James Angel of Georgetown University’s Department of Finance writes: “High quality data are essential for markets to function efficiently. This study describes the production and dissemination of U.S. stock market data and explores how much individual investors are paying for the data. The joint nature of the production of information along with trading, surveillance, and listing services makes it difficult to allocate the fixed costs to customer classes. Data from the competing exchanges are consolidated and distributed through entities known as Securities Information Processor (SIPs). The costs of providing this data fall mainly on the professional investors who value it the most. Professional traders pick up over 80% of the cost of the SIP data. The SIPs intentionally provide real-time data at very low cost to nonprofessional investors and delayed data for free. The inflation-adjusted price of real-time Tape C nonprofessional data has fallen 96.3% since 1987. Since 2008, inflation-adjusted SIP revenues allocated to the exchanges have fallen 23.7%. These cost reductions are one of the many contributors to the massive reduction in trading costs that has occurred in recent years. Not only has the price fallen, but the latency, the amount of time the SIPs take to process and report trades and quotes, has fallen as well. Latency has fallen 99.7% since 2010. The cost for nonprofessional data is quite low. For a very large broker, the cost of real-time nonprofessional data is about $0.17 per customer per month, about the same as a sip of Starbucks.”