Presented at the Cato Institute Annual Monetary Conference, 16 November 2011
Suppose for the sake of argument that we all agree to the following proposition: If we could change the monetary regime with zero switching cost, merely by snapping our fingers, we would prefer the US to be on a gold standard. In the most general terms a gold standard means a monetary system in which a standard mass (so many grams or ounces) of pure gold defines the unit of account, and standardized pieces of gold serve as the ultimate media of redemption. Currency notes, checks, and electronic funds transfers are all denominated in gold and are redeemable claims to gold. The question then arises: What would be the least costly way for the United States to make the transition to a new gold standard? We need to choose a low-cost method to insure that the agreed benefits of being on the gold standard exceed the costs of switching over.
Two transitional paths suggest themselves. (1) One path is to let a parallel gold standard grow up alongside the current fiat dollar. (2) The more conventional path, as followed after the suspension of the gold standard during the US Civil War, is to set a date after which the US dollar is to be meaningfully defined as so many grams of pure gold. Or as it is more commonly put, an effective parity is established stipulating so many dollars per fine troy ounce of gold. In our present situation, where Federal Reserve liabilities (book entries and currency notes) and Treasury coins constitute the basic dollar media of redemption, that implies converting the Federal Reserve System's liabilities and the Treasury's coins into gold-redeemable claims at so many grams of gold per dollar (or equivalently so many dollars per ounce of gold).
We see analogs to these two transitional paths when we observe how peso-using countries have made the transition to using the US dollar. In Ecuador in 1998-2000, a parallel unofficial US dollar system emerged as the annual inflation rate in the local currency rose from low to high double-digits, then to triple-digits. The private sector of the economy was already heavily dollarized when the plug was finally pulled on the heavily depreciated local currency unit in 2000. In El Salvador in 2001, the government chose to permanently lock in the dollar value of the currency—by switching from a dollar-pegged exchange rate to outright adoption of the US dollar—while inflation was low and the local currency still dominant. In a nutshell, when the official switch to the harder currency came in Ecuador, it was an act of necessity in the midst of a hyperinflation crisis. In El Salvador it was an act of foresight, to rule out such a crisis.
Allowing a parallel gold standard
Clearing away the legal barriers to a parallel gold standard is fairly simple and can be done without immediately altering existing financial institutions. Rep. Ron Paul's HR1098, the Free Competition in Currency Act of 2011, represents one straightforward approach. It would (1) ensure the enforceability of contracts denominated in units other than fiat dollars be removing legal tender status from Federal Reserve notes and Treasury coins, (2) remove taxes on gold and silver coins that FR notes do not face, and (3) remove sections of the US Code that have been used to criminalize the victimless activity of privately minting distinctive private pieces of metal intended to circulate as money. If these steps seem unprecedented, note that Federal Reserve Notes did not become legal tender until 1933. Bank of England notes are not legal tender today in Scotland or Northern Ireland, where private banknotes (also not legal tender) predominate. Note also that in Switzerland "the purchase and sale of Gold is not subject to taxes (such as value-added tax or capital gains tax) under current Swiss law." 
Further legal and regulatory changes are necessary to allow citizens who adopt the parallel gold standard to have access to gold-denominated banking services. Banking services, including the issue of gold-redeemable paper currency notes and token coins, are of special importance for the success of a gold standard given the awkwardness of making small transactions in physical gold coins. Either existing bank holding companies would have to be free to operate separate gold-denominated subsidiaries, or new gold-based institutions would have to be free to open.
The case for a level playing field between the fiat dollar and other monetary standards rests on the simple fact that the well-being of consumers is better served by competition than by monopoly. Keeping alternatives to fiat dollar at a legal disadvantage, like silver- and gold-backed bank-issued monies, or foreign currencies, limits the options of American consumers to their disadvantage. The option to use an alternative to the fiat dollar is naturally most valuable in an environment of high dollar inflation. Consumers who don't like the ongoing shrinkage of the value of the currency in their pocketbooks and wallets are then not limited to complaining, or trying to lobby the Fed or Congress for better policy, but can "vote with their pocketbooks" to protect their assets by moving into less inflationary alternative currencies.
