The risk measurement and management models used by BBVA have made it a leader in best practices in the market and in compliance with Basel II guidelines.
The Bank quantifies its credit risk using two main metrics: expected loss (EL) and economic capital (EC). The expected loss reflects the average value of the estimated losses (i.e. the cost of the business) and is associated with the Group’s policy on provisions, while economic capital is the amount of capital necessary to cover unexpected losses (i.e. if actual losses are higher than expected losses).
These risk metrics are combined with information on profitability in value-based management, including the profitability-risk binomial into the decision-making process, from the definition of business strategy to the approval of individual loans, price setting, assessment of non-performing portfolios, incentives to the different areas in the Group, etc.
There are three risk parameters that are essential in the process of calculating the EL and EC measurements: the probability of default (PD), loss given default (LGD) and exposure at default (EAD). They are generally estimated using the available historical information and are assigned to operations and customers according to their particular characteristics. In this context, the credit rating tools (ratings and scorings) assess the risk in each transaction/customer according to their credit quality by assigning them a score. This score is then used in assigning risk metrics, together with additional information such as transaction seasoning, loan to value ratio, customer segment, etc. The increase in the number of default events in the current economic situation contributes to reinforce the soundness of the risk parameters by adjusting their estimates and refining methodologies. The incorporation of data from the years of economic slowdown is particularly important for refining the analysis of the cyclical behavior of credit risk. The effect on PD estimates and the credit conversion factor (CCF) is immediate. An analysis of the impact on LGD, however, requires waiting for the outcome of the recovery processes associated with those default events.