This is the equivalent of PACKED-DECIMAL.
To improve compilation performance, specify odd numbers of numeric character positions in the picture clauses for COMP-3 (packed decimal) items. Internally, the rightmost byte of a packed decimal item contains a digit and a sign, and any other bytes contain two digits. If you use the more efficient configuration, the compiler does not need to supply the missing digit.
The contents of the leftmost (unused) digit position of the storage allocated for a packed decimal item that contains an even number of digits can change when the value of the item is changed. This might lead to unexpected results if, for example, the item is redefined, forms part of a group field, or is used as a key to an indexed file.