Wagner’s law and governments’ functions: granularity matters

Purpose - The purpose of this paper is to assess the responses of different categories of government spending to changes in economic activity. In other words, the authors empirically revisit the validation of the Wagner’s law in a sample of 61 advanced and emerging market economies between 1995 and 2015. Design/methodology/approach - The authors do so via panel data instrumental variables and time-series SUR approaches. Findings - Evidence from panel data analyses show that the Wagner’s law seems more prevalent in advanced economies and when countries are growing above potential. However, such result depends on the government spending category under scrutiny and the functional form used. Country-specific analysis revealed relatively more cases satisfying Wagner’s proposition within the emerging markets sample. The authors also found evidence of counter-cyclicality in several spending items. All in all, the Wagner’s regularity seems more the exception than the norm. Originality/value - While in the literature on the size of the public sector with respect to a country’s level of economic development has received much attention, the authors make several novel contributions: since some economists criticized Wagner’s law because of ambiguity of the measurement of government expenditure (Musgrave, 1969), instead of looking at aggregate public expenditures, the authors go much more granular into the different functions of government (to this end, the authors use the Classification of Functions of the Government nomenclature). The authors check the validity of the Law via an instrumental variable approach in a panel setting; after that, the authors take into account the phase of the business cycle using a new filtering technique to compute potential GDP (output gap); then, the authors cross-check the baseline results by considering alternative functional form specifications of the Law; and finally, the authors look at individual countries one at the time via SUR analysis.