We should not expect a spontaneous mass switchover to gold, or to Swiss Francs, as long as dollar inflation remains low. The dollar has an incumbency advantage due to the network property of a monetary standard. The greater the number of people who are plugged into the dollar network, ready to buy or sell using dollars, the more useful accepting dollars is to you. Conversely, if you are the first on your block to go shopping with gold coins or a gold-denominated debit card, you will find few stores ready to accept payments in gold. But like the benefit from using dollars in a peso economy, the willingness to accept gold-denominated money in a fiat dollar economy increases with the incumbent currency's inflation rate and its uncertainty. As Gabriele Camera, Ben Craig, and Christopher J. Waller express the general theoretical proposition, "the local currency sustains internal trade if the purchasing power risk is kept very low, but once that risk gets very high substantial currency substitution kicks in."  Should the US inflation rate return to double digits, consumers would find it very helpful to have an alternative currency network available. Potential competition might even help incentivize the Fed to keep inflation low.
Who is likely to produce private gold coins once they are recognized as legal? Gold medallions and biscuits in various sizes, from private producers around the world, are already widely held. Investors in coined gold normally pay a premium over the value of uncoined gold, which covers the cost of coining. In a recent working paper Olivier Ledoit and Sébastien Lotz, two economists at the University of Zurich, raise an interesting possibility while discussing a proposal to allow private gold coinage in Switzerland. They envision that "Gold Francs would be minted by Commercial Banks, [and] the Banks would be allowed to put their brand name and/or logo on one side of the coin. The marketing benefits from having the bank logo in every citizen's wallet would clearly cover any minting costs, so these coins could be sold at par value with the market value of their weight in Gold." It is actually not clear, but remains to be seen, that marketing benefits would cover minting costs. It is true that if the public prefers to use full-bodied coins, gold coins could circulate practically in larger denominations. Historically, however, the everyday circulation of gold coins became rare once people found banknotes more convenient and sufficiently trustworthy. At $1600 per ounce, full-bodied gold coins are completely impractical at perhaps $50 and below.
We can therefore expect most bank-issued coins to be tokens, essentially metallic banknotes, redeemable in gold (upon presentation of a minimum quantity) at the bank. Such tokens can carry the bank's logo, but they will pay for themselves by sparing the issuer the expense of using precious metal in coin production, and will save the system the burdens of incidental wear- and-tear and deliberate shrinkage that accompany full-bodied precious-metal coinage. As with paper banknotes, the float revenue rather than only the advertising value will cover the production and circulation costs. Ledoit and Lotz appear to overlook the standard historical solution to the problem of keeping small currency at par—redeemable tokens and banknotes—because they assume that payment services would be provided only by money warehouses, and do not consider that money-users might be incentivized by banks to prefer the lower-cost alternative of fractional gold-reserve bank liabilities.
Re-establishing a gold definition of the US dollar
The network property of a monetary standard supports the case for not simply legalizing a parallel gold standard, but re-establishing a gold definition for the US dollar. If network effects mean that an uncoordinated piecemeal switchover to a superior standard would not occur except during a painful period of high and uncertain inflation in the incumbent standard, there is a strong case for avoiding that pain through a coordinated switchover before high inflation occurs. That is, we would do well to follow the Salvadoran model of transition rather than the Ecuadoran model.
In considering the re-establishment of a gold dollar now, more than forty years after President Nixon closed the gold window, the question of the appropriate new parity (how many dollars per gold ounce) naturally arises. It is widely recognized that it would be foolish to try to relink the dollar to gold at the pre-1933 parity of $20.67 per ounce, the 1934-71 parity of $35 per ounce, or the post-1972 accounting price of $42.22 per fine troy ounce. It would be foolish because the US price level has risen more than 5-fold since 1971, and the real price of gold has risen in addition, so that $42.22 per ounce or anything lower implies a massive deflation not anticipated in existing nominal contracts. Great Britain's painful deflation during Churchill's ill-considered attempt to return to gold at the pre-War parity, after its high inflation during the First World War, stands as a stern warning. The purchasing power of gold was greater in the rest of the world than in Britain at that old rate, gold accordingly fled Britain, and pound-sterling values faced inescapable downward pressure. Fortunately this point is widely appreciated today, and nobody advocates returning to such a low parity.
