We test the efficient market hypothesis to see if Twitter aggregates information faster than a real-money prediction market. We use Support Vector Machines (SVMs), a supervised learning algorithm, to predict the outcome of the 2012 U.S. presidential elections via Twitter data. We then compare the prediction from SVM against the Iowa Electronic Markets (IEM). A total of 40 million unique tweets were collected and analyzed between September 29 th 2012 and November 6 th 2012. We observe: 1) The IEM is efficient on all the above days as per the semi-strong efficient market hypothesis definition [1]. SVM does not out predict the IEM. 2) The SVM prediction results are positively correlated with the IEM and predicts Obama winning the election, implying that Twitter can be considered as a valid source in predicting US presidential election outcomes. Using the Granger causality test, no causal relationship was inferred between the two-time series. 3) The candidate frequency count distribution independent of any sentiment analysis on all days is also positively correlated with IEM and SVM. Using Granger causality test, we determined that IEM statistically causes the candidate frequency count distribution in Twitter at the 1% level.