In 2017, The Economist (2017) announced that “the  world’s most valuable resource is no longer oil,  but data,” arguing that data is fuelling the modern  economy just as oil did a century earlier. In the  same year, Jonathan Haskel and Stian Westlake  (2017) published Capitalism without Capital: The Rise  of the Intangible Economy, arguing that intangible  capital, which includes data but also research and  development (R&D), patents and other intellectual  property (IP), is now the main driver of advanced  economies, rather than physical capital such as  machinery and buildings. Since then, the perceived  importance of data to business and the economy  at large has only grown, especially given the  importance of data as the raw material in artificial  intelligence (AI) systems, which promise to be a  new “general purpose technology” as consequential  and wide ranging as computers or electricity. 

The importance of data to the economy, and  particularly its increasing importance as an asset,  poses a challenge to statistical agencies. While  the current SNA standard recognizes the costs of  database management software, it does not treat  data as an asset, and does not count the costs of  acquiring the data. This absence reflects some  of the very real problems that exist in valuing  a commodity that is very situation-specific,  is usually not bought and sold in transparent  markets (or at all), and, consequently, is not  easily obtainable from firms’ balance sheets. 

Despite these problems, several national  statistical offices have risen to the challenge of  attempting to measure the value of data in a  way that is consistent with national accounts  concepts for capital. In this paper, the authors  will take stock of these attempts, assess how  plausible their estimates are, examine other  possible methods of valuation and suggest  potential ways forward for the valuation of data,  both in an SNA context and more broadly.