See P= DFA0 = 1101 1111 1010 0000 = -8288(Decimal equivalent)
P*4 = -8288*4= -33152
but as you know that 2's complement number have range between $-2^{n-1}$ to $2^{n-1} - 1$
so in 16 bit you can represent last negative number as $-2^{16-1}$ = $-2^{15}$ = -32768 to 32767
and multiplication is going out of the range so it can't be represent in 16 bit register. to get the result you need to store in 32 bit register and answer will be
11111111111111110111111010000000