Recent explanations of observed inflation have followed two main lines. The `discretionary policy' approach considers a government effecting resource transfers from the private sector through `surprise' inflation. The `inflation smoothing' approach, in contrast, stresses the optimal distribution of inflationary distortions over time and suggests that inflation will be persistent, regardless of its current level. Authors adopting this approach apply the Ramsey optimal taxation principle to government inflation plans and predict that both inflation and total government spending commitments will follow martingale processes, but they have found mixed empirical support at best.

In Discussion Paper No. 519, Research Fellow Maurice Obstfeld notes that while examples of high and seemingly chronic inflation certainly abound, there have also been notable episodes of successful inflation reduction, often coupled with government fiscal consolidation. These include the UK and US governments' returns to fixed gold parities following the Napoleonic and Civil Wars and World War I, and less dramatically the generalized disinflation in the industrial countries in the early 1980s. The available models cannot account for such important episodes of government behaviour. Most discretionary policy models include no intrinsic sources of dynamic evolution and therefore make no allowance for the dynamics of public debt or the role of government budgets in determining the public's inflationary expectations. Inflation smoothing models, in contrast, place the determination of public debt at centre stage, but the optimal plans derived from the Ramsey rule are time consistent only in very special cases, and such models' predictions will not apply if the government has discretion to set policy in each period.

Obstfeld develops a dynamic model of seigniorage with elements of both approaches. The model explicitly incorporates the intertemporal linkages among budget deficits, debt and inflation, and economies' equilibrium paths reflect the ongoing strategic interaction between an optimizing government and a private sector with rational expectations. He finds that inflation at each point in time is higher than it would be for a government unable to commit to future tax policies; but tax-smoothing behaviour may produce an inflation rate that falls over time (for plausible parameters) towards the socially preferred (possibly zero) long-run rate. The public's rational expectations of inflation based on government budgetary conditions give the authorities additional incentives to retire debt and hence reduce future seigniorage needs. Equilibria with rising or persistently high inflation may nevertheless occur when the government discounts the future much more heavily than the private sector.

Dynamic Seigniorage Theory: An Exploration
Maurice Obstfeld

Discussion Paper No. 519, March 1991 (IM)