Many mainstream economists today instinctively oppose the idea of the self-regulating gold standard because they have been trained not as economists but as “social engineers”.
Many mainstream economists, perhaps a majority of those who have an opinion, are opposed to tying a central bank’s hands with any explicit monetary rule. A clear majority oppose the gold standard, at least according to an often-cited survey. Why is that?
First some preliminaries. By a “gold standard” I mean a monetary system in which gold is the basic money. So many grains of gold define the unit of account (e.g. the dollar) and gold coins or bullion serve as the medium of redemption for paper currency and deposits. By an “automatic” or “classical” gold standard I mean one in which there is no significant central-bank interference with the functioning of the market production and arbitrage mechanisms that equilibrate the stock of monetary gold with the demand to hold monetary gold. The United States was part of an international classical gold standard between 1879 (the year that the dollar’s redeemability in gold finally resumed following its suspension during the Civil War) and 1914 (the First World War).
Why isn’t the gold standard more popular with current-day economists? Milton Friedman once hypothesized that monetary economists are loath to criticize central banks because central banks are by far their largest employer. Providing some evidence for the hypothesis, I have elsewhere suggested that career incentives give monetary economists a status-quo bias. Most understandably focus their expertise on serving the current regime and disregard alternative regimes that would dispense with their services. They face negative payoffs to considering whether the current regime is the best monetary regime.
Here I want to propose an alternative hypothesis, which complements rather than replaces the employment-incentive hypothesis. I propose that many mainstream economists today instinctively oppose the idea of the self-regulating gold standard because they have been trained as social engineers. They consider the aim of scientific economics, as of engineering, to be prediction and control of phenomena (not just explanation). They are experts, and an automatically self-governing gold standard does not make use of their expertise. They prefer a regime that values them. They avert their eyes from the possibility that they are trying to optimize a Ptolemaic system, and so prefer not to study its alternatives.
The actual track record of the classical gold standard is superior in major respects to that of the modern fiat-money alternative. Compared to fiat standards, classical gold standards kept inflation lower (indeed near zero), made the price level more predictable (deepening financial markets), involved lower gold-extraction costs (when we count the gold extracted to provide coins and bullion to private hedgers under fiat standards), and provided stronger fiscal discipline. The classical gold standard regime in the US (1879-1914), despite a weak banking system, did no worse on cyclical stability, unemployment, or real growth.
The classical gold standard’s near-zero secular inflation rate was not an accident. It was the systemic result of the slow growth of the monetary gold stock. Hugh Rockoff (1984, p. 621) found that between 1839 and 1929 the annual gold mining output (averaged by decade) ran between 1.07 and 3.79 percent of the existing stock, with the one exception of the 1849-59 decade (6.39 percent growth under the impact of Californian and Australian discoveries). Furthermore, an occasion of high demand for gold (for example a large country joining the international gold standard), by raising the purchasing power of gold, would stimulate gold production and thereby bring the purchasing power back to its flat trend over the longer term.
A recent example of a poorly grounded historical critique is provided by textbook authors Stephen Cecchetti and Kermit Schoenholtz. They imagine that the gold standard determined money growth and inflation in the US until 1933, and so they count against the gold standard the US inflation rate in excess of 20% during the First World War (specifically 1917), followed by deflation in excess of 10% a few years later (1921). These rates were actually produced by the policies of the Federal Reserve System, which began operations in 1914. The classical gold standard had ended during the Great War, abandoned by all the European combatants, and did not constrain the Fed in these years. Cecchetti and Schoenholtz are thus mistaken in condemning “the gold standard” for producing a highly volatile inflation rate. (They do find, but do not emphasize, that average inflation was much lower and real growth slightly higher under gold.) They also mistakenly blame “the gold standard” — not the Federal Reserve policies that prevailed, nor the regulatory restrictions responsible for the weak state of the US banking system — for the US banking panics of 1930, 1931, and 1933. Studies of the Fed’s balance sheet and activities during the 1930s have found that it had plenty of gold (Bordo, Choudhri and Schwartz, 1999; Hsieh and Romer, 2006, Timberlake 2008). The “tight” monetary policies it pursued were not forced on it by lack of more abundant gold reserves.
There are of course serious economic historians who have done valuable research on the performance of the classical gold standard and yet remain critics. Their main lines of criticism are two. First, they too lump the classical gold standard together with the very different interwar period and mistakenly attribute the chaos of the interwar period to the gold standard mechanisms that remained, rather than to central bank interference with those mechanisms. In rebuttal Richard Timberlake has pertinently asked how, if it was the mechanisms of the gold standard (and not central banks’ attempts to manage them) that destabilized the world economy during the interwar period, those same mechanisms managed to maintain stability before the First World War (when central banks intervened less or, as in the United States, did not exist)? Here, I suggest, a strong pre-commitment to expert guidance acts like a pair of blinders. Wearing those blinders, even if it is seen that the prewar system differed from and outperformed the interwar system, it cannot be seen that this was because the former was comparatively self-regulating and the latter was comparatively expert-guided.
Second, it is always possible to argue in defense of expert guidance that even the classical gold standard was second-best to an ideally managed fiat money where experts call the shots. Even if central bankers operated on the wrong theory during the 1920s, during the Great Depression, and under Bretton Woods, not to mention during the Great Inflation and the Great Recession, today they operate (or can be gotten to operate) on the right theory.
