Having questioned, in a previous post, some proclaimed benefits of a Strategic Bitcoin Reserve, and specifically the claim that building such a reserve would strengthen the US dollar, I now wish to say something about what the proposed Bitcoin reserve will cost. In particular, I wish to discuss the cost of the BITCOIN Act’s plan to have the US government buy up to a million Bitcoins over the course of the next five years.

Too Good to Be True

“Cost?” you may be thinking; “What cost?” After all, as a recent report on the plan points out, it and some other proposals plans for establishing a Strategic Bitcoin Reserve neither require any increase in taxes nor add to the national debt. “Instead,” the report explains, “the government could leverage its existing resources, particularly gold certificates held at the Federal Reserve’s 12 banks.”

No new taxes? No more government borrowing? Glory and hallelujah! The authors of the BITCOIN Act have discovered a way to treat the taxpaying US public to a one-million Bitcoin (or, at today’s Bitcoin price, approximately $100 billion) free lunch! It seems too good to be true!

Alas and alack, and as usual, it is. Despite not calling for any new taxes or for any addition to the national debt, the BITCOIN Act is no free lunch. Instead, the proposed Bitcoin purchase will actually end up costing the government more than if the Treasury raised the necessary funds by selling securities.

To understand why, we must delve into the details of the BITCOIN Act’s “Bitcoin purchase plan.” That plan, which would have the government acquire up to 200,000 Bitcoin every year for five years, actually consists of three components. First, for five years, it would tap into the Federal Reserve System’s Treasury remittances—the interest the Fed earns on its security holdings minus Fed banks’ interest and operating expenses—devoting up to $6 billion of those remittances per year to Bitcoin purchases. Second, it would transfer $4,425 billion from the Federal Reserve System’s capital surplus account to the Treasury’s general fund, also to be spent on Bitcoin, reducing the Fed’s surplus to less than $2.4 billion. Finally, by changing gold’s official price from its present level of just over $42.22 per fine troy ounce to something like gold’s present market price of almost $2,700 per ounce, and having the Fed monetize the resulting circa $700 billion gain in the official value of the Treasury’s gold stock, it would supply the Treasury with ample funds with which to purchase up to 200,000 Bitcoins per year for five years. Were the Treasury able to buy all those Bitcoins at an average price of $100,000, those purchases would cost $100 billion, or about one-seventh of the Treasury’s golden gain. Put another way, were it to end up having to pay an average price of $700,000 per Bitcoin, the government could still afford to buy a million coins without having to draw on existing tax revenues, raise taxes, or float more debt.

Run Dry

Because the BITCOIN Act mainly relies on the third gold-revaluation method for financing a one million coin Strategic Bitcoin Reserve, I want to devote most of this post to assessing that part of the act’s Bitcoin purchase plan. But let’s first consider the plan’s other components.

The first of those other components—the plan to devote some of the Fed’s Treasury remittances to Bitcoin purchases—can be dealt with quickly because it’s unlikely to be taken advantage of. That’s so because, instead of earning a profit as it almost always used to, thanks to having to pay higher rates on bank reserves than it earns on the long-term securities it gobbled up during past crises, the Fed has lately been losing money hand over fist. As a result, it hasn’t sent the Treasury any money since September 2023, and it won’t do so again until it has first paid down over $200 billion in accumulated paper losses. Since the Congressional Budget Office was already predicting in June that doing so would take until 2030, the BITCOIN Act’s Fed remittance component isn’t likely to fund any substantial Bitcoin purchases.

Budgetary Slight-of-Hand

That the Fed has been running in the red doesn’t altogether prevent the Treasury from taking advantage of it to fund a Strategic Bitcoin Reserve. In fact, all three components of the BITCOIN Act’s Bitcoin purchase plan have the Fed helping out in some fashion, if only unwillingly, while the remittance-diversion plan alone depends on its turning a profit.

Instead of relying on those profits, the Bitcoin purchase plan’s second component has the Treasury grabbing most of the Fed’s capital surplus. Even now, at $6.8 billion, that surplus is paper thin compared to its $29.3 billion level when the FAST (Fixing America’s Surface Transportation) Act was passed in December 2015. Having failed to agree to legislation that might have allowed the FAST Act’s $305 billion price tag to be fully paid by traditional means, Congress chose instead to limit the Fed’s surplus capital to $10 billion. Doing so forced the Fed to immediately fork $19.3 billion over to the Treasury, which added it to the Highway Trust Fund. Two 2018 appropriation bills transferred another $3.2 billion of the Fed’s surplus capital, capping it at its remaining, lowered level of just $6.8 billion.

Although the FAST Act wasn’t the first time Congress treated the Fed’s surplus as a source of off-budget funding—in 1933, it drew on it to provide the FDIC’s working capital—the 2015 measure marked the first use of the Fed’s capital to fund activities having nothing to do with monetary or bank regulatory policy. So it’s pretty obvious that the BITCOIN Act’s plan to once again raid the Fed’s capital was inspired by the FAST Act precedent.

What’s wrong with that? The same thing that was wrong with the FAST Act’s own Fed raid, namely, that even if it did the Fed no harm, it amounted to what former Fed Chair Ben Bernanke called “a form of budgetary sleight-of-hand that would count funds that are already designated for the Treasury as ‘new’ revenue.” As Bernanke explained when the FAST Act was in the works, to actually come up with the $19.3 billion it owed the Treasury, the Fed had to sell securities, thereby reducing its future earnings and future Treasury remittances. “The net effect,” Bernanke says, “is precisely the same as that resulting from the issuance of fresh government debt.” Cato’s Center for Monetary and Financial Alternatives adjunct scholar Jeff Hummel reached the same conclusion while supplying further details. So did a 2002 GAO study of the general fiscal consequences of reducing the Fed’s surplus capital. “Amounts transferred to the Treasury from reducing the [Fed’s] capital surplus account,” the study says, “would be treated as a receipt under federal budget accounting but do not produce new resources for the federal government as a whole.”

