After reading the article, answer the following questions:

1. What did you learn?
2. How did the articles relate to what you learned about in the course?
3. How did the article relate to what you are interested in?
4. How is statistics used in the real world?

Getting a fix on price-fixing cartels Carsten Crede outlines the pros and cons of cartel screens - statistical tools to help spot bad business behaviour "C onservative" estimates from the Organisation for Economic Co- operation and Development (OECD) suggest that illegal cartels of businesses - which secretly collude to set prices, allocate territory and distort market competition for their own financial benefit - harm consumers to the tune of "many billions of dollars annually"." Consumers themselves might not notice: cartel activities may only add a couple of pence or cents to the price of a product. But the cost to the collective wallet soon adds up over millions of sales. To fight cartels, competition authorities around the world impose substantial fines. Yet, there is doubt as to whether this is a sufficient deterrent. There are reasons to believe that the likelihood of detection plays an important role as well. If too few of the cartels that form are detected and prosecuted, other firms might think it profitable to form cartels. 38| SIGNIFICANCE | February 2016 Economists have suggested that authorities use statistical tools in order to enhance cartel detection methods. These "cartel screens" look for evidence of illegal activity from the traces cartels might leave in data sets - in patterns of market prices, for example. However, despite a number of successes, including the spectacular uncovering of the London Interbank Offered Rate (LIBOR) manipulation scandal, there remain mixed feelings about the use of screens. Critics argue that statistical methods of detection cannot provide reliable proof of a cartel, might falsely accuse innocent firms, are very complicated to run, and are very expensive. Yet, proponents argue that cartel screens can provide a valuable and complementary deterrent.? The concept To understand the debate surrounding the use of cartel screens, let us look at some of rau Carsten J. Crede is a PhD student at the Centre for Competition Policy and the School of Economics at the University of East Anglia. His research interests include industrial organization and applied econometrics the statistical approaches that have been proposed. To understand the prospects and limitations of these screens, there is no need to go into technical detail. What is important is a discussion of the identification strategies used. Cartel screens are based on the idea that cartels alter the way in which prices, quantities, or other market outcomes are determined. The assumption here is that the activities of a cartel - the setting of prices, say - change the data generating process (see box on page 39) in a particular industry or sector. Cartel screens use statistical methods to test whether suspicious patterns resulting from these changes can be observed in available data. In broad terms, two distinct strategies have been developed to identify suspicious patterns that result from changes in the data generating process. On the one hand, collusion can be detected by looking for observations in a data set that are very unlikely to arise under 2016 The Royal Statistical Society