In the worldview of economics as social engineering, monetary policy-making by experts must almost by definition be better than a naturally evolved or self-regulating monetary system without top-down guidance. After all, the experts could always choose to mimic the self-regulating system in the unlikely event that it were the best of all options. (In the most recent issue of Gold Investor, Alan Greenspan claims that mimicking the gold standard actually was his policy as Fed chairman.) As experts they sincerely believe that “we can do better” by taking advantage of expert guidance. How can expert guidance do anything but help?
Expert-guided monetary policy can fail in at least three well-known ways to improve on a market-guided monetary system. First, experts can persist in using erroneous models (consider the decades in which the Phillips Curve reigned) or lack the timely information they would need to improve outcomes. These were the reasons Milton Friedman cited to explain why the Fed’s use of discretion has amplified rather than dampened business cycles in practice. Second, policy-makers can set experts to devising policies to meet goals that are not the public’s goals. This is James Buchanan’s case for placing constraints on monetary policy at the constitutional level. Third, where the public understands that the central bank has no pre-commitments, chronically suboptimal outcomes can result even when the central bank has full information and the most benign intentions. This problem was famously emphasized by Finn E. Kydland and Edward C. Prescott (1977).
These lessons have not been fully absorbed. A central bank that announces its own inflation target (as the Fed has), and especially one that retains a “dual mandate” to respond to real variables like the unemployment rate or the estimated output gap, retains discretion. It is free to change or abandon its inflation-rate target, with or without a new announcement. Retaining discretion — the option to change policy in this way – carries a cost. The money-using public, uncertain about what the central bank experts will decide to do, will hedge more and invest less in capital formation than they would with a credibly committed regime. A commodity standard — especially without a central bank to undermine the redemption commitments of currency and deposit issuers — more completely removes policy uncertainty and with it overall uncertainty.
Speculation about the pre-analytic outlook of monetary policy experts could be dismissed as mere armchair psychology if we had no textual evidence about their outlook. Consider, then, a recent speech by Federal Reserve Vice Chairman Stanley Fischer. At a May 5, 2017 conference at the Hoover Institution, Fischer addressed the contrast between “Committee Decisions and Monetary Policy Rules.” Fischer posed the question: Why should we have “monetary policy decisions … made by a committee rather than by a rule?” His reply: “The answer is that opinions — even on monetary policy — differ among experts.” Consequently we “prefer committees in which decisions are made by discussion among the experts” who try to persuade one another. It is taken for granted that a consensus among experts is the best guide to monetary policy-making we can have.
Emphasis on a single rule as the basis for monetary policy implies that the truth has been found, despite the record over time of major shifts in monetary policy — from the gold standard, to the Bretton Woods fixed but changeable exchange rate rule, to Keynesian approaches, to monetary targeting, to the modern frameworks of inflation targeting and the dual mandate of the Fed, and more. We should not make our monetary policy decisions based on that assumption. Rather, we need our policymakers to be continually on the lookout for structural changes in the economy and for disturbances to the economy that come from hitherto unexpected sources.
In this passage Fischer suggested that historical shifts in monetary policy fashion warn us against adopting a non-discretionary regime because they indicate that no “true” regime has been found. But how so? That governments during the First World War chose to abandon the gold standard (in order to print money to finance their war efforts), and that they subsequently failed to do what was necessary to return to a sustainable gold parity (devalue or deflate), does not imply that the mechanisms of the gold standard — rather than government policies that overrode them — must have failed. Observed changes in regimes and policies do not imply that each new policy was an improvement over its predecessor — unless we take it for granted that all changes were all wise adaptations to exogenously changing circumstances. Unless, that is, we assume that the experts guiding monetary policies have never yet failed us.
Fischer further suggested that a monetary regime is not to be evaluated just by the economy’s performance, but by how policy is made: a regime is per se better the more it incorporates the latest scientific findings of experts about the current structure of the economy and the latest models of how policy can best respond to disturbances. If we accept this as true, then we need not pay much if any attention to the gold standard’s actual performance record. But if instead we are going to judge regimes largely by their performance, then replacing the automatic gold standard by the Federal Reserve’s ever-increasing discretion cannot simply be presumed a good thing. We need to consult the evidence. And the evidence since 1914 suggests otherwise.
Contrary to Fischer, there is no good reason to presume that expert-guided monetary regimes get progressively better over time, because there is no filter for replacing mistaken experts with better experts. We have no test of the successful exercise of expertise in monetary policy (meaning, superiority at correctly diagnosing and treating exogenous monetary disturbances, while avoiding the introduction of money-supply disturbances) apart from ex post evaluation of performance. The Fed’s performance does not show continuous improvement. As previously noted, it doesn’t even show improvement over the pre-Fed regime in the US.
A fair explanation for the Fed’s poor track record is Milton Friedman’s: the information necessary for successful expert guidance of monetary policy is simply not available in a timely fashion. Those who recognize this point will be open to considering the merits of moving, to quote the title a highly pertinent article by Leland B. Yeager, “toward forecast-free monetary institutions.” Experts who firmly believe in expert guidance of monetary policy, of course, will not recognize the point. They will accordingly overlook the successful track record of the automatic gold standard (without central bank management) as a forecast-free monetary institution.