Nonlinear rank scaling and hidden structure in NHS expenditure transparency data

• Analysis of 1.96 million financial transactions from NHS transparency data. • Financial transparency limited to >£25k conceals most transactions. • Financial transparency limited to outgoing transactions limits public oversight. • Total expenditures to suppliers resemble word use kernel and extended lexicons. A variety of transparency initiatives have been introduced by governments to reduce corruption and allow citizens to independently evaluate the effectiveness and efficiency of spending. In 2010, the UK government mandated transparency for many expenditures exceeding £25,000. The resulting data is dispersed across a range of governmental organizations and presents an opportunity to understand expenditure at scale, interrogate organizational structures and develop transparency measures. Here, we focus on data from the top two layers of the National Health Service (NHS) within England, including NHS England (NHSE) and Integrated Care Boards (ICBs). As one of the largest government run healthcare organizations in the world and potentially the sixth largest employer globally, the NHS provides a distinctive case for studying healthcare delivery, contractor dynamics, and organizational self-organization. We find that limiting transparency to larger transactions conceals a substantial share of spending from scrutiny, including most transactions. The rank-frequency distributions of suppliers, expense types, and spending categories exhibit multiple scaling regimes similar to patterns observed in word frequency and urban scaling studies indicating these methodologies can be deployed to analyze financial transparency data at scale.