Nor is it certain that reducing the Fed’s surplus capital to avoid raising taxes or increasing the national debt does no harm. Although it’s true that fiat-money issuing central banks can operate without capital, as a 2017 GAO investigation reports, academics and Fed officials worry that the practice “might lead the public and financial markets to question if the Federal Reserve was independent from the executive and legislative branches.” The GAO also concluded that such questioning becomes more likely when recurring transfers threaten to eventually reduce the Fed’s capital to zero.

Striking Gold

So we come to the third and most important component of the BITCOIN Act’s Bitcoin purchase plan: the funding of Bitcoin purchases using the proceeds from gold revaluation and monetization. Although the Federal Reserve banks haven’t owned any gold since January 1934, their assets include Treasury gold certificates worth over $11 billion. Those certificates are backed by an equal sum of gold, according to its official price of just over $42.22 per troy ounce. The gold itself is owned by the Treasury and stored at Fort Knox and various other Treasury depositories. Were the BITCOIN Act to pass, gold’s official price would be raised to its market price. Assuming a market price of $2,700 per ounce, that would make the Treasury’s gold officially worth almost 64 times its present value. The Treasury would then swap new gold certificates reflecting gold’s higher official price for the ones now in Fed banks’ possession. Upon receiving the new certificates, the Fed banks would have 90 days in which to “remit the difference in cash value between the old and new gold certificates to the Secretary for deposit in the general fund,” thereby monetizing the Treasury’s accounting profit.

Still assuming gold to be worth $2,700 per ounce, these transactions would yield the Treasury a tidy sum just shy of $700 billion. Were the Treasury able to buy a billion Bitcoin at an average price of $100,000, or somewhat more than Bitcoin’s actual market price as I write this, the gold monetization scheme would give the Treasury seven times the sum needed for the purchase, without the government having to raise more tax revenue and without adding a nickel to the national debt!

Backdoor Borrowing

As I noted before, this sounds too good to be true. The catch is that, although it’s strictly true that the plan calls for no increases in taxes or the national debt, it’s no bargain; because despite not raising the national debt, it turns out to be equivalent to having the Treasury borrow $100 billion (or whatever its Bitcoin purchases end up costing) directly from the Fed (putting-up gold as collateral), and indirectly from the nations’ banks, at interest rates that are higher, and perhaps much higher, than it could get by selling securities.

To see why, let’s consider how the various operations just described affect both the Fed’s and the Treasury’s balance. Once again I’m assuming that revaluing the Treasury’s gold stock raises the value of its gold by $700 billion. To monetize that gain, the Treasury takes back the Fed’s existing gold certificates and gives the Fed new ones worth that much more. The Fed then credits the Treasury General Account (TGA) by the same amount. So we have, in billions:

Federal Reserve:

Assets (Gold Certificates) + $700; Liabilities (TGA Balance) +$700

Treasury:

Assets (TGA Balance) + $700; Liabilities (Gold Certificates) +$700

So far, so good. But the Treasury still has to buy a million Bitcoins. Let’s assume, as before, that doing so costs $100 billion. Let’s also assume, heroically, that the Treasury resists spending any more of its apparent $700 billion windfall. The new dollars spent on Bitcoin get deposited with US banks, increasing their reserves. The changes are:

Federal Reserve:

Liabilities (TGA Balance) -$100; Liabilities (Bank Reserves) +$100

Treasury:

Assets (TGA Balance) -$100; Assets (Bitcoin) +$100

Once again, what the government has done is finance a $100 billion Bitcoin purchase with what amounts to a permanent $100 billion interest-free Federal Reserve loan, collateralized by that much of the Treasury’s increased nominal gold hoard. Although the Fed doesn’t directly charge the Treasury any interest on the $100 billion credited to the TGA, the Fed itself finances that credit by borrowing $100 billion more from the nations’ banks. And those banks do charge interest, as it were, at whatever rate the Fed pays on bank reserves. Since gold certificates earn no interest, the whole operation reduces the Fed’s profits and Treasury remittances by the full amount of those additional interest payments. The overall fiscal burden is therefore much as it might be were the Treasury to borrow $100 billion directly from the banks at the same rate the Fed pays them. But since the lending is financed by bank reserves instead of new Treasury securities.… Hey presto! It isn’t counted as part of the national debt.

Yet national debt it is, economically if not officially. And costly national debt at that. How so? At 4.65 percent, the current interest rate on bank reserves is higher than six-month and one-year Treasury bill rates of 4.26 percent and 4.27 percent, respectively. And it is much higher than the rate on three-year Treasury notes, which is now just 4.08 percent. Nor is the present situation unusual: The interest rate on reserves, being an overnight rate, is generally higher than rates on longer-term Treasury securities. Assuming that the Treasury holds all the Bitcoins it purchases for 20 years instead of selling any to pay off the national debt (the only options the BITCOIN Act allows), during that time banks will have earned, and the Fed will have lost, about $153.5 billion as a result of the government’s Bitcoin investment. Were the Treasury instead to raise the money needed to buy a million Bitcoins by selling $100 billion worth of three-year Treasury notes, the interest cost would be about $124.2 billion, or over $29 billion less.

It’s over $29 billion less, but more visible, for were the Treasury to finance its Bitcoin purchase by issuing more securities, the national debt would go up; and it is precisely in order to avoid having the debt go up, to skip past the ordinary appropriations process, and to otherwise pull the wool over Americans’ eyes, that the BITCOIN Act relies on so much gold-plated hocus pocus. What better way, after all, to gain the public’s support for a plan that may only serve to pump Bitcoin holders’ bags than by making it look like it won’t cost a thing?

This summer witnessed a major new plot twist in the Bitcoin saga. Originally conceived of as a revolutionary grass-roots alternative to established fiat currencies, including the U.S. dollar, the twist has Bitcoin serving not to counter but to fortify the U.S. dollar’s status as the world’s most popular exchange medium.

Although it has somewhat earlier roots, the new vision for Bitcoin gained prominence at this July’s Bitcoin 2024 conference in Nashville, where no fewer than three speakers, including then-presidential candidates Robert F. Kennedy, Jr. and Donald Trump, made proposals based on it.

Like Goldilocks’ Three Bears, the proposals came in three sizes. Trump’s, the most modest, would have the federal government use all of the 210,000 Bitcoins it already possesses—most of them seized by law enforcement agencies—to form “the core” of a “Strategic National Bitcoin Stockpile” that would supposedly “benefit all Americans.”

