Perhaps no issue has greater political salience in the United States today than “affordability.” Yet the task of constructing official measures of the cost of living is widely understood to be a neutral, technocratic endeavor. The choices involved—which goods to include and how to weight them, how to account for regional variation, which administrative and business data to use—are seen
as difficult yet tractable operational questions. For much of the twentieth century, however, cost-of-living measures were the subject of ongoing conflicts between capital and labor and were understood to require contestable normative judgments. By examining the historical development of the most widely used price index, the Consumer Price Index (CPI), we show how this older understanding shifted under the pressure of new kinds of economic expertise as well as changes in administrative law and procedures. Cost-of-living measures were claimed increasingly to be constructed according to neutral or objective standards, with professional economists playing a key role in devising and legitimating them. This new, more sophisticated approach, which was favored by big business, obscured the unavoidably discretionary aspects of price-index construction, and in particular the need for a price index to be underpinned by appropriate normative judgments. As a result, most recent revisions to the CPI have failed to realize its original aim: to capture changes in the requirements for living in “decency and health.” Identifying this shift in the construction of price indexes sheds light on several historical and contemporary controversies concerning inflation, wages, and changing living standards, as well as how changes in administrative law and the judicial review of agency decision-making pushed officials toward these more ostensibly neutral measures, thus laying the groundwork for this historical transformation in the CPI. This diagnosis thus suggests a better way forward: policy changes to restore a proper normative grounding to the CPI to redress its current limitations in capturing “affordability.”