Numerous normative studies have examined the optimal rate of inflation from the perspective of public finance. In the benchmark public finance models, the government can use debt, direct taxes and seigniorage to finance exogenous expenditures. To characterize the optimal inflation rate, these models focus on the potential role of seigniorage revenues in the optimal tax package. Several recent studies have reinterpreted the smoothing implications of optimal seigniorage models as positive theories of inflation and tested these implications as such. Recent empirical tests of dynamic optimal seigniorage models focus on their smoothing and long-run implications. The models also imply that the optimal policies are forward-looking, that is that seigniorage revenues depend on expected future government expenditures. In Discussion Paper No. 1121, Behzad Diba and Research Affiliate Philippe Martin report causality tests of forward-looking policies for Germany, France, Italy, the UK and the US.

The authors first investigate the stationarity properties of the series and strongly reject the presence of unit roots for the seigniorage/GDP ratio, both with and without a deterministic trend. For the causality tests, the authors consider three alternative specifications of the government expenditures/GDP ratio, depending on its stationarity properties. The tests fail to detect the implied Granger causality in the data for Germany, France, Italy and the US. This constitutes evidence against the model of optimal seigniorage under examination, and casts doubt on the empirical relevance of the public finance approach to inflation and seigniorage. The UK results, which detect causality running from the seigniorage/GDP ratio to the expenditures/GDP ratio, imply that the model may have empirical relevance for the UK.

Causality Implications of the Public Finance Approach to Inflation and Seigniorage
Behzad Diba and Philippe Martin

Discussion Paper No. 1121, January 1995 (IM)