Kennedy’s plan—the papa bear—would start with Trump's core and add 550 Bitcoin to it daily until the Treasury held at least four million Bitcoin, or more than a fifth of today's outstanding stock. According to Kennedy, his plan, which would equip the Treasury with Bitcoin worth more, at today's market prices, than its gold reserves, would put the United States “at a position of dominance that no other country will be able to usurp.”

The mama bear, finally, was unveiled by Senator Cynthia Lummis (R-WY) just after Trump’s keynote speech. It would have the U.S. Treasury establish a "Bitcoin Strategic Reserve" consisting of one million coins purchased over five years. Unlike the other proposals, Lummis’s has taken shape in actual legislation, The BITCOIN Act of 2024 (the acronym stands for “Boosting Innovation, Technology, and Competitiveness through Optimized Investment Nationwide”) introduced by Lummis a few days after the Nashville event. The proposed reserve's purposes include “strengthening the position of the United States dollar in the global financial system."

Were the idea of creating an official U.S. Bitcoin stockpile only popular with politicians, it might be brushed off as a cynical attempt to tap Bitcoin's popularity for votes. But to judge not just from the rounds of applause the three proposals garnered in Nashville, but from commentary afterward, whether or not Satoshi Nakamoto himself would have approved of them, many Bitcoin fans are convinced that the proposed Bitcoin stockpiles really will do the U.S. dollar, and the U.S. economy as a whole, a lot of good.

Politicians are not the only ones calling for the U.S. government to jump on the Bitcoin bandwagon. In a recent brief, Digital Gold: Evaluating a Strategic Bitcoin Reserve for the United States, the Bitcoin Policy Institute also recommends that the United States establish a Bitcoin Strategic Reserve to serve as "a unique complement to traditional monetary reserve assets such as gold and Treasury securities” that will help ensure “continued dollar dominance.”

Why Reserve Assets?

So it’s goodbye Bitcoin, dollar “terminator”; hello Bitcoin, dollar “supercharger”!

Or is it? Arguments suggesting that an official Bitcoin stockpile can strengthen the U.S. dollar are distinct from those that liken Bitcoin to strategic commodities like petroleum and silicon chips or those that would have the U.S. government include Bitcoin in a Sovereign Wealth Fund. Although calls for a Strategic Bitcoin Reserve tend to conflate these different arguments, I’m only concerned here with the “currency reserve” argument. Would having the government stock up on Bitcoin really “supercharge” the dollar? Would a Strategic Bitcoin Reserve serve the same purpose gold reserves serve in the U.S., or anywhere else? Is it even the case that the dollar needs “supercharging” to preserve its global status?

To answer these questions one must understand the general role that non-domestic reserves play in supporting today’s fiat monies, and the U.S. dollar in particular. Official reserve assets consist of financial assets held by fiscal or monetary authorities that can’t be created by those authorities themselves. In practice today that means foreign exchange (foreign currency itself plus foreign-currency-denominated bank deposits and securities) and gold. As of mid-2024, according to the IMF’s reckoning, the world’s monetary and fiscal authorities held $12,347 billion in foreign exchange assets and 29,030 metric tons of gold, worth around $2.2 trillion.

Why do governments hold reserves? In the days of commodity-based monies, both central and commercial banks needed reserves of the money commodity to meet redemption requests of their customers and other banks. When many nations’ monetary systems are based on the same standard commodity, as was the case during the gold-standard era prior to the Great Depression, reserves are also needed to cover international payment deficits, meaning any positive difference between nations' net foreign capital outflows, exclusive of reserve transfers, and their current account earnings (exports minus imports and net foreign transfer payments).

In today’s irredeemable fiat money systems, reserve assets obviously aren’t needed to redeem either central or commercial bank liabilities. Commercial bank deposits are instead claims to central bank paper currency or, in interbank settlements, central bank reserve credits. Provided they’re willing to let their currencies’ exchange rates vary freely, nations can rely on exchange rate adjustments to eliminate balance of payments deficits instead of having to meet deficits with foreign exchange kept on hand for that purpose.

In practice, however, even nations that issue their own fiat monies often seek to “peg” their currencies’ value to that of another nation’s currency. Small, open economies, for example, often prefer to peg their currency to that of their main trade partner to avoid exposing traders to exchange rate risk. In such cases, foreign exchange once again becomes necessary for meeting the balance of payments deficits. Other countries seek to limit movements in their currencies’ otherwise flexible exchange rates—that is, they prefer “managed” or “dirty” floats to either pegged or freely floating exchange rates. Those countries must also keep stocks of one or more foreign currencies on hand for the purpose.

Gold reserves, on the other hand, no longer serve to settle international accounts. Yet they make up roughly 15 percent of global reserve assets. The main reason for this is that gold is a good hedge against exchange-rate or “currency” risk, meaning the risk monetary authorities incur by holding reserves of foreign exchange. When kept at home, as bullion, rather than with foreign custodians, gold is also free from the political risk to which foreign exchange holdings may be subject, meaning the risk of having foreign exchange consisting of deposits in foreign banks or foreign-government securities frozen or expropriated by foreign governments.

But as we’ll see, a big chunk of the world’s official gold reserves is held for no better reason than sheer inertia. That consists of the 8,133 metric tons of gold—about two-sevenths of the world’s total—held by the United States, almost all of which is left over from a much larger pile the U.S. accumulated during the days when the dollar was still tied to gold.

The U.S. Dollar as a Global Reserve Asset

Although global foreign exchange holdings include the currencies of many countries, U.S. dollars lord over the rest, comprising over 58 percent of the total. Euros run a distant second, at 20 percent. As can be seen from the following Federal Reserve chart, a handful of currencies—the Japanese yen, British pound, Australian and Canadian dollars, and Swiss franc—account for most of the rest. Other currencies, if they’re held in reserve at all, are held in insignificant amounts.

