This statistic isn't enough to say whether unions are good or bad, even if you assume there's causation. How do you know they don't do that by slowing down productivity growth, rather than by speeding up wage growth?