The Ends of Four Big Inflations

Sargent’s central claim is that each of the four post-World War I hyperinflations in Austria, Hungary, Poland, and Germany ended abruptly, at a comparatively minor cost in foregone output and unemployment, once the government made a sufficiently binding, credible change in its fiscal and monetary regime – evidence, in his reading, for the rational-expectations view that credibility, rather than gradualism, determines the cost of disinflation. He writes that his “reason for studying these episodes is that they are laboratories for the study of regime changes”: within each of the four countries “there occurred a dramatic change in the fiscal policy regime, which in each instance was associated with the end of a hyperinflation,” and even though the events are “extreme and bizarre,” Sargent argues “it is precisely because the events were so extreme that they are relevant” to interpreting more moderate, contemporary inflations (Introduction).

The momentum view Sargent challenges holds that persistent inflation has “a stubborn, self-sustaining momentum” rooted in backward-looking expectations, so that restrictive monetary and fiscal policy raises unemployment substantially while doing little to reduce inflation except slowly and at great cost. On this view, firms and workers extrapolate past inflation into their expectations of an “underlying rate of inflation,” so that “restrictive monetary and fiscal actions in the first instance cause substantial reductions in output and employment, but have little, if any, effects in reducing the rate of inflation” (Introduction); Sargent cites the contemporary estimate that reducing U.S. inflation by one percentage point through restrictive policy would cost 220 billion dollars of annual GNP.

The rational-expectations alternative Sargent favors holds that people expect high future inflation precisely because current and prospective government monetary and fiscal policy warrants those expectations, so inflation can be stopped much more quickly and cheaply than the momentum view implies – provided the government undertakes “a change in the policy regime,” an “abrupt change in the continuing government policy…that is sufficiently binding as to be widely believed,” rather than a temporary restrictive action. Sargent is careful to note that this “is not to say that it would be easy to eradicate inflation,” since economists lack “reliable, empirically tried and true models” that can predict precisely how disruptive a given regime change will be, and how costly it proves “would depend partly on how resolute and evident the government’s commitment was” (Introduction).

Sargent draws on “recent work in dynamic macroeconomics” to argue that private agents change their own behavioral rules whenever the government changes its policy regime or strategy, so econometric relationships estimated under one regime cannot be expected to survive a change to a different one – meaning the historical hyperinflations can only be read as evidence about regime change if the observed policy shifts really were changes in the rules of the game, not merely isolated actions within an unchanged strategy. He acknowledges this is “an admittedly delicate task to interpret” from the historical record alone, since “all that we have to go on are the recorded actions actually taken, together with the pronouncements of public officials, laws, legislative votes, and sometimes constitutional provisions,” but argues the four episodes he studies are “about as close to being laboratories for studying regime changes as history has provided.”

Sargent frames government money and debt as valued the way a firm’s debt is valued – by the present value of the revenue stream (here, prospective budget surpluses) that backs it – so that under a gold or “backed” currency regime a government “must honor its debts and could not engage in inflationary finance,” while a purely “fiat,” unbacked currency has value only insofar as the public expects future tax collections, rather than future money creation, to service the government’s obligations. He explicitly invokes Keynes’s view that “the size of a government’s gold reserve was not the determinant of whether it could successfully maintain convertibility with gold: its fiscal policy was,” and treats the value of government debt as “equal to the present value of current and future government surpluses,” so assigning value to debt “necessarily requires a view about the fiscal policy regime in place” (Gold Standard section).

Austria financed enormous post-war budget deficits – “typically over 50 percent” of total government expenditures between 1919 and 1922 – by having the Austrian section of the Austro-Hungarian bank discount treasury bills at low nominal interest rates, and the resulting note issue (up 288-fold between March 1919 and August 1922) drove a “flight from the crown” that the government resisted with exchange controls administered by the Devisenzentrale; the hyperinflation, which averaged roughly 10,000 percent per annum from January 1921 to August 1922, ended abruptly in August-September 1922 once the League of Nations intervened. Three League protocols of October 1922 committed Austria to establish “a new independent central bank” barred from further advances to the government, to “cease running large deficits,” and to accept a League-appointed Commissioner General to monitor compliance, in exchange for an international reconstruction loan; the new Austrian National Bank, opened in January 1923, was “specifically forbidden from lending to the government except on the security of an equal amount of gold and foreign assets,” and the government balanced its budget within two years by discharging thousands of employees and raising prices and taxes (Austria section, pp. 9-13). Sargent notes that the exchange rate stabilized in August 1922 and prices a month later even though note circulation kept rising more than six-fold through 1924 – a pattern he attributes not to a violation of the quantity theory but to the changed, now asset-backed, character of what those notes represented; recorded unemployment-relief recipients rose from a low of 8,700 in December 1921 to 167,000 by March 1923 before receding, a cost Sargent calls “minor when compared with” the 220-billion-dollar tradeoff invoked in U.S. debate.