The dollar’s dominant role is no mystery. According to a recent Brookings Commentary, dollars are also used in 58 percent of all cross-currency-jurisdiction payments, meaning international payments minus those among Eurozone nations, as well as in most foreign exchange trading. 64 percent of world debt is also denominated in dollars, including some $13 trillion in U.S. dollar credits to nonbank borrowers outside the U.S. These statistics also explain why so many foreign governments prefer to peg, or at least limit movements in, their currencies’ exchange rates with the U.S. dollar, though doing so obliges them to hold substantial U.S. dollar reserves.

If the dollar’s status is so secure, why are people saying that it needs shoring up? That belief rests upon the fact that since the late 1990s the dollar’s share of total international reserve assets has fallen by a dozen percentage points. Were that decline matched by a corresponding increase in the share made up of euros, yen, or even British pounds—the only other serious contenders for the “dominant currency” crown—and were the shift likely to continue, it might eventually spell trouble for the dollar’s reserve-currency ranking. And were the decline a reflection of a reduced volume of USD-invoiced trade, it might suggest that the dollar is also becoming a less popular medium of exchange. But neither of those things is happening. Instead, the dollar has been losing ground not to the euro or the yen, it’s only serious if still distant rivals, but to various “nontraditional” reserve currencies, including Canadian dollars and Chinese renminbi, and gold. And it is losing not so much because a lot more trade is being invoiced in nontraditional currencies and gold—cross-border renminbi payments, for example, are still dwarfed by USD payments—but for other reasons, including the U.S. government’s “weaponization” of the dollar, meaning its resort to sanctions involving the freezing, if not outright confiscation, of foreign governments’ dollar reserves kept in U.S. financial institutions or in financial institutions in other countries that cooperate with the U.S. authorities.

It should be obvious that building a Strategic Bitcoin Reserve won't relieve foreign governments of the risk of having their U.S. dollar reserves sequestered. What’s perhaps less obvious is the fact that such a reserve would do nothing at all to shore up the dollar’s value or otherwise enhance its popularity.

U.S. Reserve Assets

As of October 2024, the United States Treasury and Federal Reserve System held $245 billion in reserve assets. Besides foreign exchange (over $37 billion worth, of which roughly two-thirds are Euros and the rest Japanese yen), and those 8,133 metric tons of gold (worth $11,041 million according to gold’s official price of just $42.22 per troy ounce, and about $691 billion at its present market price), this sum includes the United States’ IMF reserve position (not quite $29 billion) and its quota of IMF-created Standard Drawing Rights (SDRs) (a bit under $170 billion).

To judge by this total sum, the United States’ reserve asset stockpile puts it in the minor leagues. Despite being the world’s largest economy, that stockpile is only the world’s 15th largest, behind those of Hong Kong, Singapore, and Italy, among others.

Unimpressive as that raw ranking is, because it takes account of IMF reserves and SDRs, which make up over 80 percent of the United States' total reserve assets, it greatly exaggerates the importance the U.S. government attaches to those assets. For while the Treasury and Fed together decide how much foreign exchange and gold to have on hand, the United States IMF reserve position and SDR holdings are set by IMF rules. For example, the IMF periodically sets the total SDR allocation for all of its members, which it then parcels out to them according to their quota shares of the Fund.

The United States’ large SDR stockpile mostly reflects the IMF’s recent total allocation of over 660 billion SDRs (worth about $890 billion) and the United States’ hefty IMF quota share of 17.42 percent. The United States’ IMF reserve position is likewise an obligatory amount, representing a portion of each IMF member’s total mandatory contribution to the fund. Ranking nations’ reserve asset holdings after deducting both their SDR holdings and their IMF reserve positions puts the U.S. in 45th place, below Vietnam, Romania, Columbia and Qatar! And as we’ll see, because it depends on the United States’ outsize gold holdings, even this ranking exaggerates the significance of the United States’ international reserve holdings, for unlike the gold held by most central banks, the U.S. gold stock (which, incidentally, belongs not to the Federal Reserve but to the Treasury) serves no strategic purpose.

That the United States’ reserve assets, and particularly its foreign exchange holdings, are so modest is a reflection of the U.S. dollar’s unique status as the freest of free-floating currencies: other governments may tie their currencies to it, whether tightly or loosely; but so far as the United States government is concerned, maintaining those ties has mostly been, and for some time now has exclusively been, those other governments’ problem.

That the U.S. government can generally afford to let its currency float is one aspect of the “exorbitant privilege” it enjoys as a result of the dollar’s status as both a national and a global exchange medium. As I’ve noted, about 58 percent of all international trade is invoiced in dollars, and dollars make up a corresponding share of official global foreign exchange holdings. A number of countries besides the United States use actual U.S. dollars as their domestic currency. The worldwide demand for dollars, both official and private, makes it unnecessary for the United States government to borrow in other currencies, while the dollar's free-floating status rules out any need for foreign currency to settle U.S. payments imbalances.

The Exchange Stabilization Fund: A Poor Precedent

The status of the US dollar means that it really isn’t necessary for the U.S. government to hold foreign exchange at all. Since the collapse of the Bretton-Woods System in 1973, the United States hasn’t been under any obligation to take part in the maintenance of any international fixed exchange rate arrangement. And while the Federal Reserve has long been authorized to buy and sell foreign exchange, its mandate, as set forth in the Federal Reserve Act, calls for it to promote “maximum employment, stable prices, and moderate long-term interest rates,” but says nothing about stabilizing or otherwise regulating exchange rates.

Even before the collapse of Bretton Woods, responsibility for U.S. exchange rate policy has mainly rested with the U.S. Treasury rather than the Fed, with the Fed assisting the Treasury’s efforts in its capacity as its fiscal agent. (The traditionally joint nature of Treasury-Fed exchange market operations is reflected in the roughly equal division of the nation’s foreign exchange reserves between the two.) This arrangement dates from the January 1934 passage of the Gold Reserve Act, which called for the Fed to surrender its gold reserves to the Treasury in exchange for gold “certificates” in anticipation of gold’s official revaluation from $20.67 per fine troy ounce to $35 per ounce, which continued to be gold’s official price until December 1972, when it was raised to $38. (Some months later it was raised again, to its current level of $42.22 per fine troy ounce.) Of the $2.8 billion nominal profit the Treasury gained by this exchange, $2 billion went to establish an “Exchange Stabilization Fund” (ESF) “for the purpose of stabilizing the exchange value of the US dollar.” Importantly, instead of being subject to the congressional appropriations process, or to any sort of congressional scrutiny, the ESF was to be self-financing and under the exclusive control of the Secretary of the Treasury.