By similar logic, it would be foolish to declare a new parity of (say) $8000 per ounce, five times the current price. The result would be a sharp transitional inflation, and a very expensive importation of gold from around the world. Gold would rush in to take advantage of its higher purchasing power in the US, until the US price level rises approximately five-fold, to the point that $8000 no longer buys more than one ounce of gold.
The gold parity that would avoid any transitional inflation or deflation is something close to the current price dollar price of gold. "Close to" because there will be some change in the real demand for monetary gold following the stabilization of the gold value of the dollar. On the one hand, with lower expected inflation, the cost of holding non-interest-bearing money will be lower, and hence the real demand to hold money in the form of M1 dollars will rise. On the other hand, with dollar inflation risk dramatically reduced, the dollar-inflation-hedging demand for gold Krugerrands and Eagles and bullion will fall dramatically. The latter effect is likely to dominate, seeing that hedging demand is the main reason why the real price of gold is higher now than it was when the United States abandoned the last vestiges of gold redeemability in 1971.
Does the US Treasury own enough gold to return to a gold-redeemable dollar at the current price of gold? Yes, assuming that they have what they say they do. At a market price of $1600 per fine Troy oz. (to choose a recently realized round number) the US government's 261.5m ounces of gold are worth $418.4b. Current required bank reserves are only $83b. Looked at another way, $418.4b is 19.9 percent of current M1 (the sum of currency and checking account balances), a more than healthy reserve ratio by historical standards. Combined with the likelihood that US citizens' hedging demand for gold will shrink by more than banks' reserve demand will grow with larger real demand for M1 balances, I expect that the denationalization and remonetization of the US bullion stock at the current price would allow the US economy to export some excess gold. There will be a small transitional windfall for US citizens, getting imported goods and services in exchange for excess gold.
Expeditiously establishing a new gold definition for the US dollar thus requires the following two steps:
1) Withdraw most of the $1.6 trillion in non-required reserves that banks have accumulated since September 2009 by eliminating interest on reserves and selling the mortgage-backed securities that the Fed acquired in QE1, plus enough Treasuries to bring total bank reserves down to the current value of the US government gold stock.
2) Redeem Federal Reserve liabilities with the US government's gold at the then-current market price.
Why not establish 100% reserves for M1?
$8000 is the approximate figure we get if we divide October 2011's M1 ($2105b) by the stock of gold ounces held by the US government (261.5m oz.). Some economists who favor 100 percent gold reserves for currency and checking accounts have offered this approach as the way to set a new parity. As noted above, however, such a high parity implies a large influx of gold from the rest of the world, a large loss of other US wealth in exchange, and a sharp transitional US inflation. The US cannot establish 100% gold backing for currency and checking accounts without great expense. (Even more expensive, because it implies an even higher dollar-per-ounce parity, would be to set the parity by dividing M2 or any broader aggregate by the existing stock of government gold.)
To be specific, at $1600 per ounce of gold, the difference between M1 (about $2.1 t) and the current stock of Ft. Knox gold (about $400b) is about $1.7t. American taxpayers would have to buy $1.7t worth of gold, a very expensive proposition. And that is only the one-time cost. In an economy with 3 percent per annum real GDP growth, assuming a flat trend in the ratio of gold to GDP, a constant purchasing power of gold implies the importation each year of 3% of the gold stock. For a gold stock of $2.1 trillion (100 percent of M1), that would mean an annual expense of $63 billion. With a 20 percent (or alternatively 2 percent) fractional reserve against M1, the annual expense would be one-fifth (or one-fiftieth) of that figure.
It should also be noted that with 100% reserves, the historically familiar sort of currency, circulating redeemable private banknotes and token coins, are infeasible. A money warehouse would be unable to assess storage fees on anonymous currency holders. Debit cards would still be feasible, but the warehouses issuing them would have to charge storage fees.
What about the central bank?
