The Case for Open-Market Purchases in a Liquidity Trap

The paper’s thesis is that even if open-market purchases of government debt are completely ineffective at moving prices or output – the conventional “perfect substitutability” result for a zero-interest-rate liquidity trap – there remains “a powerful argument for large-scale open market operations as a fiscal policy tool,” and that the same conditions which make this fiscal argument work also make open-market operations effective for conventional macroeconomic stabilization (Abstract; Introduction, pp. 1-2). The mechanism is that converting interest-bearing public debt into money reduces future debt-service requirements and hence the distortionary taxes needed to service that debt, provided long-term interest rates are positive now or short rates are expected to be positive at some point in the future.

The authors argue that Japan’s upward-sloping term structure in the early 2000s – short rates near zero but 20-year government bond yields around 1 percent – is evidence that market participants attach some positive probability to Japan eventually escaping the liquidity trap, because none of the standard alternative explanations for an upward slope (risk premia from interest-rate volatility, default risk, or liquidity effects) is plausible if short rates are expected to stay at zero forever (Sec. II, pp. 4-6). If short rates could never rise, bonds of all maturities would be riskless perfect substitutes for money and the entire term structure would be flat at zero; since it is not flat, “the only plausible explanation… is that investors attach some positive probability to Japan’s some day having positive nominal short-term interest rates” (p. 6). This empirical premise is what licenses the model’s central assumption.

In the flexible-price version of the model, the path of nominal wages (and, with a constant markup, prices) before the economy exits the liquidity trap is pinned down by backward induction from the money stock expected to prevail at the first future date T when the nominal interest rate turns positive, so that a permanent, immediate increase in the money supply raises the entire price path, including today’s price level, even though the interest rate remains at zero throughout the intervening periods (Sec. IV, pp. 10-14, derivation via the “inverse wage Euler equation,” eq. 5-9). The authors describe the intuition as “disarmingly simple”: on the date rates first turn positive, the money stock pins down the price level conventionally, and prices before that date move toward that anchor at a rate governed by the discount rate, so raising the money stock’s level today shifts the whole terminal condition, and hence the whole price path, upward (p. 13).

The authors describe their model as “in essence just a dynamic extension of Krugman’s (1998),” but argue that Krugman’s conclusion that monetary policy is powerless rests on the specific assumption that the central bank cannot credibly commit to a permanent increase in the money supply level, because it is assumed unwilling to tolerate any permanent rise in the price level – an assumption the authors call “implausible as a total brake on policy effectiveness” (Sec. IV, p. 14; Sec. IX, pp. 32-33). Their own simulations show “there is no welfare gain to reversing a permanent money-stock increase… on the contrary, such a move would have a negative welfare impact,” so in their framework a central bank that internalizes the fiscal benefits of debt reduction has no incentive to renege on a promised expansion, making credible commitment far more plausible than Krugman’s account implies.

The two preconditions are, first, that long-term nominal interest rates be positive at some horizon, and second, that the central bank be able to carry out credibly permanent increases in the level (not merely the growth rate) of the money supply (Introduction, p. 2). The authors state the first condition “does hold in Japan today” based on the term-structure evidence in Section II, and argue in Section IX that the second is plausible for the Bank of Japan given its statutory mandate to pursue “price stability, contributing to the sound development of the national economy” alongside close coordination with the government – a mandate the authors read as “not at all inconsistent” with accepting some permanent reflation once its fiscal benefits are understood (Sec. IX, pp. 32-33).

A rough analytical calculation using Japan’s estimated “potential national burden” tax rate of 47 percent of National Income implies a permanent welfare gain of nearly 0.05 percent of National Income per year for each 1 percent increase in the money stock at the date the liquidity trap ends, which the authors call “a huge marginal welfare benefit” (Sec. VII, pp. 23-25). Numerical simulations of the full calibrated model, allowing for sticky prices and markups, find that one-time permanent money-stock increases of 1, 10, and 50 percent at date 0 yield welfare gains of 0.09, 0.84, and 2.95 percent of output per year respectively (Sec. VIII, pp. 29-30) – gains that are sizable but, as the authors note, “not linear in money stock changes”: the 50-percent-increase gain is only about two-thirds of what linear extrapolation from the 1-percent case would predict.

Raising the annual money growth rate from 2 to 7 percent through period 5 – which ends the liquidity trap immediately, since the interest rate jumps near its long-run value right away – produces a welfare gain of 0.92 percent of output per year, only slightly more than the gain from a one-time 10 percent increase in the money stock and substantially below the 2.25 percent gain that a one-time 33 percent increase (the cumulative money-stock difference by period 5) would deliver on its own (Sec. VIII, pp. 30-31). The authors attribute this to the gradual policy’s immediate effect of raising interest rates, which weakens the short-run Keynesian output surge that an unexpected inflationary burst produces under sticky prices.

Under staggered two-period nominal price-setting, an unanticipated open-market purchase remains expansionary and retains the fiscal debt-devaluation benefit, but now has an additional positive welfare effect through a temporary Keynesian increase in output, since the unexpected price-level jump is not immediately matched by proportional nominal wage and price adjustment (Sec. VI, pp. 18-23). The authors note there are also welfare costs from the relative-price dispersion that staggered pricing introduces under both expected and unexpected inflation, but judge these “unlikely to offset the primary gains” (Introduction, p. 3).

The authors argue the banking channel does not undermine their result, because what matters for the price-level effect is the money stock expected to prevail once the economy exits the liquidity trap at date T, not whether banks immediately lend out the new reserves; banks may rationally “hoard” reserves while short-term rates are zero, but should be expected to lend once positive rates return, since they would then profit from the spread between reserve and loan rates (Sec. X, pp. 36-37). They add that to the extent expansionary policy helps distressed banks write off bad loans and resume normal lending, this would only strengthen the case for central bank bond purchases.

The authors point to the United States in the late 1930s, where interest rates showed a term-structure pattern similar to Japan’s (near zero at the short end, meaningfully positive at long maturities), and where dollar devaluation in 1933 followed by large gold inflows roughly tripled the stock of high-powered money between 1933 and 1940; citing Romer (1992), they note this monetary expansion is widely credited as “the main cause of U.S. recovery from the Depression,” including the sharp 1938-1942 rise in output (Sec. IX, pp. 34-35). They conclude that, by this historical analogy, Japan may need considerably more monetary expansion to fully escape its liquidity trap, with the added motivation of the fiscal gains from public-debt reduction.