In Digital Gold, the Bitcoin Policy Institute rests its case for a Strategic Bitcoin Reserve partly on the precedent set by the ESF. “The ESF,” it says, “provides the US Treasury with a tool to stabilize currency markets during periods of exchange rate volatility, helping to ensure that the US dollar maintains its value relative to other currencies. This enables the US to intervene in foreign exchange markets, mitigate speculative attacks, and prevent sharp devaluations or appreciations that could disrupt trade balances or financial stability.”

But while the ESF was indeed established to stabilize the dollar’s international value, its exchange market interventions, almost always undertaken jointly with the Fed, have been quite limited since 1995, and it hasn’t intervened at all since March 2011. Why not? First, as we’ve seen, foreign exchange operations haven’t been necessary since the dollar was set afloat in March 1973—a fact officially (if belatedly) recognized by legislation enacted in 1976 that struck out the Gold Reserve Act's reference to stabilizing the dollar’s value and instead made the ESF responsible for undertaking whatever operations the Treasury Secretary deemed “necessary to and consistent with the United States obligations in the International Monetary Fund.”

Although, as the FRED chart below shows, the ESF took advantage of its new—and much vaguer—mandate to intervene frequently in foreign exchange markets between the late 1970s and the mid-1990s, after 1980 its interventions were, more often than not, aimed at weakening rather than strengthening the dollar. But whether they had any lasting effect on the dollar’s exchange value is doubtful, both because the interventions were too mall relative to the sizes of the markets involved to have mattered much, and because even large-scale interventions could only have a lasting effect if the Federal Reserve allowed the Treasury’s exchange-rate goals to dictate the overall course of monetary policy.

And the Fed did no such thing. Instead, its determination to clamp down on inflation during the 1980s ruled out any such subservience. So it happened that, by the 1990s, the U.S. government’s exchange rate interventions were mainly undertaken “out of a spirit of cooperation” with other governments, as when the US intervened to satisfy its part in the Plaza and Louvre Accords of 1985 and 1987, rather than in pursuit of any of US government exchange-rate targets.

Since the mid-1990s, even such “cooperative” interventions have been rare: the U.S. intervened to support the yen in 1998, to support the Euro in 2000, and to stabilize the yen following Japan’s 2011 earthquake and tsunami. But it hasn’t intervened since. In 2013 the United States took part in an agreement among the G7 nations to devote their monetary and fiscal policies to meeting domestic policy objectives instead of targeting exchange rates, and today it seems less likely than ever to change its mind. Far from serving any strategic purpose, the United States’ present foreign exchange holdings are mere leftovers from its earlier interventions.

Despite what one might expect, these developments haven’t caused the ESF—which seems to have more lives than a cat—to give up the ghost. Long before it gave up trying to influence the dollar’s exchange rate with other major currencies, the Treasury found another use for it, namely, aiding less developed nations, particularly in Latin America.

The ESF’s constitution, its vague post-Bretton-Woods mandate, and its ability to “monetize” its SDR and FX holdings (that is, to have them temporarily converted into U.S. dollars at short notice), allow it to make substantial emergency short-term loans without congressional approval. By the 1990s such lending had become the ESF’s principal undertaking. Because the ESF typically made loans by temporarily “swapping” U.S. dollars for foreign currency, they qualified as foreign exchange operations. But stabilizing exchange rates wasn’t their main purpose, and in many cases it wasn’t their purpose at all.

That Congress should have taken a dim view of the Treasury’s use of the ESF as a source of “backdoor” foreign aid is hardly surprising. Matters came to a head in 1995 when Bill Clinton used the ESF to finance a $20 billion Mexican aid package. Afterward, Congress tried but failed to substantially reduce the ESF’s capacity to fund foreign governments. But the public outcry raised by the Mexican intervention was itself enough to convince the Treasury to avoid any further, substantial use of the ESF for foreign loans. So the fund died yet again. And yet again it lived, for the Treasury found still another use for it, namely as a source of rapidly mobilized funds to deal with domestic emergencies. In 2008 it was used to insure money market fund balances, and during the COVID-19 crisis it was used both to backstop the Federal Reserve’s risky emergency lending programs and to finance the CARES Act. Inspired by these examples, the Bitcoin Policy Institute has suggested using the ESF to fund its proposed Strategic Bitcoin Reserve.

Needless to say, an ESF that serves mainly as a rapid-fire way to pay for domestic rescue operations, or to occasionally purchase Bitcoin, or oriental carpets, or Japanese real estate, doesn't need to stockpile foreign exchange. And an agency that has no need for foreign exchange also doesn’t have to hedge against the risks that go hand-in-hand with holding on to substantial amounts of foreign exchange. If the U.S. government wants to avoid losing money on its foreign exchange holdings, the sensible way for it to do so is by disposing of those holdings, gradually or otherwise, not by accumulating some other risky asset.

A Golden Relic

If the United States government has no need for foreign exchange, what, if any, need has it for gold? According to the Bitcoin Policy Institute, besides having “historically…been an important part of US global financial strategy, supporting confidence in the dollar and serving as a hedge against inflation or currency crises,” the U.S. gold stockpile serves “as a last resort financial asset that can be quickly re-monetized in extreme circumstances, providing the US with a historically reliable source of liquidity to address severe financial or geopolitical challenges that disrupt the global monetary order.” Finally, gold reserves allow the government “to subtly influence precious metal markets, ensuring price stability during periods of significant monetary or geopolitical upheaval.”

To each of these claims regarding the benefits of the United States’ gold stock there’s an obvious riposte. Concerning supporting the dollar: “historically” is doing all the lifting here, since the dollar’s value hasn’t depended on the Treasury’s gold holdings since 1971 when it became inconvertible fiat money. (The value of a free-floating fiat currency, like the value of Bitcoin, is a function of the real demand for it and the quantity supplied, not the assets its creators possess). Concerning gold’s liquidity: nothing is more “liquid” than dollars themselves, which the U.S. can create without it. Concerning helping precious metal markets: here, for once, is something a gold stockpile does do. But then the question becomes, what reason is there for the government to prop up the gold market, other than to enrich gold miners and investors at others’ expense?

