Stepping on a rake: The role of fiscal policy in the inflation of the 1970s

“The inflation of the 1970’s in the US is often discussed as if the only type of policy action that could have prevented the inflation were monetary policy actions and the only type of policy errors that might have induced the inflation were monetary policy errors. Yet fiscal policy underwent dramatic shifts in the 70’s” (Abstract, p. 1). Sims’s stated aim is to document these fiscal shifts, argue that contemporaries must have been uncertain about fiscal policy’s future course, lay out a theoretical framework for how such fiscal uncertainty affects the potency of monetary policy, and show that fiscal variables have predictive value for inflation in dynamic models even alongside standard monetary policy indicators (Abstract, p. 1).

In equilibrium models that include the government budget constraint (“fiscal theory of the price level” or FTPL models), if rational agents believe newly issued nominal debt is only partially backed by future taxes, debt issuance is inflationary, and “policy-generated increases in the interest rate increase, rather than reduce, the inflation rate” (Section I, p. 2). The mechanism: an increase in nominal debt not matched by higher expected future taxes leaves the public feeling wealthier, and they try to spend that apparent wealth until prices rise enough to erode it, or expectations of future taxes catch up; in this setting, “an increased nominal interest rate flows directly through to increased nominal government spending,” and in a fully flexible-price model “the monetary authority loses any ability to affect the price level” (Section I, p. 2).

Using the ratio of the primary surplus to the market value of privately held federal debt, Sims shows the US ran primary surpluses most of the time historically, with “the first period of large, sustained primary deficits beginning in 1975” – the Ford tax cut and rebate – during which “for one quarter, the primary deficit was at an annual rate of 20% of the outstanding market value of debt – a level not approached before or since in the period since 1950,” persisting at high levels for roughly two years (Section II, p. 3). He argues that while debt holders “had to be expecting that primary surpluses would eventually again become the norm, it seems farfetched to suppose that they would have thought there was some simple rule, based on historical behavior patterns, that would allow prediction of when and how primary surpluses would re-emerge” (Section II, p. 3) – i.e., the fiscal-backing uncertainty the FTPL mechanism requires was plausibly present.

Interest expense as a share of total federal expenditure was “generally well under 8%” until the early 1980s, when it “shot up to over 14% of the budget and stayed there for several years” (Section II, pp. 4-5). Sims speculates that “the fiscal discipline of the early Clinton years may have been engendered in part by these budget realities having forced Congress to recognize that its ability to tax and spend was increasingly limited by rising interest costs” (Section II, p. 5), offering this as one plausible, though not formally modeled, account of the swing back to sustained primary surpluses in the late 1990s.

In a simple endowment-economy model with a Taylor rule satisfying the Taylor principle (θ > 1, more than one-for-one response of the policy rate to inflation), combined with an “active” fiscal policy in which the primary deficit follows an exogenous stochastic process unrelated to the level of debt, the model has a unique equilibrium – but it is an explosive one, in which inflation and the nominal interest rate grow without bound (Section III.1, pp. 5-8). This directly contradicts the usual intuition that an active-money/active-fiscal combination is inconsistent with any equilibrium (as suggested by local-uniqueness analysis in Leeper 1991): “in this model, a unique equilibrium exists even when we combine an exogenously fixed primary surplus with a Taylor-principle Taylor rule. The result is (except for a knife-edge special case) a unique, explosive, equilibrium” (Section III.1, p. 8).

“There is a way for the monetary authority to end the explosiveness: Lower the interest rate and keep it fixed” – switching from an active Taylor rule to an interest-rate peg, which in this model delivers a unique, non-explosive equilibrium price level, because the previously unstable equation is replaced by a stable government budget constraint (Section III.1, p. 8). By contrast, “the fiscal authority cannot necessarily end the explosiveness by switching to a passive fiscal policy, because that leaves the equilibrium non-unique” (Section III.1, p. 8) – in this specific model, fixing the interest rate is the more reliable escape route, an asymmetry Sims flags as a striking, if model-specific, result.

Under an “active money/active fiscal” configuration in Leeper’s terminology (θ > γ in the paper’s linearized equation for expected inflation), if inflation is stable or only slowly exploding, “a δ-function shock to [the monetary policy disturbance] makes both [expected inflation] and [the real rate] jump upward by equal amounts, after which both decay back toward 0…In other words, monetary contraction has no effect on inflation, except to increase it” (Section III.2, pp. 9-11). A permanent fiscal tightening (a shift in the primary surplus), by contrast, “makes the price level jump downward and has no other effect” – so in this stark model, only fiscal policy, not monetary policy, controls the price level (Section III.2, p. 10).

Adding sticky prices (a Phillips curve), long-term government debt, consumption smoothing, and a countercyclical primary surplus to the model, Sims calibrates an “active fiscal, passive money” configuration (θ/γ < 1) and finds that a contractionary monetary shock still produces real effects in the expected direction – lower consumption and inflation initially – unlike the flexible-price case (Section III.3, pp. 12-15). But because monetary policy does not control the long-run price level in this configuration, “the inflation rate rises back above its steady state level by as much as it initially fell, and the rise is more sustained than the initial drop. This is the ‘stepping on a rake’ phenomenon: Apparently effective measures to reduce inflation come back, after a delay, to produce precisely the opposite of the desired effect” (Section III.3, p. 15) – the paper’s title and central illustrative result.

“The expansionary fiscal shock…creates a boom in consumption and an upward jump in the inflation rate. Monetary policy responds by increasing the interest rate, bringing the output boom and the increased inflation to an end. But the fiscal shock has permanently changed the price level and has financed the increased debt issue via delayed, but unanticipated at the time of the fiscal shock, inflation” (Section III.3, p. 15). In other words, monetary policy can manage the short-run cyclical consequences of a fiscal expansion, but cannot prevent the fiscal shock from eventually being financed through a permanently higher price level rather than higher future primary surpluses.

Estimating a seven-variable Bayesian VAR on 1960-2010Q1 U.S. data (real GDP, the PCE deflator, one-year and ten-year Treasury rates, the primary-deficit-to-debt ratio, debt-to-GDP, and interest expense over receipts), Sims finds “the variance of the price level is determined mainly by two disturbances, the [output] and [price] innovations,” and that the difference between these two shocks produces a response pattern – rising prices alongside declining projected future primary deficits – that “looks like a possible candidate for fiscal effects on inflation,” with a statistically significant positive inflation response (Section IV, pp. 16-19). He cautions, however, that this shock’s negative output response “cannot be interpreted as a pure random fiscal disturbance,” and instead may reflect “the kind of expansionary fiscal surprises that accompanied the oil price shocks of the 1970’s,” so this fiscal-like channel accounts for “a non-trivial, but far from dominant, component of historical price variation,” leaving open whether the fiscal changes were “an essential part of the transmission mechanism to inflation, or were instead a passive reaction to the output contraction” (Section IV, pp. 18-19).

“There is no excuse for econometric models intended for monetary policy analysis to continue to omit serious treatment of fiscal behavior. It is clear from theoretical analysis that fiscal policy can be a primary transmission mechanism or a primary source for changes in the inflation rate” (Section V, “Conclusion,” p. 19). Writing in 2010, Sims argues the point is especially pressing because “we are…entering into a period of remarkable shifts in fiscal policy and remarkable uncertainty about fiscal policy,” and because “central bank balance sheets have expanded to the point that possible effects of monetary policy on the fiscal situation cannot be ignored” – a central bank considering the full range of its own and fiscal authorities’ impacts on output and inflation “should be using a quantitative model that treats explicitly and realistically the potential impacts of fiscal policy on the price level” (Section V, pp. 19-20).