We've already mentioned that it is important to understand the origin of missing values and whether they can offer us some information. In the following example, we will analyze the missing values presented in the UBPRE628 variable. This variable measures the amount of total long-term debt of a bank divided by the total bank equity capital. The capital of a bank is important because in the case of losses faced in the operations, this will be used by the bank to absorb them and to avoid a future insolvency. The higher the capital, the higher the buffer of a bank to face economic problems.

In general, the highest proportion of debt related to the bank capital, the more problems the bank could experience in the future if, for example, a new crisis occurs. In the event of a new financial crisis, a bank couldn’t repay its debt, even by selling its assets.

According to our analysis, this variable displays a high percentage of missing values, specifically this ratio is not informed for the 23.97% of banks in our data:

missingdata[missingdata$colnames=="UBPRE628",]

 ##     colnames ncol nmsg percmissing
 ## 281 UBPRE628 7091   17   0.2397405