Far from having been deliberately acquired to serve some strategic purpose, the Treasury’s current gold stock is a legacy of the days, subsequent to the dollar’s devaluation in January 1934, when the U.S. was perceived as a uniquely safe haven for the precious metal. As the FRED chart below shows, between the passage of the Gold Reserve Act and the 1950s the U.S. gold stockpile grew more than six-fold—passively, if not to officials’ dismay—to a peak of more than 20,000 metric tons. During the 1950s, and especially after the Bretton-Woods System became operational in 1958, the flow reversed.

When Nixon closed the gold window in August 1971, the U.S. gold stock was under 8,700 metric tons. Since then, despite serving no obvious purpose, it has changed very little. Yet no statute keeps the Treasury from auctioning off its gold, provided that it redeems a corresponding nominal value of gold certificates held by the Fed. In fact, the Treasury has sold some gold, including about 491 metric tons it disposed of during the late 1970s to take advantage of gold’s then-record price. But when, in June 2011, some economists and politicians proposed selling the rest to get around that year’s debt ceiling, then-Treasury Secretary Tim Geithner replied that besides “damaging” financial markets, the proposed sale would “undermine confidence in the United States." Although Geithner specifically had in mind the effects of what he called a gold “fire sale,” and not those of disposing of the U.S. gold stock in a more gradual and otherwise orderly fashion, the possibility of substantial gold sales hasn’t been raised again since then. In fact there is no reason why the U.S. couldn’t sell-off its entire gold stock provided it did it slowly, much as Canada has done over the last few decades.

Who Needs Another Unnecessary Reserve Asset?

While the U.S. government has resisted reducing its gold stockpile, it has never entertained the idea of adding to it. After all, if the 8,133-plus tons of gold now kept in deep storage in Fort Knox and elsewhere is merely a vestige of the gold-standard and Bretton-Woods days, proposing that the U.S. acquire more gold would be like proposing that, because they have coccyxes, human beings would benefit by growing tails.

So what about a Strategic Bitcoin Reserve? If there’s no reason save inertia for the government’s vast gold stockpile, there’s no reason at all for it to acquire Bitcoin. (A “second-best” argument for having it do so—that adding Bitcoin to its portfolio might lower the risk associated with those gold holdings—founders on the fact that Bitcoin isn’t a particularly good gold hedge.) No reason, that is, unless it’s “to [not so] subtly influence the market” for that virtual precious metal, specifically by driving up its price.

And—let’s be honest—although some Bitcoin fans may sincerely believe that a Strategic Bitcoin Reserve will strengthen the U.S. dollar, others favor it despite not caring a fig about the dollar's future because they expect it to make them rich, and don’t mind if it does so at others’ expense.

Subscribe now

I never thought I’d see the day when free banking seemed like it might become part of a President Elect’s reform agenda. But now that day has come, sort of.

I say “sort of” because although the skimpy Republican Party platform itself says nothing about it, a call for free banking forms part of the Heritage Foundation’s Project 2025, a 900-page book that’s the closest thing we have to a detailed Trump administration reform blueprint. Although Project 2025 isn’t an official Republican publication—the Heritage Foundation, a non-partisan think tank, publishes a similar work every presidential election year, and Trump denies that this one represents his own agenda—the fact remains that quite a few people involved in the project took part in Trump’s previous administration or are otherwise close to him. Besides having been the first Trump administration’s White House budget policy director, Paul Winfree, the credited author of Project 2025’s chapter on “The Federal Reserve,” was also among those Trump considered nominating for an opening on Federal Reserve Board. It’s in that chapter that Project 2025’s call for free banking occurs.

Although the Republican Party platform itself says nothing about free banking, or about banking of any sort, or even about the Federal Reserve, its very first plank calls for “revers[ing] the worst Inflation crisis in four decades.” Concerning just how such a reversal might be accomplished, the platform doesn’t say diddly, forcing us to wonder just what solution the incoming administration might have in mind. That’s where Project 2025 and free banking come in, for Winfrey’s chapter, which is supposed to represent what the best and brightest conservative minds have to say on the subject, puts free banking at the top of a list of “monetary rule reform options” offered “in decreasing order of effectiveness” for combating both inflation and boom-and-bust cycles.

What is Free Banking?

What does Project 2025 mean by “free banking”? How is allowing free banking supposed to limit, or even “reverse,” inflation? Could it actually work?

The expression “free banking” has long been a cause of confusion, because it has more than one meaning. Lawrence White, Kevin Dowd, and I, along with other writers sometimes bunched together as the “modern free banking school,” have followed the European practice of defining it and its foreign equivalents (le banque libre, bankfreiheit, etc.) simply as banking without any special government regulations: according to this view, banking is “free” when all banks enjoy equal privileges, including, most controversially, equal freedom to issue their own circulating paper currencies, subject to no limitations save those implicit in the requirement that they honor the terms of their currency and deposit redemption commitments. Although free banking so defined is a hypothetical arrangement only—no government has ever refrained altogether from regulating banks—several past banking systems, including those of Scotland during the first half of the nineteenth century and Canada for most of the Dominion’s first half century— came close enough to inform general conclusions regarding the likely workings of a genuinely free banking system.

In discussions of U.S. banking history, on the other hand, “free banking” has had a very different meaning. There it has been used to refer to arrangements that provided for relatively free entry into the banking business. In particular, prior to the Civil War eighteen U.S. states passed so-called “free banking” laws, allowing banks to be established according to general rules, with no need for a special legislative charter. But not having to plead for a charter didn’t mean not being regulated. In fact, banks formed under U.S. “free banking” laws were all subject to strict regulations. None could have branches, and all had to back their paper currencies with particular securities instead of being able to back them with the same assets that backed their deposit balances. It isn’t even clear that such “free banking” laws generally eased entry into the banking business. The requirements of ten of the eighteen state “free banking” statutes were so onerous that no banks were ever established under them!

