When we don't know the proportion ahead of time, we assume that p=0.5. This assumption gives us the largest possible sample size since we multiply p(1-p).
The formula you need is
n=1/4 (z/E)2
where 1/4 is p(1-p) when p=1/4
z = z-score for a 98% confidence level = 2.33
E = margin of error = 0.01
n= (1/4)(2.33/.01)2 =13572.25 = 13573
When determining sample size, always round up, even when normal rounding rules would suggest otherwise.