Hungary’s hyperinflation followed a similar script – large deficits from 1919-1924 financed through the State Note Institute, compounded by especially generous, below-market-rate loans to private borrowers that Sargent identifies as “a much larger” source of high-powered-money growth than in the other three countries – and prices (up 263-fold between January 1922 and April 1924) and the krone’s exchange rate stabilized abruptly in March 1924 once a League of Nations-brokered reconstruction plan, paralleling Austria’s, created the Hungarian National Bank in mid-1924 and bound it against unsecured lending to the government. As in Austria, note and deposit liabilities of the central bank kept expanding after stabilization (more than three-fold between March 1924 and January 1925), which Sargent again attributes to the shift from unbacked to gold- and sterling-backed liabilities once the League’s reconstruction protocols took effect; unemployment data, available only from just after stabilization, show no clear increase in the following year or two, which Sargent finds “more plausible” to read as evidence of a limited employment cost than as evidence of a cost too long-lasting to appear in the record (Hungary section, pp. 13-17).

Poland ran large deficits through 1924, financed by borrowing from the Polish State Loan Bank (whose outstanding notes rose 523-fold between January 1922 and December 1923, alongside a roughly proportional collapse in the mark’s dollar exchange rate), but unlike Austria and Hungary achieved its initial stabilization in January 1924 without a foreign loan or League of Nations intervention – the Minister of Finance was instead granted emergency powers to found the Bank of Poland, which was barred from further unsecured lending and capped in the credit it could extend to the government, alongside a new gold-linked currency unit, the zloty. Sargent describes essentially the same post-stabilization pattern found elsewhere – central bank note circulation continued to rise (up 3.2-fold in 1924 alone) even as prices and the exchange rate held roughly stable, a pattern he again attributes to the notes becoming effectively backed by gold, foreign exchange, and private paper; unemployment rose noticeably around the January and July 1924 stabilization dates but, in Sargent’s assessment, was “not nearly as bad” as U.S.-style estimates of the inflation-output tradeoff would predict. He also notes that Polish stabilization proved less durable than the other cases: the zloty depreciated again from late 1925, which Sargent attributes to the government’s “premature relaxation” of exchange controls and continued below-market lending by the central bank, requiring a further foreign loan in 1927 (Poland section, pp. 18-20).

Germany’s hyperinflation, “the most spectacular” of the four, was dominated by the burden and uncertainty of reparations and, from January 1923, by the government’s policy of financing “passive resistance” to the French occupation of the Ruhr through Reichsbank discounting of treasury bills; wholesale prices rose 2,038-fold between January 1922 and July 1923 and 25,723-fold by August 1923, while Reichsbank notes rose “only” 378-fold and then 5,748-fold over the same intervals – a gap Sargent reads as evidence that Germans were aggressively economizing on marks by holding foreign currency instead. The inflation stopped abruptly in November 1923 following the October 15, 1923 decree creating the Rentenbank and the Rentenmark, which Sargent argues mattered not for its “cosmetic” unit change but because it imposed hard, credible ceilings on the total volume of Rentenmarks and on credit to the government, and because the government simultaneously moved to balance its budget by cutting the civil service – including a 25 percent cut in government employees under the October 27, 1923 personnel decree – and by obtaining relief from reparations under the Dawes Plan. By “all available measures,” Sargent reports, stabilization was followed by “increases in output and employment,” and although 1923 was a bad year for German production, substantially because of the Ruhr occupation and passive resistance rather than the anti-inflation policy as such, 1924 was better than 1923 and 1925 was a good year – even though Sargent also finds evidence, drawing on Graham’s data, of real economic distortion from the episode, including “overinvestment” induced by the collapse in the real return to money and government debt (Germany section, pp. 20-23).

