So you want to avoid loops?
Here you have it:
public static String repeat(String s, int times) {
    if (times <= 0) return "";
    else return s + repeat(s, times-1);
}
(of course I know this is ugly and inefficient, but it doesn't have loops :-p)
You want it simpler and prettier? use jython:
s * 3
Edit: let's optimize it a little bit :-D
public static String repeat(String s, int times) {
   if (times <= 0) return "";
   else if (times % 2 == 0) return repeat(s+s, times/2);
   else return s + repeat(s+s, times/2);
}
Edit2: I've done a quick and dirty benchmark for the 4 main alternatives, but I don't have time to run it several times to get the means and plot the times for several inputs... So here's the code if anybody wants to try it:
public class Repeat {
    public static void main(String[] args)  {
        int n = Integer.parseInt(args[0]);
        String s = args[1];
        int l = s.length();
        long start, end;
        start = System.currentTimeMillis();
        for (int i = 0; i < n; i++) {
            if(repeatLog2(s,i).length()!=i*l) throw new RuntimeException();
        }
        end = System.currentTimeMillis();
        System.out.println("RecLog2Concat: " + (end-start) + "ms");
        start = System.currentTimeMillis();
        for (int i = 0; i < n; i++) {
            if(repeatR(s,i).length()!=i*l) throw new RuntimeException();
        }               
        end = System.currentTimeMillis();
        System.out.println("RecLinConcat: " + (end-start) + "ms");
        start = System.currentTimeMillis();
        for (int i = 0; i < n; i++) {
            if(repeatIc(s,i).length()!=i*l) throw new RuntimeException();
        }
        end = System.currentTimeMillis();
        System.out.println("IterConcat: " + (end-start) + "ms");
        start = System.currentTimeMillis();
        for (int i = 0; i < n; i++) {
            if(repeatSb(s,i).length()!=i*l) throw new RuntimeException();
        }
        end = System.currentTimeMillis();
        System.out.println("IterStrB: " + (end-start) + "ms");
    }
    public static String repeatLog2(String s, int times) {
        if (times <= 0) {
            return "";
        }
        else if (times % 2 == 0) {
            return repeatLog2(s+s, times/2);
        }
        else {
           return s + repeatLog2(s+s, times/2);
        }
    }
    public static String repeatR(String s, int times) {
        if (times <= 0) {
            return "";
        }
        else {
            return s + repeatR(s, times-1);
        }
    }
    public static String repeatIc(String s, int times) {
        String tmp = "";
        for (int i = 0; i < times; i++) {
            tmp += s;
        }
        return tmp;
    }
    public static String repeatSb(String s, int n) {
        final StringBuilder sb = new StringBuilder();
        for(int i = 0; i < n; i++) {
            sb.append(s);
        }
        return sb.toString();
    }
}
It takes 2 arguments, the first is the number of iterations (each function run with repeat times arg from 1..n) and the second is the string to repeat.
So far, a quick inspection of the times running with different inputs leaves the ranking something like this (better to worse):
- Iterative StringBuilder append (1x).
- Recursive concatenation log2 invocations (~3x).
- Recursive concatenation linear invocations (~30x).
- Iterative concatenation linear (~45x).
I wouldn't ever guessed that the recursive function was faster than the for loop :-o
Have fun(ctional xD).