What can 100,000 books tell us about the international public library e-lending landscape?

Introduction. We investigated the relative availability of e-books to libraries for e-lending in five English-language countries & analysed their licence terms and prices. Method. We created a unique dataset recording author, publisher, price and terms for 100,000 titles and 388,045 e-lending licences across Australia, New Zealand, Canada, the United States and United Kingdom through the aggregator Overdrive. We developed new algorithms to estimate the original publication year for each title & to match titles across jurisdictions. Analysis. We examined the relationships between title price, age, terms, jurisdiction, publisher and publisher type using various statistical analyses and machine learning. Results. Price and licence differences across countries are largely attributable to the Big 5 publishers. Prices are largely independent of title age (unless the title is in the public domain) or the rights libraries obtain in exchange. Licence terms are not affected by age either, meaning that the most restrictive terms are often applied to older, less demanded books. Conclusions. By setting terms independent of titles' value to libraries, publishers may discourage libraries from adding older and less-demanded books to their collections. We will test this hypothesis in a follow-up library survey.