Sargent resolves this apparent paradox by distinguishing “unbacked” or “outside” money – central bank liabilities backed mainly by government treasury bills that carried no real commitment to future taxation – from “backed” or “inside” money, created through open-market operations in gold, foreign exchange, and commercial paper that were fully backed at the margin; once each country’s reconstruction protocols took effect, the “proper interpretation” of the central bank’s note issue changed even though the number of notes in circulation kept rising, so the further note growth should not, in his view, be read as inflationary. He illustrates the underlying logic with a simple money-demand model in which a previously unexpected, permanent, sudden drop in the rate of money creation forces a discrete, once-and-for-all jump upward in the level of the money stock in order to keep the price level from falling – and suggests that the gradual, rather than instantaneous, rise in money actually observed after each stabilization is best explained either by the public adjusting its expectations to the new regime only gradually (an explanation Sargent says he finds “hard to accept”) or by adjustment lags in real money demand, though his own preferred explanation “stresses the distinction between backed and unbacked money.”

Sargent reports that every country’s stabilization was accompanied by some rise in measured unemployment, but characterizes these increases as comparatively “minor” set against the cost implied by the 220-billion-dollar-per-percentage-point estimate invoked in contemporary U.S. discussions, while explicitly cautioning that how much of the recorded unemployment reflects the stabilization itself, as against other real dislocations, “cannot be determined” from the available data. In Austria, relief recipients climbed from 8,700 to a peak of 167,000 before receding; in Hungary, unemployment shows no clear rise in the year or two after stabilization; in Poland, unemployment rose around the 1924 stabilization dates but Sargent judges it “not nearly as bad” as U.S.-style estimates would predict; and in Germany, “by all available measures,” stabilization brought “increases in output and employment,” with the poor 1923 performance attributed mainly to the Ruhr crisis rather than to the anti-inflation policy itself.

Czechoslovakia serves as a comparison case demonstrating that the same credible-regime-change logic worked even before a hyperinflation had a chance to develop: under Finance Minister Alois Rasin, Czechoslovakia “adopted the conservative fiscal and monetary policies which its neighbors did only after their currencies had depreciated radically,” stamping and taking over the Austro-Hungarian notes circulating within its borders, statutorily capping the central banking office’s unbacked note issue in April 1919, and running only modest deficits from 1920 on. As a result, Sargent reports, Czechoslovakia “avoided the hyperinflation experienced by neighbors” and even experienced a mild deflation in 1922-1923, with the crown eventually stabilizing at about 2.96 cents after Rasin’s assassination forced abandonment of his original goal of restoring the pre-war gold parity (Czechoslovakia section, pp. 23-25).

Sargent’s conclusion is that “the essential measures in ending hyperinflation in each of Germany, Austria, Hungary, and Poland were, first, the creation of an independent central bank that was legally committed to refuse the government’s demand for additional unsecured credit, and, second, a simultaneous alteration in the fiscal policy regime,” with the two measures “interrelated and coordinated” so as to bind the government to place its debt with parties who would value it according to whether it was backed by adequate prospective taxes relative to expenditures. He explicitly notes that “it was not simply the increasing quantity of central bank notes that caused the hyperinflation,” since note circulation kept growing after each stabilization, and that earlier, non-regime-changing stabilization attempts – citing Hungary under Hegedus, and earlier attempts in Germany – “failed precisely because they did not change the rules of the game.” In a footnote, Sargent also flags a deep objection to his own interpretation, associated with his and Neil Wallace’s other work, that in a genuinely rational-expectations framework it may be “impossible to conceive…[of] a change in regime” at all, since what looks like a regime change could instead be the realization of events under one more complicated, state-contingent regime; Sargent states he believes the data “could be described using this view” but that doing so “would substantially complicate the language…without altering the main practical implications” (Conclusion, and accompanying footnote).

Sargent presents the four hyperinflations as a direct challenge to the view – which he associates with the “220 billion dollars of annual GNP” estimate then current in U.S. policy discussion – that ending embedded inflation necessarily requires a long, gradual, and very costly disinflation; because each historical hyperinflation ended “abruptly, rather than gradually” and at what he judges a comparatively low cost once a credible regime change occurred, he suggests contemporary policymakers may be overestimating both the time and the output sacrifice required to end more moderate inflation, provided they can make an equally credible commitment. At the same time, Sargent is careful to note, in the paper’s closing lines, that while the four episodes were “extreme and bizarre,” they remain relevant precisely because their extremity makes the underlying “elemental forces that cause and can be used to stop inflation” easiest to identify, and he expresses the hope that “these incidents are full of lessons about our own, less drastic predicament with inflation, if only we interpret them correctly” (Introduction; Conclusion, final paragraph).