This week we learned about prediction markets and what I found most interesting is how accurate these predictions are.
In politics, Berg, Nelson & Rietz (2008) found that 74% predictions since 1988 until 2004 were very close to the actual outcome. Now, why does this matter? Well most political campaigns use these numbers to get supporters and in addition this analysis can be used to see where the candidate lacks supporters. For instance, a candidate might have a strong support from a certain neighbourhood, but not from another, which in turn can help the campaign focus their assets on getting supporters from the other neighbourhood. Or, the numbers can show the campaign where the other candidate has the most support from, so that the campaign can either attempt to take away some of that support or focus more on the opposing side for themselves.
When it comes to the stock market, I found an interesting article that explains how Google Trends can be used in predicting the events of the stock market. Preis, Moat & Stanley (2013) argue that when a lot of people Google the terms: “debt” and “inflation” it is very likely that they are concerned about the current market and that they are likely to start selling their shares. When the researchers applied this theory it lead them to a 326% profit within 7 years. However, the article also warned that the more the people use this strategy the less accurate it will be. In addition, in a more modern world new keywords will appear and more research on this topic will be needed.
However, the prediction market is not perfect and it can be manipulated by outside forces. Simple rumours about a company, product or a person can change the stock price temporarily and affect the predictions.
So, its overall accuracy is hard to determine and aside from just using prediction market the researcher should be aware of its flaws.
References:
Berg, J. E., Nelson, F. D., & Rietz, T. A. (2008). Prediction market accuracy in the long run. International Journal of Forecasting, 24(2), 285-300.
Preis, T., Moat, H. S., & Stanley, H. E. (2013). Quantifying trading behavior in financial markets using Google Trends. Scientific reports, 3.
