losses) where available. Additional indicators should be considered for the following types of exposure:
- exposures to corporate and retail SMEs – sectoral/industry indices;
- exposures to residential real estate – house prices, region-specific indices;
- exposures to other retail – consumer leverage ratio or similar information.
- The specified downturn period should be a minimum of one year. Longer periods could be also be considered in order to account for cases where the historical data show longer stress periods for some indicators, or where the peaks or troughs of different economic indicators are not reached simultaneously but are nonetheless the effect of one single overall downturn. In such cases, the downturn period should be long enough to reflect the continued stressed situation.
Evidence for elevated realised LGD
Consequently, the specified downturn conditions are evidenced by elevated levels of realised LGD including treatment for incomplete recovery processes (in accordance with paragraph 110 above) at portfolio level, or at the relevant sub-range of application, driven by stressed levels of the relevant economic indicators (as specified in paragraph 120 above).
Institutions should derive LGD estimates which are appropriate for the downturn conditions specified, following the principles set out in paragraphs 120 and 121 above. Any lag between the beginning of the downturn period and the date of the impact on the realised LGDs should be taken into account. This means that even where high levels of realised LGD are not experienced simultaneously with the stress seen in economic indicators, but are still the result of such stress, they should be considered as the LGD estimates appropriate for the economic downturn.
In assessing the accuracy of those of the LGD estimates that are appropriate for an economic downturn, institutions should compare the LGD estimates derived in accordance with paragraph 122 above with a reference value derived according to the following steps.
- First, institutions should identify, from the most recent 20 years, the two individual years with the highest observed losses considering the defaults observed in those years.
- Given the current circumstances (adverse economic conditions experienced in many countries since 2008), the most recent 20 years can
be replaced with the most recent 10 years for estimations made during 2017. Thereafter, this period should be increased by one year each year until the period of 20 years is reached, provided representativeness requirements are met. Institutions should be able to provide evidence that the period considered actually contains years which include adverse economic conditions.
- To identify the two individual years referred to above, institutions should
- group all defaults within the RDS and corresponding exposures and losses by the year in which the default occurred and obtain the ratio of total losses to total exposure; and
- select the two individual years with the highest ratio of total losses to total exposure. This analysis should consider years for which the maximum length of recovery process has been observed.
- Second, institutions should calculate reference values as the average realised LGD from those two individual years (see paragraph 123(a)
above) for each facility grade or pool that they use. When the LGD estimates result from combining different components (for example, secured and unsecured), the reference values can be calculated at the level of each of the components and the comparison made at this level
Where the downturn LGD estimates (by facility grade or pool) or, if applicable, estimates of model components (including MoC) obtained by the institution are lower than those resulting from the reference value described above, the institution should be able to provide evidence that its downturn LGD methodology is aligned with the target of elevated LGDs driven by economic conditions (as specified in paragraph 121 above).
The reference value referred to in this paragraph should not be considered as a valid methodological option. Institutions are expected to develop internal methodologies compliant with paragraphs 120 to 122 above to estimate LGD appropriate for an economic downturn.
Where an institution does not have a data series with the length described above or cannot provide evidence that the available data include adverse economic conditions, to comply with the requirement to estimate LGD appropriate for an economic downturn the approach described above should be applied with the available data series and an add-on or MoC should be applied.
- ECB guide to internal models - Credit Risk, Sep 2018