Just as the regulatory requirements of U.S. “free banking” laws differed from those of genuinely free banking, so too did the performance of those U.S. systems differ from that of banking systems that most closely approximated genuinely free banking. While the record of U.S.-style “free banking” systems was mixed, with some working reasonably well and others failing miserably, during their heydays the Scottish and Canadian banking systems were noteworthy for their success at avoiding panics and sparing bank customers from loss. (For this reason, it's frustrating to find otherwise competent economists declaring that “The free banking episodes in the 19th century were ones of considerable instability,” as if they were only aware of “free banking” in its less fortunate U.S. guise.) That the more regulated U.S. arrangements were also less successful may seem paradoxical, but the explanation is straightforward: the regulations imposed on U.S.-style free banks often made them more rather than less likely to fail.

So, to assess Project 2025’s call for free banking, we need to know which sort of free banking the project is proposing. Alas, the answer isn’t clear. Project 2025’s definition of free banking as an arrangement in which “neither interest rates nor the supply of money is controlled by the government” and “banks typically issue liabilities (for example, checking accounts) denominated in dollars and backed by a valuable commodity” doesn’t help: according to it, almost any past commodity-money based banking system might qualify, including ones that denied most banks the right to issue paper currency. Nor does Project 2025 refer to either U.S. state “free banking” systems or any of the arrangements usually treated as case studies by the modern free banking school. The one historical arrangement it does mention, the “Suffolk System” that operated in New England during the decades prior to the Civil War, doesn’t quite fit either definition. The Suffolk System was established in 1824, thirteen years before Michigan passed the first U.S.-style “free banking” law. In it, one particular bank—Boston’s Suffolk Bank—came to serve as a central clearinghouse for the notes issued by other banks in New England. The Suffolk System was remarkable in many respects. By actively returning participants’ notes to them for redemption, it strictly limited their ability to extend credit, and to that extent put a lid on New England’s money stock. It also established the United States’ first “uniform” currency system, keeping all New England banknotes circulating at par within the region. But the Suffolk System wasn’t a free banking system. On the contrary, what made it remarkable was its ability to compensate for restrictions on branch banking and bond-backing requirements that were the main reason why banknotes issued elsewhere in the country could seldom travel across state lines without losing value.

Constraining Inflation

The common ground between Project 2025’s treatment of free banking and the arguments of the modern free banking school concerns the role of active banknote redemption in constraining money creation. The argument here is that, if all banks enjoy equal rights to issue paper currency, then rival banks will treat rivals’ banknotes that come their way like so many paper checks, returning them for payment. Offsetting interbank dues would be netted-out, multilaterally in the presence of a central clearinghouse like the Suffolk System; bilaterally otherwise. But net dues would have to be settled through a transfer of reserves, which back in the day meant gold or silver. As long as reserves were themselves scarce, and no bank had it in its power to create more of them, interbank rivalry would limit systemwide credit expansion, thereby tying-down the price level.

On this point, Project 2025’s understanding aligns with those of myself and other modern free banking school thinkers. But the project goes on to say that

As a result of this stability and lack of inflation inherent in fully backed currencies, free banking could dramatically strengthen and increase both the dominant role of America’s financial industry and the use of the U.S. dollar as the global currency of choice. In fact, under free banking, the norm is for the dollar’s purchasing power to rise gently over time, reflecting gains in economic productivity.

Here, in passing from a context in which bank reserves meant gold or silver to current reality, Project 2025 takes a logical leap. For while it’s true that back in the nineteenth century free banking kept banks on a tight leash, limiting the amount of liabilities that could be supported by a given amount of reserves, the rate of inflation ultimately depended on the scarcity of the reserve medium itself. Had the U.S. been on a silver rather than a gold standard during the last decades of the nineteenth century, instead of falling steadily prices might have risen gradually. And had the U.S. remained on a greenback standard there’s no telling just how much prices would have risen, with or without free banking. A tight leash is, after all, incapable of keeping a dog from wandering if one ties it to something that itself wanders. Just how prices would behave today were free banking of any sort provided for will likewise depend on the scarcity of the bank reserve medium that’s in use.

What Monetary Standard?

Alas, Project 2025 is as vague about what that reserve medium would consist of as it is about the nature of free banking. At one point it says that under its free banking proposal “The Federal Reserve is effectively abolished, and the Department of the Treasury largely limits itself to handling the government’s money.” It follows that those institutions would no longer supply banks with reserves, as the Fed does today and as the Treasury sometimes did in the past. That rules out a fiat or greenback standard. But it doesn’t rule out much else. Instead, Project 2025 seems to have in mind a system that would allow for more than one reserve medium, and (consequently) as many “parallel” monetary standards. “Today,” it says, “we might expect most banks to back with gold, although some might prefer to back their notes with another currency or even by equities or other assets such as real estate. Competition would determine the right mix of assets in banks’ portfolios as backing for their liabilities.”

Such a radical scenario raises more questions than it answers. Were banks left free to make their deposits and circulating media (whether paper or digital) redeemable in whatever reserve asset they wished, one could no longer speak of a distinct “U.S. dollar,” let alone claim that free banking would “strengthen and increase” its use as a global currency. Instead there would be multiple, unofficial monetary standards, like those Friedrich Hayek envisions in his controversial pamphlet on Denationalisation of Money. At very least, Project 2025 needs to make up its mind: are we to have competing monetary standards, and kiss the official U.S. dollar goodbye, or are we to keep that official dollar and somehow strengthen it?

Free vs. Central Banking, Then and Now

Whatever the merits of Hayek’s radical proposal, two things must be said concerning it. One is that the arrangement it envisions is far removed from free banking in any of its historical forms. The other is that it’s unlikely to ever be realized in practice. Both things are true for the same reason, to wit: that banks simply aren’t in the the business of concocting new monetary standards. Their business has always consisted, and is likely to go on consisting, of dealings in established monetary standards. Ordinary banks thrive by getting the public to treat their IOUs as substitutes for standard “outside” monies, meaning monies they themselves don’t create. To do this, they make those IOUs more convenient, for example, by making them out of lightweight pieces of paper rather than heavy bits of metal, and by making sure that their value, unlike that of metal bits, isn’t affected by wear and tear. They also stand ready to redeem those IOUs in the standard money they represent whenever they’re asked to do so. Although regulations may reinforce these outcomes, market forces tend to promote them as well.