Because the nation's stock of money becomes endogenous under a gold standard, no monetary policy is needed. Retaining a central bank committee to "manage" the gold standard undermines its automatic operation, creates uncertainty by opening the door to policies that lead to devaluation or suspension, and thus does more harm than good. A central bank inevitably faces political pressures to pursue monetary policies inconsistent with redemption for gold at a fixed rate. It can endanger or suspend redemption so with legal impunity, and it faces no competitive pressure to maintain its reputation. When the central bank runs a policy inconsistent with maintaining the gold standard, typically the gold standard gives. Competing private banks, which do face legal and competitive constraints, have a better historical track record than central banks for maintaining gold redemption. The classical gold standard of 1879-1914 functioned quite well without a central bank in countries like Canada that did not weaken their commercial banks with legal restrictions. Even in the United States, despite several financial panics that (to judge from the Canadian example) could have been avoided by banking deregulation, the business cycle was not worse than it has been under the Fed's watch since 1914. 
Nor does the gold standard require a central bank for other purposes. Many of the banks that issue checking accounts may also be relied upon to issue gold redeemable circulating currency notes, as they did before the Federal Reserve monopolized banknote issue, and token coins. The Fed's other useful functions can be returned to private clearinghouse associations, namely the clearing and settlement of payments, the setting and enforcement of membership standards for solvency and liquidity, and the last-resort lending of temporary liquidity support to solvent member banks. Because their members' own money is at stake and they cannot simply print fiat money, clearinghouse associations do not and cannot bail out insolvent banks at taxpayer expense, whether through direct capital injections, asset purchases at above-market prices, or loans at below-market rates.
The journalist Martin Wolf has written that "the obvious form of a contemporary gold standard would be a direct link between base money and gold. Base money — the note issue, plus reserves of commercial banks at the central bank (if any such institution survives) — would be 100 per cent gold-backed. The central bank would then become a currency board in gold, with the unit of account (the dollar, say) defined in terms of a given weight of gold." Actually, although irredeemable central bank notes are base money today, under a gold standard only coined gold and bullion reserves are base money. Notes in circulation are redeemable liabilities of the issuers and not part of actual or potential bank reserves. And although a currency board is less likely than a central bank to undermine the gold standard, there is no need for either. The most efficient form of a contemporary gold standard makes gold the base money, i.e. the medium of redemption and unit of account, while currency and other common media of exchange are the fractionally backed gold-redeemable liabilities of commercial banks. Wolf rightly recognizes that "It is wasteful to hold a 100 per cent reserve in a bank, if depositors do not need their money almost all of the time," but does not draw the obvious conclusion that "a currency board in gold" is therefore less efficient than fractional-reserve banking under a gold standard.  Wolf expresses the common worry that "Such a system is unstable. In good times, credit, deposit money and the ratio of deposit money to the monetary base expands. In bad times, this pyramid collapses. The result is financial crises, as happened repeatedly in the 19th century." But the banking system is more robust than he suspects, as seen in Scotland, Canada, Sweden, and other less-regulated systems without central banks under the gold standard. Repeated financial crises were a feature of the nineteenth-century banking systems in the United States and England, weakened as they were by legal restrictions, but not of the less restricted systems elsewhere.
Barry Eichengreen's recent critique of reinstating the gold standard
In a recent critique of proposals for reinstating a gold standard, the economic historian Barry Eichengreen has repeated the often-made but nonetheless absurd claim that a gold numeraire is equivalent to a commodity price support, writing: "Surely a believer in the free market would argue that if there is an increase in the demand for gold, whatever the reason, then the price should be allowed to rise, giving the gold-mining industry an incentive to produce more, eventually bringing that price back down. Thus, the notion that the U.S. government should peg the price, as in gold standards past, is curious at the least." Surely Professor Eichengreen understands that if there is an increase in the demand for gold under a gold standard, whatever the reason, then the relative price of gold (the purchasing power per unit of gold over other goods and services) will in fact rise, which will in fact give the gold-mining industry an incentive to produce more, which will in fact eventually bring the relative price back down. That one unit of gold continues be worth one unit of gold does not involve the pegging of any price.
"More curious still," Eichengreen continues, "is the belief that putting the United States on a gold standard would somehow guarantee balanced budgets, low taxes, small government and a healthy economy." Of course "guarantee" is too strong a term, and a budget balanced each and every fiscal year is not the right goal. But a gold standard does help to ensure budget balance in the desirable present-value or long-run sense, by requiring a government that wants to sell its bonds in the international market to stay on a fiscal path consistent with full repayment in gold.
