When you use BigDecimal, you’ll need to specify the RoundingMode for the result, which can be UP (away from zero), DOWN (toward zero), CEILING (toward positive infinity), FLOOR (toward negative infinity), HALF_UP (toward nearest neighbor, or up if both neighbors are equal), HALF_DOWN (toward nearest neighbor, or down if equal), HALF_EVEN (toward nearest neighbor, or to the nearest even neighbor if equal), or UNNECESSARY (result should be exact). [4] X Research source