It follows that “free banking,” understood to mean freeing banks of most regulatory constraints, can’t itself be relied upon to usher in one or more new monetary standards. Were one to allow U.S. banks today to issue their own circulating currency, whether paper or digital, on any terms they wish, and to otherwise deregulate them, they would still denominate their liabilities in U.S. dollars, meaning the U.S. dollars now supplied by the Federal Reserve System, and they would go on redeeming their own dollars in Federal Reserve dollars. So a free banking reform won’t itself change the nature of the U.S. dollar, and something other than free banking—some constraint upon the Federal Reserve’s capacity to create official dollar reserves—will be needed to rule out, let alone “reverse,” inflation.

Here it’s essential to understand that a free banking reform today no longer has the same implications as such a reform would have had in, say, England ca. 1860. Back then, the Bank of England enjoyed something close to a paper currency monopoly: its notes were legal tender in England and Wales (but not Scotland or Ireland) for sums of £5 or more, and no other joint-stock bank was allowed to issue notes within a 65-mile radius circle centered upon the City of London. Since the passage of Peel’s 1844 Bank Act no new bank of issue of any sort could be established even outside of that circle, and existing issuers had to stay within their average issue for the twelve weeks leading to that act’s passage.

So far as the present discussion is concerned, the Bank of England’s privileged position had one particularly crucial consequence. It caused other banks in England and Wales to treat its liabilities as reserves, as good or (for many purposes) better than gold itself. Doing so exempted it entirely from the discipline of adverse interbank clearings, putting it, and the English banking system as a whole, on a very long leash indeed. When the Bank expanded credit too generously, it led other banks, pied-piper like, to expand along with it. The long leash would run out only once international interest rate or price level differentials inspired a drain of specie to the rest of the world, whereupon the Bank would reverse course to save itself, at the risk of crashing England’s economy.

Under these old English circumstances, a switch to free banking would have had far-reaching consequences. It would have eliminated any incentive for banks to treat the liabilities of one of their rivals (for now they would all be equal rivals) as equivalent to, let alone better than, gold itself. All banks’ notes would have been actively returned for redemption, and all banks would therefore have found themselves constrained by relatively short leashes. The leashes being shorter, the corrections would have come sooner, and the damage resulting from any banks’ misconduct, if there were any worth speaking of, would have been far less severe.

This is why the early nineteenth-century free- vs. central- banking debates mattered. (It is also why Walter Bagehot, who is so often wrongly portrayed at a champion of central banking, actually favored a natural or “many reserve” banking system to a “one reserve” system overseen by a privileged bank of issue.) Back then, the difference was indeed one between a less cyclically-stable system dominated by a central institution and one subject to no central control that was for that very reason better at avoiding both short-run inflation and business cycles. But both options involved the same commodity standard and, therefore, the same ultimate constraint on money creation.

Today, though it still might prove beneficial, a switch to free banking would be far less consequential. For better or worse, champions of central banking won that old debate. Because the central-bank plus-specie-standard cocktail was about as stable nitrocellulose, once central banking became the order of the day, it was only a matter of time before specie standards gave way to inconvertible fiat standards. Once that happened, “free banking” ceased to be an effective, let alone efficient, way to limit inflation, because allowing it still left the established fiat standard in place. “Free” banks would be able to issue their own circulating claims to fiat paper money; but the former fiat money would continue to serve as the system’s reserve medium, and would remain under the central bank’s control. For that reason free banking in the U.S. today would not spell the end of “[p]ublic control of money creation through the Federal Reserve System.”

A Return to Gold?

Although free banking heads Project 2025’s list of possible ways to “take the monetary steering wheel out of the Federal Reserve’s hands and return it to the people,” that list includes several other options, starting with that of reverting to a specie standard by once again making Federal Reserve dollars redeemable in gold. Unlike a move to free banking, and assuming it could be done at all, such a reform would meaningfully constrain the Fed, provided it lasted.

And therein lies a problem. Assume, for the sake of argument, that Congress is determined to go ahead with the proposed change, and that the details have all been ironed out. Forget that the classical gold standard was an international standard, many of the merits of which would not be replicated by a U.S.-only version. The fact remains that the trust that held the old gold standard together died with it: people no longer trust central banks to honor their redemption promises. Hence the notorious vulnerability of fiat-era pegged exchange rate regimes to speculative attacks. If there is any reason for supposing that a Fed-administered gold standard would be immune to such attacks, I have yet to hear it. Even arranging, through a change in gold’s official price and other means, to have every dollar on the Fed’s balance sheet fully backed by gold would not rule out a run on the dollar of the sort that brought the U.S. banking system to its knees in late February and early March, 1933. The merest hint that the dollar might be devalued would suffice to trigger it; and by now everyone knows that, unlike most private bankers, central bankers can devalue their liabilities with impunity.

To its credit, Project 2025 acknowledges the vulnerability of a Fed-administered gold standard. “[T]here is no guarantee,” it says, “that the government will stick to the price peg.” (It adds, less coherently, that “allowing a commodity standard to operate along with a fiat dollar opens both up for a speculative attack,” when in fact it’s the commodity standard itself, and not the presence of a parallel fiat standard, that invites such attacks.) Another passage appears to downplay the problem, by noting that “People would redeem [gold-backed Federal Reserve dollars] en masse only if they feared the government would not be able control itself, for which the only solution is for the government to control itself.” But far from bolstering the case for reviving the gold standard, that passage suggests that a revived gold standard would be otiose!

Fiat-Money Rules

Project 2025 also considers various other monetary rules that would tie the Fed’s hands without altering the U.S. dollar’s current status as an irredeemable fiat money. Unlike a convertible money, a rule-based but otherwise free-floating fiat money isn’t vulnerable to speculative attacks. These other alternatives are therefore likely to prove more durable than a new gold standard. Just how durable will depend on each particular rule’s ability to cope with various sorts of macroeconomic shocks, and particularly with shocks to real output and the velocity of various monetary aggregates.

Addressing the pros and cons of the alternative fiat-money rules Project 2025 considers would require another long essay. The point of this one is to insist that, while free banking might make a good complement to such a rule today, the days when it could be considered a substitute are long gone.