"Most curious of all" to Eichengreen "is the contention that under twenty-first-century circumstances going back to the gold standard is even possible." This time is somehow different, apparently. But going back to the gold standard by re-establishing a dollar-gold parity requires today only what it has always required: (1) a sufficient real gold stock, which as shown above the US government already has on hand, and (2) the political will to do so. Developing a parallel gold standard, using present-day technologies for money transfer, would probably be easier today than it has ever been.
 For the generic definition and supply-demand analytics of a gold standard see Lawrence H. White, The Theory of Monetary Institutions (Oxford: Basil Blackwell, 1999), ch. 2.
 See Lawrence H. White, "Statement on HR 1098, The Free Competition in Currency Act of 2011," http://financialservices.house.gov/UploadedFiles/091311white.pdf.
 Olivier Ledoit and Sébastien Lotz, "The Coexistence of Commodity Money and Fiat
Money, University of Zurich Department of Economics Working Paper No. 24 (August 2011), p. 2.
 Gabriele Camera, Ben Craig, and Christopher J. Waller, "Currency competition in a fundamental model of money," Journal of International Economics 64 (Dec. 2004), pp. 535–36.
 Ledoit and Lotz, op. cit., p. 5.
 Note: In counting all the gold as bank reserves I'm assuming that coins in circulation would become redeemable tokens, not become full-bodied gold coins. The current numbers update Lawrence H. White, "Will the Gold in Fort Knox Be Enough?" in Prospects for a Resumption of the Gold Standard: Proceedings of the E. C. Harwood Memorial Conference [Economic Education Bulletin vol. 44, no. 9] (Great Barrington, MA: American Institute for Economic Research, 2004), 23-32.
 The Fed's gold certificate entry as reported on its balance sheet (H.4.1, 6 October 2011) is $11,041 million, the product of the bookkeeping price of $42.22 times 261.511 million oz. Au. See also Federal Reserve Bank of New York, "The Key to the Gold Vault" (2008), p. 17, which notes: "A majority of these reserves are held in depositories of the Treasury Department at Fort Knox, Kentucky, and West Point, New York. Most of the remainder is at the Denver and Philadelphia Mints and the San Francisco Assay Office." I ignore the US share of IMF gold.
 Lawrence H. White, "Accounting for Fractional-Reserve Banknotes and Deposits," The Independent Review 8 (Winter 2003), pp. 423– 441.
As Alan Greespan told Jon Stewart on the Daily Show, 19 Sept. 2007: "You didn't need a central bank when we were on the gold standard, which was back in the nineteenth century. And all of the automatic things occurred because people would buy and sell gold, and the market would do what the Fed does now."
 George Selgin and Lawrence H. White, "Credible Currency: A Constitutional Perspective," Constitutional Political Economy 16 (March 2005), pp. 71-83.
 George A. Selgin, William D. Lastrapes, and Lawrence H. White, "Has the Fed Been a Failure?," Cato Institute Working Paper no. 2 (November 2010), forthcoming in Journal of Macroeconomics.
 Martin Wolf, "Could the World Go Back to the Gold Standard?," Martin Wolf's Exchange blog (1 Nov. 2010), http://blogs.ft.com/martin-wolf-exchange/2010/11/01/could-the-world-go-back-to-the-gold-standard/#axzz1aPjGlRV6
Wolf incidentally remarks that "Economists of the Austrian school wish to abolish fractional reserve banking," but this is true only of a fraction of Austrian-school economists.
 See Kevin Dowd, ed., The Experience of Free Banking (London: Routledge, 1992); also George Selgin, "Bank-lending 'Manias' in Theory and History," in Selgin, Bank Deregulation and Monetary Order (London: Routledge, 1996).
 Barry Eichengreen, "A Critique of Pure Gold," The National Interest (Sept. –Oct. 2011), http://nationalinterest.org/article/critique-pure-gold-5741.
 The recent Nobel laureate Thomas Sargent made this point in "An Interview with Thomas Sargent," The Region (Federal Reserve Bank of Minneapolis), September 2010.