Liudas Giraitis, George Kapetanios, Ana Wetherilt, and Filip Zikes, "Estimating the Dynamics and Persistence of Financial Networks, with an Application to the Sterling Money Market", Journal of Applied Econometrics, Vol. 31, No. 1, 2016, pp. 58-84. The data come from three sources: 1) Proprietary payments database 2) Proprietary reserve account balances database 3) Bank of England website 1) Proprietary payments data The main source of our data is the proprietary payments database maintained by the Bank of England as a sole operator of the UK's real-time large-value payment system CHAPS. These data contain all messages (payments) between the member banks of CHAPS, the so-called settlement banks, in the period between 2 January 2003 and 31 March 2012. Each message contains infomation about the sender and receiver of funds, the payment value in pound Sterling, and the time and date of the transaction. From these payment messages, we select those that are deemed to be associated with overnight unsecured lending using the Furfine (2001) algorithm as described in the Appendix. Given the inferred overnight loans transactions, we construct, for each pair of settlement banks, daily aggregate gross values lent and borrowed expressed in pound Sterling. For example, for a pair of banks A and B we calculate the total amount lent by A to B and the total amount lent by B to A, for each day in our sample period. Note that we model the gross amounts rather than the net lending between A and B, as close-out netting is not allowed in case of default, and hence gross funds borrowed have to be repaid in full by the surviving banks. This gives us a time-series of daily matrices, which are then modelled as described in the paper. In addition to the daily values lent and borrowed, we construct a daily time series of average overnight interest rate, which is reported in Figure 2. For each day, the average overnight interest rate is calculated as a value-weighted average interest rate using all overnight loan transactions on that day. The resulting average rate is then annualised using the 365/365 convention. 2) Proprietary reserve account balances In Figure 2, we report daily aggregated reserves. These are obtained by summing up the end-of-day balances of settlement banks in their reserve accounts with the Bank of England. These data are proprietary and are maintained by the Bank of England. 3) Bank of England website In Figure 2, we report the Bank rate, i.e. the main policy rate of the Bank of England. In Section 5, we discuss a robustness check that is based on the Sterling Overnight Index Average rate (SONIA). Both time series are obtained from the Bank of England website.